[lazarus] Forcing fonts
Hi all, Okay, the WIKI states that font setting is very limited (unless you use canvas drawn components) since GTK overrrides your settings. Is there ANY way I can get around this ? I am getting desperately annoyed that after I create a really good looking program, the first time a user changes his desktop theme, the captions no longer fit on the buttons :( All my apps now use GTK2. I have tried to hack gtkrc-2.0 by hand - no luck there, it seems to ignore the font line entirely. Resizing the buttons is not an option, I HAVE to support 640x480 resolution and the buttons only just manage to all fit on screen as it is. -- A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Forcing fonts
I would say a more sane approach is that application settings should override theme defaults. I know the gnome guys have this thing about deciding how every human being on the planet ought to think but surely they didn't make it impossible to do this ? Did they ? The gnome guys think, that fonts and themes should be controlled by the user, i.e. every human being on the planet, not by the application, i.e. programmer. Oh I've heard their arguments, I even agree with them. It's their METHODS I don't approve off. I think they go about their goal of consistency in completely the wrong way and fail exactly at what they set out to do - controll by the user. Where KDE gets it right is letting the user configure everything, then sticking with it as DEFAULT which the user can overwrite PER APPLICATION because not all apps do the same tasks and thinking that there can be a completely universal good standard is just stupid. Some apps NEED to look different in order to actually work usably. I have actually had a lead gnome dev (who shall remain nameless) say: any setting that would only be usefull to one application must never be implementable. This I dissagree with 100%. As far as I am concerned, too many options for configuration is a lesser crime than too few, and every real user agrees with me - and I have a rather particular point of authority in this. I have thousands of users, and 95% of them were first time computer users when they touched my system for the first time. They all had the option of gnome or KDE, and they all tried both, and they all preffered KDE - because KDE could look the way they wanted it to. In the end TRUE userfriendlyness is recognizing that every user is unique, and allowing every bit of the system to be adaptable to the USER's preffered way of working - when you have that the computer learns the user and no longer vice versa. In fact, before I began on OpenLab 4 I did a survey among my paying customers, since not a single user was using the gnome desktop by choice. I won't be including it in OpenLab4 - I see no point of supporting something my customers don't want anyway. The gnome libs are there, cos there are some cool gnome apps. But that's it. A.J. -- A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Forcing fonts
For starters I created in olpack(http://silentcoder.co.za/viewsvn/viewsvn.php?project=lazaruspath= %2Folpack%2F) the screensize component, these particular apps, by their very nature are used alone, so they run maximized. Since lazarus doesn't (yet?) obey ws_maximized, I use the screensize to get the Tell me how the code to set ws_maximized looks, and I can check if the gtk backend of lazarus handles it, and fix it. Is it just via windowstate ? That's all I've tried, windowstate in the object inspector. Bear in mind I've never actually USED delphi. I first learned a bit GUI programming in Java around 5 years ago before then I had spent about 10 years doing nothing more GUI than ncurses :p I got sick of java FAST ! So I actually wrote the freepascal support for tcl/tk and maintained it for some time (nobody else seemed very interested though) so I stopped bothering, haven't actually even looked at that code for about 3 years. Anyway, at that point I had no use for doing any GUI programming in my career that couldn't be handled by the odd bit of SDL. Things changed and suddenly the same job was involving a LOT of GUI coding, so I studied the principles of it. When it came to actually using a language though, I opted for the one I knew best (pascal based, been using it since I was 9 years old after all) and chose Lazarus. Been doing that ever since. Effectively though, what I know about how to actually CODE GUI's I learned from using a delphi book and playing with lazarus by myself. Oh and constantly reading this mailing list ! A.J. -- A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] forcing a component to repaint itself
Okay, I have a small problem with my tfunkymemo component, it doesn't repaint after loading a new file into it. This is a bit of an issue if you look at what I use it for (what I CREATED it for) which is live help in the app that changes as you navigate. The help files load perfectly, but because it's canvas drawn it doesn't show until the form is repainted. moreover since it's on a sepperate panel to where the actual navigation occurs, it doesn't seem to get repainted by events in the app. If you minimize and then maximize the screen it does update though. I tried calling form1.paint and form1.panel1.paint after the loadFromFile, neither had any effect (not to mention this should really be handled in module code) - how can I force the component to repaint itself after loadFromFile is called ? I honestly don't know which bit of the component source is relevant to snip, so I'll just link to all of it (asuming that it may be needed): http://silentcoder.co.za/viewsvn/viewsvn.php?project=lazaruspath=%2Folpack%2Ffunkymemo.pas TIA for any advice. A.J. -- A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Bugs ? Startlazarus increases link times
Right, I can't be certain the drive is fine of course, but it is brand new - I've have this laptop less than a month. It isn't a great work horse but it's all I got until my desktop CPU comes back from repairs. That said, dmesg and messages are both empty, and it does finish, machine has 180mb of ram. A.J. -- A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Bugs ? Startlazarus increases link times
Ok some stats from my setup then: Lazarus version: Subversion checkout on saturday (same happened with 0.9.8 however) CPU: AMD Sempron 2800+ CPU Mhz: 1600 Acer Aspire 3000 ld version - from binutils 2.15.92.0.2 *cast confusion. on . A.J. On Tuesday 30 August 2005 20:01, L505 wrote: Tested linking times (not including compiling time) Lazarus: - lazarus.new.exe 2 minutes : 37 seconds StartLazarus - lazarus.new.exe 2 minutes : 54 seconds Celeron 400mhz, 327MB RAM, Windows 2000 Lazarus Version: 0.9.7 RB Alpha LD.exe version: not sure, it is 279,552B in size -- A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Menu icons in KDE / Gnome
That's highly distribution-specific. There are some efforts to make a standard, though. For Fedora: install the desktop-file-utils. Make a desktop-file (see an example in lazarus/install/gnome.ide.desktop) and then run desktop-file- install desktopfile.desktop. (this also makes a kde-icon and menu-entry) Aahh, a question the distribution developer gets to answer :) Actually, the freedesktop standard is very specific, although gnome fails to follow it completely unless you tweak a bunch a stuff, KDE does follow it perfectly so the easy way for the programmer is to install for gnome since KDE will pick it up. KDE and gnome will both check /usr/share/applications/ for desktop files, so that's where your desktop file should go, to prevent inter-distro issues, install your icon picture somewhere global as well, and set a full path in the desktop file. Finally, the crunch is that to get your menu item in the right place, you need to use categories, the reason for those are that the distro's can rename menu sections, but the application entries will be still be sorted like-by-like. The easy way to do those, are to drag an item from the same part of the menu you want to be in onto your desktop, and look at it with a text file, the categories are usually the last line, just copy and paste it into the bottom of your own file. A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] TMemo only shows English
Help. I'm working on integrating the translations for the OpenLab installer now a.t.m. unfortunately they are not working. All the files are in plain text, which is the same I had used for earlier versions of OpenLab, except now the installer is using GTK2, whenever I try to load one of the translations, it just shows an empty window though, in the shell I get a lot of errors from Pango complaining about invalid UTF-8 strings. How do I fix this please ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: Help - TMemo shows only english text
So nobody has an answer :( Oh well, I'll just redo them as TFunkyMemo's, that handles every language I've thrown at it so far. Was hoping to avoid that though, it's a lot of overtext for plain text. A.J. _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: Help - TMemo shows only english text
On Thursday 08 September 2005 00:27, Marc Weustink wrote: A.J Venter wrote: So nobody has an answer :( Oh well, I'll just redo them as TFunkyMemo's, that handles every language I've thrown at it so far. Was hoping to avoid that though, it's a lot of overtext for plain text. TBH, At first I had no clue of what your problem was. THere are only Lower ascii chars in you lang files or do they have some ISO encoding ? In the latter case, what is yout LANG setting ? The files are plain text, but almost certainly not plain ASCII, after all they contain other languages, thus far French, Afrikaans and Portoguese. They show up perfectly in vim and also in ever graphical editor I've tried. The LANG setting is probably on EN-US and would be a tricky one to change as the language is supposed to be altered live by a selection during the installer. My current design is simple, each language has all the message files in a tarbal named after the language, I read all the file names and show a drop down list, the user chooses, I extract his chosen selection and and move on with the next dialog. The actuall installer is written in bash, but I use lazarus created dialogs to ask questions, show menu's etc. Simliar to using something like xmessage actually, but I get to make all the dialogs look alike and follow the OpenLab GUI designs. I'm open to suggestion here, though I'm getting close to my deadlines :( - trouble is I couldn't even ASK for translations until all the messages were finalized (for obvious reasons) so their the last thing to happen, and I wasn't expecting difficulties. Is there some way I can tell a TMemo to handle UTF-16 instead ? Possibly with a fileheader ? I'm pretty sure there's nothing in there that can't be displayed by UTF-16, it's not like I've had any translations to asian languages. A.J. _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] TListBox - Multiselect
Hi all, A Quick question, I know how to find out which item in a TListBox was select with single selection. Now however, I wish to do a multi-selecting one, how can I determines which ITEMS were selected. Ideally I need to be able to feed all the selected items into a loop and step over them one by one. Does A TlistItem have a ''selected'' property ? Would doing a for loop over the TListBox.Items testing for selected work ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListBox - Multiselect
A listbox also has a selected property. You can check if an item is selected by checking the corresponding selected instance. Like: begin for I := 0 to (ListBox1.Items.Count - 1) do begin if ListBox1.Selected[I] do ShowMessage(ListBox1.Items[I] + ' is selected'); end; end; Thanks, this is exactly what I needed, now a related question, how can I determine the LAST thing that was selected ? In the order the user clicked it, I need to hook up something to the onSelectionChange event that will let me show a description of the current selected item, and if another is selected, change to the description for the new one etc. etc. A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Re: Graphics basics on gtk
The timer problem is non-graphics related. I would like to receive data from the ISA card at 500khz (or I will lose data). But thank you all for your help so far. I will research if there is a way to create a very efficient timer. If not, I'll just use another thread. Felipe, you may also want to look at gamepack, which provides a double-buffered drawing component as well as sprite support. Mail me off-list if you want the details. Ciao A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Re: Graphics basics on gtk
On Sunday 02 October 2005 04:01, Felipe Monteiro de Carvalho wrote: A.J. Venter wrote: Felipe, you may also want to look at gamepack, which provides a double-buffered drawing component as well as sprite support. Mail me off-list if you want the details. Sounds very interesting. TDoubleBuffer seams very similar to what I am trying to do. I must be doing a simple error ... I really am going to write a good example program about this once I have time and get this right. But why don't you give the details on the list? This way they might benifit someone reading the mail archives... Because I have given it on-list before and don't wish to be accused of spamming. Since you ask though, you can check it out from SVN with svn co svn://silentcoder.co.za/lazarus/gamepack Or you can browse the code at http://www.silentcoder.co.za/viewsvn It's under the lazarus license. A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TApplication.Terminate
On Tuesday 25 October 2005 17:42, Alain Michaud wrote:
Hi,
I had the same problem a while ago. There were some discussions on
this list and this seems to be a deep bug.
In my case, I noticed that any timer runing would not stop.
Finally I found the ultimate weapon: I send the kill command to the
shell.
1 -find the PID of your program
2 - send exec(kill pid);
(something like that. I can give you the exact code if you want
It works!
I've had the same bug, but I think THAT answer lacks a certain elegance as an
interim measure, I chose to remember some of my old pascal, and my close
handlers look like this:
{Insert code to try and cleanup as much as possible here}
application.terminate;
halt;
The halt command of course is technically a pascal command but it DOES seem to
work. The only thing I'm not sure of is what it actually DOES under linux,
hopefully it uses signals. I do know that it's alternative form
[halt(status:integer);] is apparently the only way to set an exit status
(granted that is more commonly needed for commandline than GUI apps however).
Ciao
A.J.
--
80% Of a hardware engineer's job is application of the uncertainty principle.
80% of a software engineer's job is pretending this isn't so.
A.J. Venter
Chief Software Architect
OpenLab International
http://www.getopenlab.com
http://www.silentcoder.co.za
_
To unsubscribe: mail [EMAIL PROTECTED] with
unsubscribe as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Incredibly weird canvas.TextOut bug
On Wednesday 09 November 2005 04:22, A.J. Venter wrote: Hi all, I'm totally stumped by this one, for some reason, within my one custom component (a TFunkyMemo) the textouts are working just fine, however, those on the OTHER side of the screen (which are done through tsprites on top of tdoublebuffer) are behaving in a seriously weird way. Only the one on the far left shows up. This happens in BOTH gtk1 AND gtk2 on linux. mmm, I realized later I was unclear here, when I say only the one on the far left, I do NOT mean the TFunkMemo's contents, I mean only the text printed on the far left SNOWMAN shows up. A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com http://www.silentcoder.co.za _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TrayIcon preview
Hi Felipe, Thank you for this, something I've been VERY much wanting in lazarus was trayicon support. A few quick questions though 1) Is it a component ? Visual ? Nonvisual ? I admit I've only had a cursory glance at your sources, not a sufficient study yet. 2) Is there a README or a short howto available ? And one for Mattias et all - this is SUCH a widely useable thing and since Felipe made it nicely multiplatform for ALL the currently active Lazarus platforms, shouldn't it go into Lazarus itself ? Ciao A.J. On Monday 16 January 2006 23:49, Felipe Monteiro de Carvalho wrote: The test program works on Linux (tested on Gnome, KDE, IceWM) and Windows (tested on XP). -- Felipe Monteiro de Carvalho -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] platform independent application launching
Hi all, I have an interesting dillema, I need to write a platform independant program which needs to be able to launch other programs. It must NOT wait for them to exit but merely launch them, specifically we`re talking about launching things like a web-browser or pdf viewer. The choice of the actual commands I will handle, the bigger problem is how to do it. Obviously the method only works on unixes (I definitely need win32 support) and as far as I know, tprocess cannnot LAUNCH a program and then forget about it and pass control back to my own app. If there is no single way to do it, I am happy to use IFDEFS, but of course I still need to know what I can do on windows. Any ideas ? -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] New lazarus based program
For those interested. http://silentcoder.co.za/tiki/tiki-view_blog_post.php?blogId=2postId=214 Sidenote: Since the blog went out, a friend of mine has already started on a Mac port, gotta love multiplatform compilers ! -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] Access violation ?
Hi all,
I am trying to build a native gzip uncompressor into a unit for use by
programs, no difficulty right, I just use paszlib and stole (with some
adaptation) the code from the minigzip example to come up with this function
(I chose a function in order to feed the output back to a graphical programs)
Trouble is, as soon as I try to use it, the program gets an access violation.
In the terminal I can pick up the following output:
TApplication.HandleException Access violation
Stack trace:
$4048B4C0 LAZARUSPACKAGEINTF_finalize, line 49 of gtkglobals.ppiphtml.pas
--
It doesn't actually matter WHICH packages are required, if I comment out the
functions in the same unit that use iphtml and remove the requirements it
still happens. If I remove the file from the package and try compiling a test
program with the unit just in the same directory, it still fails.
All that changes is the unit in WHICH it dies (seems to fall back to plain
gtkglobals if iphtml is not present).
I assume the mostly plain pascal code this was based on is somehow not
lazarus/package compatible but I am at a loss to explain why.
Any advice greatly appreciated, the function follows below
-
function gz_uncompress(inputfile,outputfile :string):Integer;
Var len:Int;
Written:Uint;
ioerr :Integer;
err:int;
Busy :Boolean;
Status :Integer;
infile :gzfile;
outfile:file;
Begin
infile := gzopen(inputfile,'r');
Assign(OutFile,OutputFile);
Rewrite(OutFile,1);
Status := 0;
Busy := True;
writeln('test');
While Busy do begin
len := gzread(infile, @buf, BUFLEN);
if (len 0) then
Begin
Busy := False;
Status := GzReadError;
end;
if len = 0 then
Busy := False;
if busy then
Begin
{$I-}
blockwrite (outfile,buf,len,written);
{$I+}
if (written len) then
Begin
Busy := False;
Status := WriteError;
end;
end;
end; {While}
If Status = 0 then
begin
{$I-}
close(OutFile);
{$I+}
ioerr := IOResult;
if ioerr 0 then
Status := IOResult;
if Status = 0 then
if gzclose(infile) 0 then
Status := GzCloseError;
end;
gz_uncompress := Status;
end;
---
--
A.J. Venter
Chief Software Architext
OpenLab International
www.getopenlab.com
www.silentcoder.co.za
+27 82 726 5103
_
To unsubscribe: mail [EMAIL PROTECTED] with
unsubscribe as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Access violation ?
Another note on this, I stripped the function into a unit all of it's own (no
package), then I called this function from a lazarus program and got an
access violation as soon as I called gz_uncompress.
I then created a very simple little plain pascal program calling the same unit
with paramstr(1)'s - worked perfectly with no errors.
E.G. Whatever is wrong is indeed something in the pascal code clashing with
something in the Object Pascal of lazarus - any ideas ?
A.J.
On Saturday 28 January 2006 15:01, A.J. Venter wrote:
Hi all,
I am trying to build a native gzip uncompressor into a unit for use by
programs, no difficulty right, I just use paszlib and stole (with some
adaptation) the code from the minigzip example to come up with this
function (I chose a function in order to feed the output back to a
graphical programs)
Trouble is, as soon as I try to use it, the program gets an access
violation. In the terminal I can pick up the following output:
TApplication.HandleException Access violation
Stack trace:
$4048B4C0 LAZARUSPACKAGEINTF_finalize, line 49 of
gtkglobals.ppiphtml.pas --
It doesn't actually matter WHICH packages are required, if I comment out
the functions in the same unit that use iphtml and remove the requirements
it still happens. If I remove the file from the package and try compiling a
test program with the unit just in the same directory, it still fails.
All that changes is the unit in WHICH it dies (seems to fall back to plain
gtkglobals if iphtml is not present).
I assume the mostly plain pascal code this was based on is somehow not
lazarus/package compatible but I am at a loss to explain why.
Any advice greatly appreciated, the function follows below
-
function gz_uncompress(inputfile,outputfile :string):Integer;
Var len:Int;
Written:Uint;
ioerr :Integer;
err:int;
Busy :Boolean;
Status :Integer;
infile :gzfile;
outfile:file;
Begin
infile := gzopen(inputfile,'r');
Assign(OutFile,OutputFile);
Rewrite(OutFile,1);
Status := 0;
Busy := True;
writeln('test');
While Busy do begin
len := gzread(infile, @buf, BUFLEN);
if (len 0) then
Begin
Busy := False;
Status := GzReadError;
end;
if len = 0 then
Busy := False;
if busy then
Begin
{$I-}
blockwrite (outfile,buf,len,written);
{$I+}
if (written len) then
Begin
Busy := False;
Status := WriteError;
end;
end;
end; {While}
If Status = 0 then
begin
{$I-}
close(OutFile);
{$I+}
ioerr := IOResult;
if ioerr 0 then
Status := IOResult;
if Status = 0 then
if gzclose(infile) 0 then
Status := GzCloseError;
end;
gz_uncompress := Status;
end;
---
--
A.J. Venter
Chief Software Architext
OpenLab International
www.getopenlab.com
www.silentcoder.co.za
+27 82 726 5103
_
To unsubscribe: mail [EMAIL PROTECTED] with
unsubscribe as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] PasZlib still
I am still unable to get paszlib to work.
I have since gone to some pretty big lengths toward my goal, but I am afraid I
am stuck.
I think the ideal way would be to be able to read the file into a TMemstream,
decompress it into another Tmemstream and then write that out, but I must
admit I have no idea how to go about achieving that.
So considering that, for the moment I decided that I may have made a mistake
in my code conversions, so I fell back on minigz's own code virtually
unaltered, but wrapped inside a component.
All I get is that when I call the extract method - the program throws:
TApplication.HandleException Access violation
Stack trace:
$40489AE0 GZ_EXTRACT_finalize_implicit, line 54 of gtkglobals.pp
What does an error in finalize MEAN ?
I did an strace on the program, and it certainly reads and writes a LOT of
data once you trigger the method, so SOMETHING is happening before the crash.
I attach my full source code this time, if somebody would like to see the
trace I will mail that offlist as it is likely to be huge.
I urgently need to get this working to do transparent decompression for
OpenBook on windows (calling gunzip only works where gunzip works well).
Ciao
A.J.
--
A.J. Venter
Chief Software Architext
OpenLab International
www.getopenlab.com
www.silentcoder.co.za
+27 82 726 5103
{ Copyright (C) 2006 A.J. Venter
This library is free software; you can redistribute it and/or modify it
under the terms of the GNU Library General Public License as published by
the Free Software Foundation; either version 2 of the License, or (at your
option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU Library General Public License
for more details.
You should have received a copy of the GNU Library General Public License
along with this library; if not, write to the Free Software Foundation,
Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
}
unit gz_extract;
{Note - you must add turbopoweripro to your project requirements for
this to work}
{$mode objfpc}{$H+}
interface
uses
Classes, SysUtils,Process,zutil,gzIo, Forms, Controls, Graphics, Dialogs;
Type
TGZ_Extractor = class(TComponent)
private
FInputFile : String;
FOutPutFile : String;
public
destructor destroy; override;
Procedure Extract;
published
property InputFile : String write FInputFile;
property OutPutFile : String write FOutPutFile;
constructor create(AOwner: TComponent); override;
end;
procedure register;
Implementation
{TGZ_Extractor}
const
BUFLEN = 16384 ;
GZ_SUFFIX= '.gz' ;
{$DEFINE MAXSEF_64K}
var
buf : packed array [0..BUFLEN-1] of byte; { Global uses BSS instead of stack }
prog : string;
procedure error (msg:string);
begin
showMessage(msg);
end;
procedure Register;
begin
RegisterComponents('OLPack',[TGZ_Extractor]);
end;
procedure gz_uncompress (infile:gzFile; var outfile:file);
var
len : int;
written : uInt;
ioerr : integer;
err : int;
begin
while true do begin
len := gzread (infile, @buf, BUFLEN);
if (len 0)
then error (gzerror (infile, err));
if (len = 0)
then break;
{$I-}
blockwrite (outfile, buf, len, written);
{$I+}
if (written len)
then error ('write error');
end; {WHILE}
{$I-}
close (outfile);
{$I+}
ioerr := IOResult;
if (ioerr 0) then begin
writeln ('close error: ',ioerr);
halt(1);
end;
if (gzclose (infile) 0{Z_OK})
then error ('gzclose error');
end;
procedure file_uncompress (filename,outname:String);
var
inname : string;
infile : gzFile;
outfile : file;
ioerr : integer;
len : integer;
begin
len := Length(filename);
inname := filename;
infile := gzopen (inname, 'r');
if (infile = NIL) then begin
writeln (prog,': can''t gzopen ',inname);
halt(1);
end;
Assign (outfile, outname);
{$I-}
Rewrite (outfile,1);
{$I+}
ioerr := IOResult;
if (ioerr 0) then begin
writeln ('open error: ',ioerr);
halt(1);
end;
gz_uncompress (infile, outfile);
end;
Procedure TGZ_Extractor.Extract;
Begin
File_UnCompress(FInputFile,FOutPutFile);
end;
destructor TGZ_Extractor.destroy;
Begin
end;
constructor TGZ_Extractor.create(AOwner: TComponent);
begin
inherited create(AOwner);
end;
end.
Re: [lazarus] PasZlib still
Thanks for this, I am now evaluating all the suggestions. Just a question Graeme, does dzlib deal with winzip or gzip format ? The files I need access to are all gzipped. Ciao A.J. On Monday 30 January 2006 18:51, Graeme Geldenhuys wrote: Hi AJ, Not sure if paszlib is the same as dzlib? I think the dzlib just uses the same zlib interface as what is included in Dephi. Anyway, I attached a zip file containing 5 units as I use them. tiCompress: Defines a abstract class and a Factory for registering different compression algorithms with my app. tiCompressZlib: Implements the interface with zlib compression tiCompressNone: Implements the interface with no compression. Handy for unit testing and debugging. tiConstans: contains some of the constans I use dzlib: zlib compression library Hope that helps a fellow ZA citizen. Regards, - Graeme - On 1/29/06, A.J. Venter [EMAIL PROTECTED] wrote: I am still unable to get paszlib to work. I have since gone to some pretty big lengths toward my goal, but I am afraid I am stuck. I think the ideal way would be to be able to read the file into a TMemstream, decompress it into another Tmemstream and then write that out, but I must admit I have no idea how to go about achieving that. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] lazarus components
Andrea, while I do not have a tutorial on it (though I was planning to write one when I have time actually) the process is very similar to Delphi, you can browse my SVN tree http://www.silentcoder.co.za/viewsvn For several component packs with a number of components, being fairly small components - they are easy to base others from. A.J. On Wednesday 01 February 2006 17:59, Andrea Mauri wrote: In internet I found a lot of tutorial on how to write components in delphi. I found also a tutorial that explain how to convert delphi component for lazarus. I would like to try to write components with lazarus, there is a wiki page or a tutorial to do that? Can I follow the tutorial dedicated to delphi also for lazarus? Which are the main differences? Thanks, Andrea _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Window Icon again - Proposal
The linux/gtk forms have icons. What is missing, are icons showing in the file managers. This is normally done by copying an image of the executable name to a directory like /usr/shar/pixmaps/. Aye which will usually set the window icon anyway so normally it doesn't bother people - except my program is meant to run uninstalled from a DVD so that doesn't entirely work anymore. But then... if the linux forms have icons... where/how do you set them ? Or is the ONLY way currently to install an icon to $ICONDIR. I can't find any place to do so under project options (latest lazarus from svn). The trickiest bit is likely to be reading an X11 Pixmap variable out of a TPixmap. See thread Application Icon under Win32 (modified patchs) from 9 Feb 2005 Thanks for that. A.J. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Window Icon again - Proposal
On Wednesday 01 February 2006 19:28, A.J. Venter wrote: The linux/gtk forms have icons. What is missing, are icons showing in the file managers. This is normally done by copying an image of the executable name to a directory like /usr/shar/pixmaps/. Aye which will usually set the window icon anyway so normally it doesn't bother people - except my program is meant to run uninstalled from a DVD so that doesn't entirely work anymore. But then... if the linux forms have icons... where/how do you set them ? Or is the ONLY way currently to install an icon to $ICONDIR. I can't find any place to do so under project options (latest lazarus from svn). Scratch that, found it in the form properties, been using lazarus too long, I'm too used to what isn't there... A.J. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Dumb question - updating a progressbar
On Wednesday 01 February 2006 16:40, A.J. Venter wrote: Hi all, Why is it that I cannot seem to get my progressbar to update ¨live¨ I am trying to show the progress of decompressing files when the user selects a multifile book in OpenBook. Don't worry about it, I decided to do the right thing(tm) and I used a thread. All good now :) A.J. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] GetFileList
I have been trying all night through a thousand variations on the theme of findfirst and findnext to try and come up with what should be a simple algorythm, start at a specified toplevel directory, and add every file below it with full path to a stringlist. Of course it does need to recurse subdirectories. Constantly failing, I tried googling but all I can find is delphi components which while they ought to be simple enough all throw access violations at runtime. Any advice ? Or perhaps somebody feels bored... TIA A.J. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] GetFileList
', 'The mask '+FilterAsText+' is not a valid expression.'); end; end; end; var Directory: String; begin Result:=false; RemoveFilterRegExpr:=nil; KeepFilterRegExpr:=nil; List:=nil; try // get directory Directory:=DirName; //if (Macrosnil) and (not Macros.SubstituteStr(Directory)) then exit; Directory:=AppendPathDelim(Directory); // setup filters SetupFilter(RemoveFilterRegExpr,false, Filter); if length(exclude) 0 then SetupFilter(KeepFilterRegExpr,false, exclude); // search files List:=TStringList.Create; if not SearchInDirectory(Directory,0) then exit; Result:=true; finally RemoveFilterRegExpr.Free; //KeepFilterRegExpr.Free; if not Result then begin List.Free; List:=nil; end; end; end; = Any idea what else I need to change to make this general-use version of the function work ? Ciao A.J. -- A.J. Venter Chief Software Architect, OpenLab International www.getopenlab.com +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] GetFileList
Macros: To search for '$(ProjectDir)/*.lfm' SimpleSyntaxToRegExpr: To use masks like '*.txt' For example SimpleSyntaxToRegExpr(FilterAsText) converts '*.txt' into the regular expression '^.*\.txt$'. If you don't want to use regular expressions you can use FileInFilenameMasks See the comment at the function body. Thanks the function is working now with regex support, including simpleformat regex´s. Now I have some new questions though. According to the tutorial on lazarus-ccr a thread must never update the form, does this also apply to things like messagedialog and showmessage ? Because if so, then this function is not currently threadsafe and I need to hack it a little more (I at least intend to use it in threads). For the record, the final function is going into olpack, possibly with a wrapper component later in case anybody else wants it. A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] GetFileList
Now I have some new questions though. According to the tutorial on lazarus-ccr a thread must never update the form, does this also apply to things like messagedialog and showmessage ? Yes. The complete painting and mouse/keyboard stuff belongs to one thread. Thanks, so implemented, I removed the calls for those, and simply made sure the function will return false, to keep it threadsafe, error HANDLING will have to be the task of the calling programs, at least error CATCHING is there. Ciao A.J. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Re: TrayIcon preview
There is no executable xev, and no package named xev available on the repositories of my Mandriva 2006. Xev is part of the Xorg project but not all distros include it in the default packages, look for something called xorg-utils or such which isn't installed, includes all the stuff like xcalc and xev and xlsfonts which normal users almost never use anymore. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Re: TrayIcon preview
On the other hand, on IceWM instead of my icon, I see some random colors painted and it doesn't respond to mouse events. This may or may not be related but do an ldd on your icewm executable, icewm by default has ONLY support for xpm image formats (though there is a configure option to enable png as well). If you don´t see libpng in the libraries list, try doing a test program using only xpm images to your trayicon, that may help, barring that it may also imply that icewm´s image handling isn´t compatible to the LCL ways. Of course this is just a guess based on your description, I could be wrong. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] Am I being too ambitious?
Or am I just doing it the wrong way around ?
A very large part of my dayjob involves writing frontends to commandline
programs under linux, so of course I figured I could make my life a lot
simpler if I had an abstract class I could use to run processes in threads
and retrieve the output live instead of having to reimplement essentially the
same code over and over.
So I began with my class, derived it from thread, built the TProcess into
the execute method and added the right synchronizations to have it update
the output as it arrives back from tprocess.
It compiles just fine.
But the moment I try to use it in a simple test program, it throws an
async-reply exception, something I would normally only expect if threaded
code tried to update the display directly.
Any ideas where my thinking went wrong ?
The unit is attached. The call in my test program is as follows (yes I know
I am not utilizing the power of my own class, I just wanted to see if it
WORKED).
procedure TForm1.Button1Click(Sender: TObject);
begin
MyProcess := TThreadedProcess.Create(True);
MyProcess.Command := Edit1.Text;
Button1.Enabled := False;
MyProcess.Resume;
While MyProcess.Running do
sleep(100);
Button1.Enabled := True;
Memo1.Lines.Assign(MyProcess.OutPut);
MyProcess.Free;
end;
Ciao
A.J.
--
A.J. Venter
Chief Software Architect, OpenLab International
www.getopenlab.com
+27 82 726 5103
unit threadedprocess;
{$mode objfpc}{$H+}
interface
uses
Classes, SysUtils,process;
Type
TThreadedProcess = class(TThread)
private
FRealOutPut : TStringList;
FCommand : String;
FOutPut : TStringList;
procedure UpdateOutPut;
procedure Started;
procedure AllDone;
FRunning : Boolean;
protected
procedure execute; override;
public
destructor destroy; override;
property Command : String write FCommand;
property Output : TStringList read FOutPut;
Property Running : Boolean read FRunning;
Constructor Create(isSuspended : boolean);
end;
//procedure register;
implementation
//procedure Register;
//begin
// RegisterComponents('OLPack',[TThreadedProcess]);
//end;
Procedure TThreadedProcess.Started;
Begin
FRunning := True;
end;
procedure TThreadedProcess.UpdateOutPut;
Var I : Integer;
Begin
For I := FRealOutput.Count to FoutPut.Count do
FoutPut.Add(FRealOutPut[I]);
end;
Procedure TThreadedProcess.AllDone;
Begin
FRunning := False;
end;
Procedure TThreadedProcess.Execute;
const
READ_BYTES = 2048;
Var
Process :Tprocess;
MemStream : TMemoryStream;
n: LongInt;
BytesRead: LongInt;
begin
FRealOutPut := TStringList.Create;
Process := TProcess.create(nil);;
Process.CommandLine := Fcommand;
{Actually run the thing and catch the output}
MemStream := TMemoryStream.Create;
BytesRead := 0;
Process.Options := [poUsePipes];
Process.Execute;
synchronize(@Started);
while Process.Running do
begin
// make sure we have room
MemStream.SetSize(BytesRead + READ_BYTES);
// try reading it
n := Process.Output.Read((MemStream.Memory + BytesRead)^, READ_BYTES);
if n 0
then begin
Inc(BytesRead, n);
FRealOutPut.Clear;
FrealOutPut.LoadFromStream(MemStream);
synchronize(@updateOutPut);
end
else begin
// no data, wait 100 ms
Sleep(100);
end;
end;
// read last part
repeat
// make sure we have room
MemStream.SetSize(BytesRead + READ_BYTES);
// try reading it
n := Process.Output.Read((MemStream.Memory + BytesRead)^, READ_BYTES);
if n 0
then begin
Inc(BytesRead, n);
end;
until n = 0;
MemStream.SetSize(BytesRead);
FRealOutPut.LoadFromStream(MemStream);
synchronize(@UpdateOutPut);
MemStream.Free;
Process.Free;
synchronize(@AllDone);
end;
constructor TThreadedProcess.Create(isSuspended : boolean);
begin
FRunning := False;
FOutPut := TStringList.Create;
FreeOnTerminate := False;
inherited Create(isSuspended);
end;
destructor TThreadedProcess.Destroy;
Begin
FOutPut.Free;
FRealOutPut.Free
end;
end.
Re: [lazarus] Am I being too ambitious?
The synchronize calls shouldn't be placed inside your TThreadedProcess code... they should be used in the main application thread to provide mutual exclusion for the shared visual components. It's the other way round, at least in delphi: you call synchronize from a thread when you want the method executed in the context of the main thread. That is correct, I know this much at least :) after all how can you synchronize from the main thread it doesn´t know what the other thread is doing. Anyway I think the problem is right at the design level, the right way I think would be to create a tthread descendent in a unit, with an extra property FCommandLine which can be overwritten at instantiation time. Pretty much what I had, but it should NOT be instantiated in the program. Instead the same unit should contain a tcomponent descendent which instantiates the tthread descendent, takes its values from poperties to pass on to the ththread descendent (which itself is a property) and THEN the thread can synchronize with a property of the tcomponent for output ? Am I getting closer you think ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Am I being too ambitious?
Mmmh, looking more closely to your code (I didn't before, I was just replying to the assertion that you call synchronize from the main thread) I don't think you should use synchronize at all (maybe a mutex to coordinate the access to FRunning, but even that may be overkill for a boolean): since you don't access the TStringList unless the process has ended there's no problem of simultaneous access from the main thread and the thread (and even if you wanted to do that I'd prefer a mutex over synchronize). In fact (not knowing the internals of the synchronize implementation, so I may be wrong) I cannot see how your thread can work since you're in a busy loop waiting for the thread to finish, so there's no way the main thread can schedule your synchronize call. Okay, let me explain exactly what I wish to achieve. I want to be able to stick a tmemo on a form, call tthreadedprocess with a command that takes a long time to run and has a lot of output (lets say ¨find /¨ for an example) - and have the output scrolling into the tmemo LIVE as it arrives, hence the call to synchronize every time the tmemstream reaches 2048 bytes. The main thread is not meant to be ANYWHERE in this class, this class is meant to create a thread, and then FEED the output of the tprocess to the main thread (in the calling appplication) LIVE. So while it´s true that I don´t use FOutPut in this class -that is because I want to be able to use it in the calling application. Having explained this, does it clarify why the code looks as it does ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] Really quick question
Under linux, when passing a filename with spaces/special characters to a shell command, I would simply make sure it is wrapped in codes eg. commandline := Comand+' '+FileName+''; How (if at all) would I do the same under win32 ? If I don't need to do anything - will the above still work or do I need to ifdef it ? A.J. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Really quick question
On Tuesday 07 February 2006 21:56, Michael Van Canneyt wrote: On Tue, 7 Feb 2006, A.J. Venter wrote: Under linux, when passing a filename with spaces/special characters to a shell command, I would simply make sure it is wrapped in codes eg. commandline := Comand+' '+FileName+''; How (if at all) would I do the same under win32 ? The same, if you pass it to cmd. If I don't need to do anything - will the above still work or do I need to ifdef it ? Not if you pass it to cmd. Well I am passing to a tprocess, and I did notice it opening cmd windows, so worst case scenario it will cause an issue if we're talking windows 98 or earlier ? Which I am not very concerned about (and look at me, I LEARNED to code on linux and I'm lost whenever I try to port anything :p ) -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Really quick question
I can second this, don't forget on WIn98 the shell is named command, on 200 and XP it is cmd and if you start it with command /? it'll tell you what options are accepted. Some of them are really helpful e.g. for debugging. Okay now two related questions 1: Can I somehow tell a tprocess under windows which is launching a non-GUI backend command to NOT open a cmd/command window ? Especially since the command being run is non-interactive, the window closes as soon as it's done but it just looks ugly. 2: So how do I specify advanced vfat file/directorynames in OP code ? E.G. to let my installer create an installdir under Program Files - how would I mask out the space so OP knows its ONE directory name ? A.J. -- A.J. Venter Chief Software Architext OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Who Discussed Linker Slowness?
On Monday 13 February 2006 17:55, Lord Satan wrote: Is anybody using a way around or are you all day waiting? The way around is running under Linux where linking is much faster. As often said the speed problems are caused by the GNU linker and AFAIK there is no way to solve this other than writing a new linker. I was the one who first pointed this out, and I can give you a few insights, a large part of the speed issue is with the way GNU ld seems to treats (lazarus?) sources and their libraries. Linking speed become a direct factor of how much ram you have. My little 128mb laptop starts swapping like a maniac while linking and can take 30 minutes on anything with more than a few basic widgets, same code on my desktop (2GB of ram) takes about 3 seconds. What I did find is that the best way to up the linker speed when you don´t have a lot of ram is to write a wrapper script in /usr/bin (I called mine nicefpc) which will launch ppc386 with a much lower nice value (I used -18) then give it the suid bit so it can actually be launched and pass ppc386 a $* (so it gets all the parameters) and then tell lazarus to use nicefpc as it´s compile command. This ONLY works under linux but with the nice value way below everything else ppc386 and it´s spawns (like ld) all get maximum priority for memory and CPU access. Just don´t try to read mail while compiling :) Of course this works only under linux, dunno if there is a way to do something similiar on win32. A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: RES: [lazarus] Who Discussed Linker Slowness?
On Tuesday 14 February 2006 04:56, Henrique de Paula Faria wrote: Vincent, can you give some tip on how to configure the publish project to strip and upx the executable ? Henrique. You can strip and upx from the project options as well, (using the postbuild command and the compiler settings in conjunction). But it is NOT an answer ! Strip - rips out debugging code - NOT static linking ! UPX compresses the executable - in a NOT AS COMPATIBLE/SAFE AS THEY SAY way. I know, I used to use UPX, I have seen how badly it goes wrong when it does - apps that simply won´t run after being compiled... These are work-arounds to the REAL issue which is that by default lazarus links statically. Smartlinking takes longer but it is the right way(tm). You should STILL strip any program you are distributing but THAT is just standard software distribution policy - debugging symbols are for building, not for shipping. In other words, it´s the size of the executable AFTER stripping we should be concerned about. An outside executable compresor may or may not be usefull to you but depending on one is a bad thing(tm) and CERTAINLY not as a work-around for static linking. The fact that lazarus is meant to be multiplatform and that the number of support platforms are increasing makes this doubly important. I know that ppc386 can take a commandline option to tell it whether or not to smartlink, though I don´t know if this can just disable compiler directives for smartlinking or if it can replace them. Correctly used however, that is the answer, publish project should smartlink the code, if smartlinking is to slow then the normal run-build-run version should static link your program. The LCL needs to be smart linkable, but I don´t think it always needs to be smartlinked. You should be able to static link during development. Ciao A.J. PS. Dynamic linking is not even an option to me - the last thing I want is for my lazarus programs to start behaving like kylix programs and insist you load about 12 other library packages which may or may not have been built for your glibc depending on whether borland was in a good mood that day. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] forms maximizing
On Thursday 16 February 2006 11:42, Pēteris Bērzs-Bērziņš wrote: Hi maybe some one can help me with advice. I wana make my main form like Delphi, Lazarus. On top on the screen, but on maximizing it goes under start bar witch is on top. See attached file. Strange, I have programs which does this (see olpack´s tscreensize component on silentcoder.co.za) and this has never happened with them. I would guess it ´s the window manager that is responsible for it, I know that under KDE it won´t go behind or over the taskbar but will just stop dead against it. Unless you are trying to go for fullscreen mode which I am pretty sure will require a call to gtk_wm_hint stopping against the taskbar has always been how I defined maximized. Mind you it may be possible to use a wm_hint call in your oncreate event to tell the window manager to maximise it for you ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] lightweight http
On Wednesday 22 February 2006 03:33, Juan Martinez wrote: Ales Katona [EMAIL PROTECTED] escribió: Juan Martinez wrote: Hi all I'm looking for a very lightweight http component, I don't need browser capabilities, even not html procesing, just sending, receiving and parsing http headers are needed. Anyone knows about such a component? I've seen httpclient unit in the fcl, but I don't know how to use it. Have you considered CurlPas - it's not exactly lightweight but it can be used in a lightweight manner and it's available and it works :) Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] lightweight http
On Wednesday 22 February 2006 18:21, A.J. Venter wrote: On Wednesday 22 February 2006 03:33, Juan Martinez wrote: Ales Katona [EMAIL PROTECTED] escribió: Juan Martinez wrote: Hi all I'm looking for a very lightweight http component, I don't need browser capabilities, even not html procesing, just sending, receiving and parsing http headers are needed. Anyone knows about such a component? I've seen httpclient unit in the fcl, but I don't know how to use it. Have you considered CurlPas - it's not exactly lightweight but it can be used in a lightweight manner and it's available and it works :) Sorry, I pasted the wrong link - this one is more appropriate to your needs: http://curlpas.sourceforge.net/curlpas/index.html -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] TListView.CustomSort
Hi, It seems that the lazarus tlistview's do not implement the customsort method - not even as a virtual, the result is that all the tutorials I could find on how to sort a tlistview by a columnclick do not work with lazarus since they all rely on tcustomsort (and I presume OnCompare). Is this a TODO ? Or is there a different way to do this in Lazarus ? Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListView.CustomSort
On Wednesday 22 February 2006 21:14, Marc Weustink wrote: A.J. Venter wrote: Hi, It seems that the lazarus tlistview's do not implement the customsort method - not even as a virtual, the result is that all the tutorials I could find on how to sort a tlistview by a columnclick do not work with lazarus since they all rely on tcustomsort (and I presume OnCompare). Is this a TODO ? Or is there a different way to do this in Lazarus ? This is a todo. I tried to implement sorting, read about it on how to do it on gtk and found that it differs much from win32. So I tried to find a nice way were both would fit in a VCL compatible way. And then I had other business or so. Okay then, so basically right now the only way would be to clear, create a sorted list and refill ? I suppose that can work as a temporary measure since sorting in my app is actually quite critical, unless somebody has a better idea. (This is for OpenBook - to allow people to sort search results by a field of their choice - obviously it would hugely improve the overall usability). Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListView.CustomSort
On Wednesday 22 February 2006 21:30, Jesus Reyes wrote: Have you considered using a stringgrid? or some derived from it like TSemaphorGrid (lazarus/examples/grid_semaphor) I looked at them right in the beginning, but I don't think they will be able to do what I am after look and feel wise (this: http://www.silentcoder.co.za/tiki/show_image.php?id=568) Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListView.CustomSort
I don't see anything in the screenshot that TStringGrid can't do :D Jesus Reyes A. Okay, I must be stupid then :) I always rather thought of TStringGrid as a tool for letting users INPUT data, while TListView is a tool for PRESENTING data that must not be editable - I think users might be confused by the spreadsheet like cells if they cannot edit them. Though to be honest the biggest single reason is probably that I have hardly ever used the stringgrid components so I don't actually know them very well and the idea of rewriting the population code in it scares me a little :) Mind you if you feel like contributing to a free software project by doing the conversion for me - well I never say no to useful contributions :p Okay, enough contrib-whoreing, you may just be right though - and I should probably just go ahead and figure it out. Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListView.CustomSort
On Wednesday 22 February 2006 22:29, Micha Nelissen wrote: On Wed, 22 Feb 2006 22:22:04 +0200 A.J. Venter [EMAIL PROTECTED] wrote: Mind you if you feel like contributing to a free software project by doing the conversion for me - well I never say no to useful contributions :p Okay, enough contrib-whoreing, you may just be right though - and I should probably just go ahead and figure it out. I do have to remark that a TListView is more native than a StringGrid. At least, on windows xp a listview looks much nicer. A stringgrid is always completely LCL-drawn. Having now taken a much more indepth look at tstringgrid, I think it could work - but only at the cost of hugely complexifying my program - unless somebody is really eager to do it for me, I don't think it's worth it. I would rather help with implementing sorting for tlistview so I can have it for my next version. My feel btw: to keep code platform neutral, don't use the native sorting methods, but do a sort based on pure data manipulation. I could even do that in the app (as opposed to a clear and refill method) although that is harder to do it is certainly not impossible and a quicksort is practically the first algorithm any of us learned after all :) Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListView.CustomSort
On Wednesday 22 February 2006 22:55, Jesus Reyes wrote: Okay, enough contrib-whoreing, you may just be right though - and I should probably just go ahead and figure it out. No problem, send me some code to convert and I will try, if you doesn't like it (see below), you can always try to do the sorting as you alredy planned and continue with Tlistview :) Well in that case, you can check out the code from my subversion server: svn co svn://silentcoder.co.za/tools/openbook You will also need to checkout and install my component packs: svn co svn://silentcoder.co.za/lazarus In order to actually compile it as it uses both olpack and gzpack quite extensively (though not for anything that affects what you are looking at). The code tree includes samples of all the needed support files including an index (though not the nearly 4gb of data it is an index FOR of course - but you shouldn't need that for testing - just verifying that the thing opens with the right parameters ought will do it). For the rest it should be fairly simple to make sense of but shout if you want to check anything, I am also on IM: Yahoo: silentcoder Jabber: [EMAIL PROTECTED] MSN: [EMAIL PROTECTED] Ciao and thanks A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListView.CustomSort
Ciao and thanks A.J. Sorry about that, I intended to mail Jesus directly and forgot that lazarus uses reply-to setting. A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Enhanced versions of TStringGrid
Being said that, I think your final question remains open to others wanting to participate on ideas of enhancement, deriving or whatever :) I think, Michael already gave a nice solution: Let's implement a plugin system for grids. See SynEdit for an example. For the hook lists see TControl.AddHandler. It should be possible to write plugins for cell drawing, cell editors, sorting, navigation and data connections. I like this idea as well (sorry if this is a me-too post :p), the concept of having a single top-level grid component, into which you can then plug those features needed by your application makes far more sense to me than having a dozen grid components each with one extra feature. I would even suggest that TDBGrid should be dropped and it's functionality turned into a plugin for this top-level grid component. Unless of course people feel that would break too much compatibility/familiarity for former delphi users. Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] Why is this throwing an exception ?
So there I was, thinking today, what can I do to once more save future time...
and I realised that something I would really like is to have a component that
can wrap around ipro and implement a simple file-based html viewer with image
support, jpeg loading and such built-in.
Easy right, after all just that exists in the example shipped for it, so all I
have to do is plagiarise the code a bit and create a component out of it.
So I created my component and added the bits I wanted, mostly just a property
for specifying an html file and a method for showing it.
And it works, even jpg's load fine.
But immediately after loading it throws an exception, ignore the exception and
it throws it again, and again and again (no idea is given of WHAT exception).
The frequency and timing does however suggest that the exception is thrown
during the onpaint event (though I never TOUCHED that :p )
Anyway, here is the code:
unit simple_html;
{A simple file based html viewer component based on ippro with image viewing
added}
{$mode objfpc}{$H+}
interface
uses
Classes, SysUtils, LResources, Forms, Controls, Graphics,IPHtml,lazjpeg;
type
TSimpleIpHtml = class(TIpHtml)
public
property OnGetImageX;
end;
TSimpleHTML = class(TIpHtmlPanel)
private
FHtmlFile : String;
FNewHTML: TSimpleIpHtml;
public
{ public declarations }
procedure HTMLGetImageX(Sender: TIpHtmlNode; const URL: string; var
Picture: TPicture);
procedure ShowFile; {Loads a designtime set html file}
destructor destroy; override;
published
{ published properties }
constructor create(AOwner: TComponent); override;
property HTMLFile : String Write FHtmlFile;
end;
procedure register;
implementation
{TSImpleHTML}
procedure Register;
begin
RegisterComponents('OLPack',[TSimpleHTML]);
end;
procedure TSimpleHTML.HTMLGetImageX(Sender: TIpHtmlNode; const URL: string;
var Picture: TPicture);
var
PicCreated: boolean;
begin
try
if FileExists(URL) then begin
PicCreated := False;
if Picture=nil then begin
Picture:=TPicture.Create;
PicCreated := True;
end;
Picture.LoadFromFile(URL);
end;
except
if PicCreated then
Picture.Free;
Picture := nil;
end;
end;
procedure TSimpleHTML.ShowFile;
var
fs: TFileStream;
begin
if FileExists(FHtmlFile) then
try
fs:=TFileStream.Create(FHtmlFile,fmOpenRead);
try
FNewHTML.OnGetImageX:[EMAIL PROTECTED];
FNewHTML.LoadFromStream(fs);
finally
fs.Free;
end;
SetHtml(FNewHTML);
except
on E: Exception do begin
Writeln('Unable to open HTML file',FHtmlFile);
end;
end;
end;
destructor TSimpleHTML.destroy;
Begin
inherited;
end;
constructor TSimpleHTML.create(AOwner: TComponent);
begin
FNewHTML:=TSimpleIpHtml.Create;
inherited;
end;
end.
---
Here is the relevant parts from a simple demo program I did to test it:
TForm1 = class(TForm)
Button1: TButton;
Button2: TButton;
Edit1: TEdit;
OpenDialog1: TOpenDialog;
SimpleHTML1: TSimpleHTML;
procedure Button1Click(Sender: TObject);
procedure Button2Click(Sender: TObject);
procedure TForm1.Button2Click(Sender: TObject);
begin
if OpenDialog1.Execute then
Edit1.Text := OpenDialog1.Filename;
end;
procedure TForm1.Button1Click(Sender: TObject);
begin
SimpleHTML1.HTMLFile := Edit1.Text;
SimpleHTML1.ShowFile;
end;
---
Anybody here familiar enough with ipro to know what I am doing wrong ?
Ciao
A.J.
--
A.J. Venter
Chief Software Architect
OpenLab International
www.getopenlab.com
www.silentcoder.co.za
+27 82 726 5103
_
To unsubscribe: mail [EMAIL PROTECTED] with
unsubscribe as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
Re: Why is this throwing an exception ?
Sorry, forgot to add, here is the stack trace as well: TApplication.HandleException Access violation Stack trace: $0820B073 TIPHTMLINTERNALPANEL__HTMLPANEL, line 15979 of iphtml.pas $08209DA0 TIPHTMLINTERNALPANEL__MOUSEMOVE, line 15523 of iphtml.pas $0810D156 TCONTROL__WMMOUSEMOVE, line 2882 of ./include/control.inc $0805CEF7 $0810A30E TCONTROL__WNDPROC, line 1542 of ./include/control.inc $08101D49 TWINCONTROL__WNDPROC, line 3347 of ./include/wincontrol.inc $08109224 TCONTROL__PERFORM, line 1014 of ./include/control.inc $08100C14 TWINCONTROL__ISCONTROLMOUSEMSG, line 2823 of ./include/wincontrol.inc $08101CD9 TWINCONTROL__WNDPROC, line 3328 of ./include/wincontrol.inc $0811E965 DELIVERMESSAGE, line 3457 of gtkproc.inc $08126CC2 DELIVERMOUSEMOVEMESSAGE, line 867 of gtkcallback.inc $08126D67 GTKMOTIONNOTIFY, line 920 of gtkcallback.inc $401E164D SYNREGEXPR_finalize_implicit, line 936 of gtkglobals.ppiphtml.pas $40210E5A SYNREGEXPR_finalize_implicit, line 936 of gtkglobals.ppiphtml.pas $4021016A SYNREGEXPR_finalize_implicit, line 936 of gtkglobals.ppiphtml.pas $4020E234 SYNREGEXPR_finalize_implicit, line 936 of gtkglobals.ppiphtml.pas $40243FD9 SYNREGEXPR_finalize_implicit, line 936 of gtkglobals.ppiphtml.pas -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: Why is this throwing an exception - ANSWER
It's a bug in ipro. I checked the code at the stack trace and found this: function TIpHtmlInternalPanel.HtmlPanel: TIpHtmlCustomPanel; begin Result := TIpHtmlPanel(Parent); while (Result.ClassType TIpHtmlPanel) do Result := TIpHtmlPanel(Result.Parent); end; The exception gets thrown in the while line. Now WHY they do it this way I do not know, but it means you get a real problem if you derive a component from it. I commented out the lines in the function like so: function TIpHtmlInternalPanel.HtmlPanel: TIpHtmlCustomPanel; begin //Result := TIpHtmlPanel(Parent); // while (Result.ClassType TIpHtmlPanel) do //Result := TIpHtmlPanel(Result.Parent); end; And now my test program works as beautifully as you could wish. From the context I get the idea this is somehow used if you wish to embed two htmlpanels inside one another, which doesn't affect me - but the method they are using is NOT good because it causes crashes if you derive anything (which changes the classtype of course). So do I report a bug ? Or how does this get fixed ? Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: Why is this throwing an exception - ANSWER
So do I report a bug ? Or how does this get fixed ? Please check above fix. If it works, I will change it. Well I have since redesigned my entire component so it is no longer descended from TIpHtmlPanel, and it is in a fully working state based on a very different approach. However I do think the fix should be applied because I am doubtlessly not the last person to try and descend the class :) Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] transparent forms
How does it work? Does it scan the form and build a region or so? Sounds like a horrible hack to me TBH. How do you use it ? Well there are a few ways one could achieve this, the worst hack being to have the code scan the form and copy whatever is behind it on screen onto the section you selected, that is a horrible way. Now on MacOSX and Linux (at least with newer versions of X) it should be possible to do TRUE translucency. On linux by tapping into the Xcomposite and Xrender extensions, not sure how you get to these features on Mac. On windows I don't know but different implementations for supported platforms is quite okay if the interface to the component is the same. I do think it would be a nice-to-have but right now I doubt there is a clean multiplatform approach so a per-platform implementation would be needed and it's not lazarus style to add components that work only in one place. I do reckon that a lazarus component which can tap into the features of xcomposite manager would be really neat though, not sure what it would take to create one, especially if you don't want to use a tprocess, I suspect our X-header conversions don't include the functions it uses so it may require converting some code from the c project by X.org. Of course it won't work on all versions of X only those that have the right support compiled in (which is alpha code right now) but it would mean that where available lazarus apps can be xgl enabled, users could have a slider right in the app for adjusting transparency... neat nay (of course I think this will be whole-form only though it could surely be mapped into a component as well to affect only that which is in the bounds of the component - personally I would like both - the component could have a slider letting the user adjust the transparency of the region it encapsulates (and otherwize look and work like say a tgroupbox) and another component which would be simpler and probably non-visual but would adjust full form transparency to whatever value you pass it in properties) Wow that was a lot of bracing :p Anyway, just some thoughts from me on the topic. Others ? Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: Why is this throwing an exception - ANSWER
Patch as per Mattias' suggestion attached, I tested it and verified that it fixes the problem allowing derivative components of ipHtmlPanel to be created and used. Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 Index: components/turbopower_ipro/iphtml.pas === --- components/turbopower_ipro/iphtml.pas (revision 8812) +++ components/turbopower_ipro/iphtml.pas (working copy) @@ -15976,7 +15976,7 @@ function TIpHtmlInternalPanel.HtmlPanel: TIpHtmlCustomPanel; begin Result := TIpHtmlPanel(Parent); - while (Result.ClassType TIpHtmlPanel) do + while not (Result is TIpHtmlPanel) do Result := TIpHtmlPanel(Result.Parent); end; Index: components/fpcunit/Makefile.fpc === --- components/fpcunit/Makefile.fpc (revision 8812) +++ components/fpcunit/Makefile.fpc (working copy) @@ -7,7 +7,7 @@ [compiler] unittargetdir=lib/$(CPU_TARGET)-$(OS_TARGET) unitdir=../../lcl/units/$(CPU_TARGET)-$(OS_TARGET)/ ../../lcl/units/$(CPU_TARGET)-$(OS_TARGET)/$(LCL_PLATFORM)/ ../../packager/units/$(CPU_TARGET)-$(OS_TARGET)/ -options=-dLCL -dLCLgtk -S2 -gl +options=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl [target] units=fpcunittestrunner.pas @@ -18,6 +18,16 @@ $(wildcard $(COMPILER_UNITTARGETDIR)/*$(RSTEXT)) \ $(wildcard $(COMPILER_UNITTARGETDIR)/*.compiled) \ $(wildcard *$(OEXT)) $(wildcard *$(PPUEXT)) $(wildcard *$(RSTEXT)) +[prerules] +# LCL Platform +ifndef LCL_PLATFORM +ifeq ($(OS_TARGET),win32) +LCL_PLATFORM=win32 +else +LCL_PLATFORM=gtk +endif +endif +export LCL_PLATFORM [rules] .PHONY: cleartarget all Index: components/fpcunit/Makefile === --- components/fpcunit/Makefile (revision 8812) +++ components/fpcunit/Makefile (working copy) @@ -1,5 +1,5 @@ # -# Don't edit, this file is generated by FPCMake Version 2.0.0 [2006/01/19] +# Don't edit, this file is generated by FPCMake Version 2.0.0 [2005/12/07] # default: all MAKEFILETARGETS=i386-linux i386-go32v2 i386-win32 i386-os2 i386-freebsd i386-beos i386-netbsd i386-solaris i386-qnx i386-netware i386-openbsd i386-wdosx i386-emx i386-watcom i386-netwlibc i386-wince m68k-linux m68k-freebsd m68k-netbsd m68k-amiga m68k-atari m68k-openbsd m68k-palmos powerpc-linux powerpc-netbsd powerpc-macos powerpc-darwin powerpc-morphos sparc-linux sparc-netbsd sparc-solaris x86_64-linux x86_64-freebsd x86_64-win64 arm-linux arm-wince powerpc64-linux @@ -231,6 +231,14 @@ PACKAGESDIR:=$(wildcard $(FPCDIR) $(FPCDIR)/packages/base $(FPCDIR)/packages/extra) override PACKAGE_NAME=fpcunittestrunner override PACKAGE_VERSION=0.1 +ifndef LCL_PLATFORM +ifeq ($(OS_TARGET),win32) +LCL_PLATFORM=win32 +else +LCL_PLATFORM=gtk +endif +endif +export LCL_PLATFORM ifeq ($(FULL_TARGET),i386-linux) override TARGET_UNITS+=fpcunittestrunner.pas endif @@ -454,115 +462,115 @@ override CLEAN_FILES+=$(wildcard $(COMPILER_UNITTARGETDIR)/*$(OEXT)) $(wildcard $(COMPILER_UNITTARGETDIR)/*$(PPUEXT)) $(wildcard $(COMPILER_UNITTARGETDIR)/*$(RSTEXT)) $(wildcard $(COMPILER_UNITTARGETDIR)/*.compiled) $(wildcard *$(OEXT)) $(wildcard *$(PPUEXT)) $(wildcard *$(RSTEXT)) endif ifeq ($(FULL_TARGET),i386-linux) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-go32v2) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-win32) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-os2) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-freebsd) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-beos) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-netbsd) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-solaris) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-qnx) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-netware) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-openbsd) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL
Re: Why is this throwing an exception - ANSWER
I applied only the ippanel change. As far as I know that should have been the only one anyway though it seems the tutorial on patch creation should be expanded to include make clean before creating a patch. A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TListBox and Font properties
On Sunday 26 February 2006 21:36, Panagiotis Sidiropoulos wrote: AFAIK it is not yet implemented in the LCL gtk2 interface, only in the LCL gtk1 and win32 interface. Does it need any special implementation? Why this for gtk1 can't work on gtk2? Maybe it is a silly question, I don't know. How someone can find infromation on how to do such things? The short version is because the LCL is NOT the interface it is using, lazarus compiling to GTK1 does not mean you are using GTK1. You are still compiling your program against the LCL, the LCL uses some or other widget set for rendering but the LCL itself has to implement the calls to let that widget set do whatever is needed. Right now I believe the win32 and GTK1 interfaces are by far the most complete since they are the oldest ones, carbon, qt and gtk2 are all in a much more immature state. The big thing is that you have the Object Pascal way of telling components how to behave on one side, and whatever method is typical in the widget set on the other side - the LCL basically translates the one to the other - so the translation to GTK1 for ownerdraw isn't the same as for GTK2 because GTK2 doesn't use the same calls as GTK1 did. With me so far ? Right now GTK2 is actually more complete than QT and Carbon mostly because a lot of the calls are similar to the highly mature GTK1 which allowed for a faster conversion, but it is not I think production ready. I have tried it and there are some glaring problems, the worst for me is that the GTK2 LCL has some kind of issue in event handling - a LOT of events just disappear into the bitbucket and nothing annoys a user as much as clicking a button and nothing happens, it is also MUCH slower than the GTK1 version but it IS improving and I don't think it will be very long before you CAN use it for production work. Now, you seem interested in helping to shorten that time, the place to start is the layout and readme text files in the LCL subdirectory of your lazarus source tree, get an idea of how the LCL interfaces are laid out, then go into the subdirectory that implements your target widget set and start hacking the code, when you are done, recompile lazarus and try it out, repeat. Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] Pretty looking TEdit
Hi all, Probably the most requested feature for tappytux is that the entry line where users type answers needs to be bigger (I mean huge) and prettier, using something like Comic Sans font at size 32 should do it, but of course lazarus provides no means of setting fonts and gtkrc seems to only be able to do it globally - not for one widget type only. Which is why I am thinking that what I would need to do is create a component that is canvas drawn and knows how to output text in the font type of your choosing, in fact I think textout to a color-painted region might almost do it, however there is a catch, the component HAS to be UTF-8 enabled because tappytux is VERY internationalised and UTF is just about the only way to do that cleanly. Now for the most part textout DOES display UTF-8 but the hard bit is input. I have no idea how UTF-8 is mapped to keycodes, if I know how to handle an onkeydown event for a component in such a way as to ensure that the character displayed is what was actually TYPED (which includes handling things like Del and Backspace right) then my problem is solved. Of course such a component is sufficiently generically usefull that I imagine others would bennifit as well and as always I will lgpl it. Any advice on how it should work ? What it should be descended from and how to get from input to screen ? Ciao A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
On Monday 27 February 2006 10:47, Micha Nelissen wrote: A.J. Venter wrote: Hi all, Probably the most requested feature for tappytux is that the entry line where users type answers needs to be bigger (I mean huge) and prettier, using something like Comic Sans font at size 32 should do it, but of course lazarus provides no means of setting fonts and gtkrc seems to only be able to do it globally - not for one widget type only. Are you sure gtk is really incapable of showing a different font for certain widgets ? IIRC Darek implemented Color property for some widgets, so I guess fonts should be possible as well. (But I have no gtk experience). If this is possible, then this is *the* way to go. I spent about 4 hours last night reading gtkrc howto´s and studying existing theme files and thus far I cannot find anything to do this, even then it STILL requires an external gtkrc file, I think a component for those cases where you do need it would be very usefull, the same way a tstatictext can be used instead of a tlabel when you just need output. A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
On Monday 27 February 2006 11:40, Mattias Gaertner wrote: On Mon, 27 Feb 2006 11:15:03 +0200 A.J. Venter [EMAIL PROTECTED] wrote: On Monday 27 February 2006 10:47, Micha Nelissen wrote: A.J. Venter wrote: Hi all, Probably the most requested feature for tappytux is that the entry line where users type answers needs to be bigger (I mean huge) and prettier, using something like Comic Sans font at size 32 should do it, but of course lazarus provides no means of setting fonts and gtkrc seems to only be able to do it globally - not for one widget type only. Are you sure gtk is really incapable of showing a different font for certain widgets ? IIRC Darek implemented Color property for some widgets, so I guess fonts should be possible as well. (But I have no gtk experience). If this is possible, then this is *the* way to go. I spent about 4 hours last night reading gtkrc howto´s and studying existing theme files and thus far I cannot find anything to do this, even then it STILL requires an external gtkrc file, I think a component for those cases where you do need it would be very usefull, the same way a tstatictext can be used instead of a tlabel when you just need output. I just tried under linux/gtk1: Setting a TEdit font size to 30 gave a a huge font. Then this may be the right way, although I am using GTK2 here (I need the pango UTF support), and gtk_rc_parse will allow me to force a specific theme file. How exactly does your line look ? Any idea how I need ot change it to use GTK2 ? Or did you do this in code ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
Any idea how I need ot change it to use GTK2 ? Or did you do this in code ? I did it in the designer and it works at runtime with gtk1. For gtk2: This needs to be implemented in gtkobject.inc TGtkWidgetSet.SetWidgetFont. I am studying that file now, thinking off it, my way wouldn´t work anyway, without SetWidgetFont there isn´t any real way for a custom component to force it´s font either (except perhaps for direct GTK2 calls in which case I may as well just do it for the LCL and be done with it). Not being a GTK guru I am a little lost but mostly understanding what the code is meant to be doing and I am fairly certain that (some of) the code in PangoDrawText can be used to implement this (since I need it, I am happy to try at least), what I do need however is some idea of how it should work, where is it defined ? Procedure ? Function (result type) ? Unfortunately the lazarus-ccr seems to be down (could be my ISP though) so I cannot look it up there and I cannot find such a function in the GTK1 file of the same name at all. I assume it´s specified somewhere else ? Presumably the same place where the debugln for setWidgetFont comes from. Anybody up for helping me ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] Cannot run r8833
Not sure how far back this goes because I only update ever few days, but I cannot run this version, after a clean build, lazarus begins to start, shows the splash screen and the lazarus editor window and freezes (nothing ever gets painted IN the editor window). On the shell the last output is: TMainIDE.DoNewEditorFile A Filename= Thereafter absolutely nothing happens, I did an strace but I couldn´t figure out what it was trying to do, just that it didn´t actually do anything beyond that point. At first I thought it was just my home machine but now I have the same result at the office. Obviously I cannot try to do the font stuff until this happens, it could mean some new incompatibility with openlab (I did test - removing $HOME/.lazarus had no impact nor does trying to run it as root), though I wouldn´t know where to start looking, let me know if I can provide any usefull information like LDD outputs or something. Note that at this stage we are just talking about a clean make, no GTK2 or anything yet, not even any extra packages. Ciao A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
Well I did some research and the GTK way of doing it is really very simple, except that the method (modify_font) doesn´t seem to exist in PGTKWidget which I surmise is because the definition of THAT is taken from the GTK1 definition. According to the GTK2 docs, all GTK2 widgets derive from GTKWidget and should inherrit modify_font which can set any pangoFontDescription, the only remaining part is to convert a tfont TO a PangoFontDescription (well PPangoFontDescription actually). So the RIGHT way I believe is to enable the method, presumably by descending PGTKWidget to create PGTK2Widget (oddly TGK2Widget exists but TGKWidget does not) - which may have the inadvertent side effect of meaning that damn near every bit of GTK needs to have it´s type definitions altered to call and use the derived class, either that or GTK2 should simply overWRITE the PGTKWidget definition with it´s own version that enables the GTK2 specific features. Thoughts anyone ? Ciao A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: Why is this throwing an exception - ANSWER
On Monday 27 February 2006 18:19, Vincent Snijders wrote: A.J. Venter wrote: I applied only the ippanel change. As far as I know that should have been the only one anyway though it seems the tutorial on patch creation should be expanded to include make clean before creating a patch. That would not have helped in this case, because most of your changes were because of local changes to files in the repository. Vincent. Then I really dunno what I did wrong because that was the ONLY changed file at the time, at least the only one *I* had changed - does lazarus change files during runtime or installing of packages ? How SHOULD I have created the patch ? A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
If this is possible, then this is *the* way to go.
At least I know it is possible for gtk2 - I implemented dynamic fonts in
one of my (C++, unfortunately) programs. The documentation seems to talk
about the font code as something New and Improved (tm), so I'd say it's
quite hard in gtk1.
Well I could find the documentation on how to do it, but the actual method
isn't implemented somewhere, it is not impossible that the PLACE where it is
missing is in the FPC units, but since those were autogenerated my gutt tells
me that it is more likely to be somewhere in the LCL.
I am not exactly sure how the GTK2 objects were turned into OP classes for the
LCL, I would have to look at the wrapper code first, my feeling however is
that the likely scenario is that somewhere an external declaration needs to
be added that can enable GTKWidget.Modify_font from the GTK2 libs.
GTK1 already has support for this, and if I I didn't REALLY need UTF-8 support
for text-out I would have just compiled my app for GTK1 instead but since I
don't know how, and it would be GREAT if this works, I am happy to help write
code to make it happen, if I can get help from some of the gtk2 guru's it
would just happen faster.
My current code:
In gtk2object.inc add this (I put mine just above TGTKWidget.SetCallBack):
Procedure TGTK2WidgetSet.SetWidgetFont(const AWidget: PGTK2Widget; const
Afont: tFont);
Var
PangoDescStr,DescOpts : String;
NewFontDescription : PPangoFontDescription;
Begin
{$IFDEF GTK2}
PangoDescStr := AFont.Name;
DescOpts := '';
If FSBold in AFont.Style then
DescOpts := DescOpts + ' bold';
If FSItalic in AFont.Style then
DescOpts := DescOpts + ' italic';
If FSUnderline in AFont.Style then
DescOpts := DescOpts + ' underline';
If FSStrikeOut in AFont.Style then
DescOpts := DescOpts + ' strikethrough';
PangoDescStr := PangoDescStr+DescOpts+' '+intToStr(Afont.Size);
//Pango does not appear to have a way to set the character set in the
//font description but seems to default to UTF-8 this probably
//requires some or other todo item.
try
NewFontDescription := PPangoFontDescription(PangoDescStr);
AWidget.Modify_Font(NewFontDescription);
except
debugln('COULD NOT SET FONT');
end;
{$ENDIF}
end;
In GTK2Int.pas change
protected
procedure AppendText(Sender: TObject; Str: PChar);
to:
protected
procedure SetWidgetFont(const AWidget: PGTK2Widget; const Afont: tFont);
procedure AppendText(Sender: TObject; Str: PChar);
Note that the current code DOES NOT COMPILE due to the missing Modify_Font.
PS. Thanks Darek, I managed to get THIS far by studying your patch and will
look at it again when doing the per-widget bits.
PPS. As you can see, right now tfont.color is NOT yet handled at all.
Ciao
A.J.
--
A.J. Venter
Chief Software Architect
OpenLab International
www.getopenlab.com
www.silentcoder.co.za
+27 82 726 5103
_
To unsubscribe: mail [EMAIL PROTECTED] with
unsubscribe as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
Correct. There must be already such a function, but I'm not sure if it works good. There are a lot of font issues in the gtk2 interface. Well if it's there, it's not there under the same name :) So the RIGHT way I believe is to enable the method, presumably by descending PGTKWidget to create PGTK2Widget (oddly TGK2Widget exists but TGKWidget does not) - which may have the inadvertent side effect of meaning that damn near every bit of GTK needs to have it´s type definitions altered to call and use the derived class, either that or GTK2 should simply overWRITE the PGTKWidget definition with it´s own version that enables the GTK2 specific features. I'm not sure what this TGTK2Widget is. It is defined in gtk2int.pas and seems to to be the parent widget for gtk2 widgets, except it's a LOT shorter than the TGTKWidget definition in gtkint.pas (this was where I had to add the procedure definition for setWidgetFont), and it doesn't seem to be derived, I have not found any reference to PGTK2Widget however leading me to presume it doesn't exist, the procedure reference right now uses PGTKWidget since that is what I got from Darek's patch, but I am fairly certain that it DOES need to change to be a pointer to a TGK2Widget, I am still trying to figure out where to DO that though, it is not impossible that defining PGTK2Widget and making whatever changes is needed to allow SetWidgetFont calls under GTK2 to pass the right type will solve the problem entirely - of course a GTK2 only method won't exist if we are using a GTK1 class definition to refer to the widget objects in calls. Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
try NewFontDescription := PPangoFontDescription(PangoDescStr); AWidget.Modify_Font(NewFontDescription); I will take a look ... Mucios Gracias Amigo. A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Pretty looking TEdit
On Monday 27 February 2006 18:51, Mattias Gaertner wrote: font_desc:=pango_font_description_from_string(PChar(PangoDescStr)); gtk_widget_modify_font(AWidget,font_desc); pango_font_description_free(font_desc); Right my patch is ALL BUT complete, one thing is missing though, This code FAILS is tfont.name is set to Default, I need to CATCH that and give it a SANE default font name in that case, what should I use ? Ideally it should be whatever lazarus normally uses. A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] GTK2 Font patch - build 1
This is very nearly ready, but it still has two bugs I don't know how to fix - hopefully though it will provide enough of a basis that somebody can help me finish it off. Bugs: 1) Tedit doesn't display unless the font is EXPLICITELY set (e.g. my current catch code doesn't seem to work for it) 2) TLable sets the font at designtime but not at runtime -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 Index: components/fpcunit/Makefile.fpc === --- components/fpcunit/Makefile.fpc (revision 8836) +++ components/fpcunit/Makefile.fpc (working copy) @@ -7,7 +7,7 @@ [compiler] unittargetdir=lib/$(CPU_TARGET)-$(OS_TARGET) unitdir=../../lcl/units/$(CPU_TARGET)-$(OS_TARGET)/ ../../lcl/units/$(CPU_TARGET)-$(OS_TARGET)/$(LCL_PLATFORM)/ ../../packager/units/$(CPU_TARGET)-$(OS_TARGET)/ -options=-dLCL -dLCLgtk -S2 -gl +options=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl [target] units=fpcunittestrunner.pas @@ -18,6 +18,16 @@ $(wildcard $(COMPILER_UNITTARGETDIR)/*$(RSTEXT)) \ $(wildcard $(COMPILER_UNITTARGETDIR)/*.compiled) \ $(wildcard *$(OEXT)) $(wildcard *$(PPUEXT)) $(wildcard *$(RSTEXT)) +[prerules] +# LCL Platform +ifndef LCL_PLATFORM +ifeq ($(OS_TARGET),win32) +LCL_PLATFORM=win32 +else +LCL_PLATFORM=gtk +endif +endif +export LCL_PLATFORM [rules] .PHONY: cleartarget all Index: components/fpcunit/Makefile === --- components/fpcunit/Makefile (revision 8836) +++ components/fpcunit/Makefile (working copy) @@ -1,5 +1,5 @@ # -# Don't edit, this file is generated by FPCMake Version 2.0.0 [2006/01/19] +# Don't edit, this file is generated by FPCMake Version 2.0.0 [2005/12/07] # default: all MAKEFILETARGETS=i386-linux i386-go32v2 i386-win32 i386-os2 i386-freebsd i386-beos i386-netbsd i386-solaris i386-qnx i386-netware i386-openbsd i386-wdosx i386-emx i386-watcom i386-netwlibc i386-wince m68k-linux m68k-freebsd m68k-netbsd m68k-amiga m68k-atari m68k-openbsd m68k-palmos powerpc-linux powerpc-netbsd powerpc-macos powerpc-darwin powerpc-morphos sparc-linux sparc-netbsd sparc-solaris x86_64-linux x86_64-freebsd x86_64-win64 arm-linux arm-wince powerpc64-linux @@ -231,6 +231,14 @@ PACKAGESDIR:=$(wildcard $(FPCDIR) $(FPCDIR)/packages/base $(FPCDIR)/packages/extra) override PACKAGE_NAME=fpcunittestrunner override PACKAGE_VERSION=0.1 +ifndef LCL_PLATFORM +ifeq ($(OS_TARGET),win32) +LCL_PLATFORM=win32 +else +LCL_PLATFORM=gtk +endif +endif +export LCL_PLATFORM ifeq ($(FULL_TARGET),i386-linux) override TARGET_UNITS+=fpcunittestrunner.pas endif @@ -454,115 +462,115 @@ override CLEAN_FILES+=$(wildcard $(COMPILER_UNITTARGETDIR)/*$(OEXT)) $(wildcard $(COMPILER_UNITTARGETDIR)/*$(PPUEXT)) $(wildcard $(COMPILER_UNITTARGETDIR)/*$(RSTEXT)) $(wildcard $(COMPILER_UNITTARGETDIR)/*.compiled) $(wildcard *$(OEXT)) $(wildcard *$(PPUEXT)) $(wildcard *$(RSTEXT)) endif ifeq ($(FULL_TARGET),i386-linux) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-go32v2) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-win32) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-os2) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-freebsd) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-beos) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-netbsd) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-solaris) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-qnx) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-netware) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-openbsd) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-wdosx) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL$(LCL_PLATFORM) -S2 -gl endif ifeq ($(FULL_TARGET),i386-emx) -override COMPILER_OPTIONS+=-dLCL -dLCLgtk -S2 -gl +override COMPILER_OPTIONS+=-dLCL -dLCL
[lazarus] GTK2 SetFont Patch - final (for now)
Right then, I tested this a bit and it seems to work pretty well unless I am
still missing something, this version ALSO includes font.color support.
As a whole this should probably ALL be considered experimental code but it
(probably) won't crash your apps.
At least I believe it provides a good base from which it can now be improved
as any special cases reveal themselves and if any bugs are reported I will do
my best to help fix them.
For now though, that's three patches in one night so I think I tortured
everyone's mailboxes quite enough :p
Ciao
A.J.
--
A.J. Venter
Chief Software Architect
OpenLab International
www.getopenlab.com
www.silentcoder.co.za
+27 82 726 5103
Index: lcl/interfaces/gtk/gtkwscontrols.pp
===
--- lcl/interfaces/gtk/gtkwscontrols.pp (revision 8836)
+++ lcl/interfaces/gtk/gtkwscontrols.pp (working copy)
@@ -30,7 +30,7 @@
uses
{$IFDEF GTK2}
- Gtk2, Glib2, Gdk2,
+ Gtk2, Glib2, Gdk2,Pango,
{$ELSE}
Gtk, Glib, Gdk,
{$ENDIF}
@@ -392,6 +392,14 @@
const AFont: TFont);
var
Widget: PGtkWidget;
+{$IFDEF Gtk2}
+ PangoDescStr,DescOpts : String;
+ NewFontDescription : PPangoFontDescription;
+ Size : Integer;
+ Name : String;
+ xfg: TGDKColor;
+ State : Longint;
+{$ENDIF}
begin
if not AWinControl.HandleAllocated then exit;
Widget:=pGtkWidget(AWinControl.handle);
@@ -399,13 +407,59 @@
exit;
if AFont.IsDefault then exit;
+{$IFDEF Gtk}
DebugLn('TGtkWSWinControl.SetFont ',DbgSName(AWinControl));
GtkWidgetSet.SetWidgetFont(Widget,Afont);
GtkWidgetSet.SetWidgetColor(Widget,AWinControl.font.color, clNone,
[GTK_STATE_NORMAL,GTK_STATE_ACTIVE,
GTK_STATE_PRELIGHT,GTK_STATE_SELECTED]);
+{$ENDIF}
+{$IFDEF Gtk2}
+
+ If Afont.IsDefault then Exit;
+
+ If AFont.Size = 0 then
+ Size := 10
+ else
+ Size := AFont.Size;
+
+ if (length(AFont.Name) = 0) or
+(upperCase(AFont.Name) = 'DEFAULT') then
+ Name := 'Sans'
+ else
+ Name := AFont.Name;
+ DebugLn('TGtkWSWinControl.SetFont(GTK2) ',DbgSName(AWinControl));
+ PangoDescStr := Name;
+ DescOpts := '';
+ If FSBold in AFont.Style then
+ DescOpts := DescOpts + ' bold';
+ If FSItalic in AFont.Style then
+ DescOpts := DescOpts + ' italic';
+ If FSUnderline in AFont.Style then
+ DescOpts := DescOpts + ' underline';
+ If FSStrikeOut in AFont.Style then
+ DescOpts := DescOpts + ' strikethrough';
+
+ PangoDescStr := PangoDescStr+DescOpts+' '+intToStr(Size);
+//Pango does not appear to have a way to set the character set in the
+//font description but seems to default to UTF-8 this probably
+//requires some or other todo item.
+try
+NewFontDescription := PPangoFontDescription(PangoDescStr);
+NewFontDescription :=pango_font_description_from_string(PChar(PangoDescStr));
+ gtk_widget_modify_font(Widget,NewFontDescription);
+pango_font_description_free(NewFontDescription);
+//And now the font color
+ xfg:=AllocGDKColor(colorToRGB(AFont.Color));
+for State := GTK_STATE_NORMAL to GTK_STATE_SELECTED do
+ gtk_widget_modify_fg(Widget, State ,@xfg);
+except
+debugln('COULD NOT SET FONT');
end;
+{$ENDIF}
+end;
+
procedure TGtkWSWinControl.SetPos(const AWinControl: TWinControl;
const ALeft, ATop: Integer);
var
Index: lcl/interfaces/gtk2/gtk2int.pas
===
--- lcl/interfaces/gtk2/gtk2int.pas (revision 8836)
+++ lcl/interfaces/gtk2/gtk2int.pas (working copy)
@@ -54,6 +54,7 @@
TGtk2WidgetSet = class(TGtkWidgetSet)
protected
+procedure SetWidgetFont(const AWidget: PGTKWidget; const Afont: tFont);
procedure AppendText(Sender: TObject; Str: PChar);
function CreateComponent(Sender : TObject): THandle; override;
function GetText(Sender: TComponent; var Text: String): Boolean;
@@ -120,6 +121,7 @@
property Sorted : boolean read FSorted write SetSorted;
property Owner: TWinControl read FOwner;
end;
+
{$IfDef GTK2_2}
procedure gdk_display_get_pointer(display : PGdkDisplay; screen :PGdkScreen; x :Pgint; y : Pgint; mask : PGdkModifierType); cdecl; external gdklib;
Index: lcl/interfaces/gtk2/gtk2object.inc
===
--- lcl/interfaces/gtk2/gtk2object.inc (revision 8836)
+++ lcl/interfaces/gtk2/gtk2object.inc (working copy)
@@ -959,6 +959,69 @@
Applies a Message to the sender
--}
+
+Procedure TGTK2WidgetSet.SetWidgetFont(const AWidget: PGTKWidget; const Afont: tFont);
+Var
+ PangoDescStr,DescOpts : String;
+ NewFontDescription : PPangoFontDescription;
+ Size : Integer;
+ Name : String;
+ xfg: TGDKColor;
+ State: Longint;
+Begin
+{$IFDEF GTK2}
+If Afont.IsDefault then Exit;
+
+ If AFont.Size = 0 then
+ Size
Re: [lazarus] TListview sorting
On Tuesday 28 February 2006 12:45, Graeme Geldenhuys wrote: Hi AJ, I might be to late for you, but this is how I did my sorting in a TListview. It doesn't need to clear the Listview while sorting. You might need to reset the .Selected though. This code goes into the ColumnClick event of the TListview: Thanks for this, though indeed it won´t help me now. Having USED a grid and seen the power of using cell-draw events to display rather than maintaining a list (instant speed increases all around - and BIG NOTICEABLE ones) I am certainly not going back, not while my program needs to handle an average of 12000 entries per session. That said, I haven´t studied your code in detail but the thought occurs to me that nearly every listview sort tutorial for delphi uses an assignment to the customsort event, just about nobody seems to use built-in sorts so perhaps the easiest way to get more compatible is simple to export the right properties from the LCL to ALLOW people to assign a customsort method - which would of course still have to be written by the user. This won´t implement tlistview.sort (for which mind you, your code may be a good start if you feel like working out how to make it a patch) but it WILL allow all those customsort based projects to work on lazarus. Thoughts ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] Somewhat OT: The massive db-less search
Right, the final missing feature of OpenBook is to be able to search for phrases INSIDE the books, now the logical way would be to just check each book and determine whether or not it contains a matching phrase... except that there are 12000 of them. For the author/title searching it´s easy, I have a prebuilt index of files matching them to these details and I JUST search the index, I cannot rely on any kind of SQL style datbase (except maybe SQLite or something else that can work without a server) since the program must be able to run from DVD. So the question arises how best to do it, searching for a phrase in 12000 books one by one will take FOREVER. Ideally I need to somehow build a wordlist/phrase index and search that, something like what htt://dig does, but I need to pass the results back to my lazarus code in a format I can easilly map back to the index (so I can display author/title information with the results) the last part is easy if the indexing contains a filename, I can just look up the filename from the author/title index. So my question is can anybody recommend a good tool for creating such a search index ? It should either create it as some kind of easilly parseable text format, or alternatively something like an sqlite database could work as well. I just reckoned before I test a hundred apps, I would ask if anybody has any suggestions on which to try first. One crucial element is that OpenBook is multiplatform with both windows and linux versions so it´s vital that whatever method I use to do this search behind the scenes is ALSO multiplatform. TIA A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Somewhat OT: The massive db-less search
On Tuesday 28 February 2006 17:24, Vincent Snijders wrote: If java is an option for you: http://lucene.apache.org/java/docs/ If not, maybe you can port it to fpc. We use this (the .NET port) at work to index all publications of Statistic Netherlands. Searching is fast. Thanks, I am looking now, there is of course a nice catch most search engins do word-list indexing, which is FINE for web-pages, but NOT for searching 12000 books as just about every search would match nearly every book - a book is MUCH more data than a web-page. So litterally the only in data search that would give more or less usefull results is full-sentence searches - e.g ALL the words you entered IN THE ORDER you entered them DIRECTLY juxtaposed - easier in one sence since a substring search will either find an exact match or none at all, but harder in that wordlist indexing simply will not work. Looking at things like lucene and egothor it seems that they actually want to search the files themselves... all good and well except for a catch - all the files are gz compressed, openbook has on-demand decompression built-in - so users don't even need to know about it, the file just appears to open from a users PoV. Now this is not to say that using the indexes from such a search index will not work - I can index on the uncompressed copy and then just use the data - but somehow I just don't see keyword based searching as being truly usefull here, the data is just too different. Most large document warehouses have fairly diverse data in each document, but this is a disk full of books - most of them fiction, in other words the data you are talking about here is several megabytes per file, highly repetitive (in computing terms) and not very diverse (again in computing terms). A character name will probably get you only a few books, but a search like Here's looking at you kid is supposed to get pretty much only cassablanca, not every book that ever used the words looking and kid (which are the ones in that phrase which typical keyword searches would consider uncommon). Frankly I am ready to tell my boss it cannot be done, doing per-file searching on the DVD is likely to take a few DAYS per result, and I just don't think you can DO this kind of search from metadata. Well maybe if I could stick wikiquotes in there and then compare the results to my available book list - of course wikiqoutes is about 20GB and needs a webserver etc. - so it cannot exactly run from a DVD. Basically unless somebody already knows how to do this, I am happy to admit I am not smart enough to solve THIS one :) A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] GTK2 SetFont Patch - final (for now)
On Wednesday 01 March 2006 10:36, Panagiotis Sidiropoulos wrote: Are all these changes available just by svn update and recompiling Lazarus or anything more is needed? Panagiotis Yes, since Mattias applied the patches to the main tree. At this stage I am waiting for more clarity on where Darek's GTK1 work is going before advancing mine so that the two can be kept in design sync. Ciao A.J. PS. Darek, thanks for your note earlier, but my name is AJ not AI (though that might have been cool :p ) -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] The Latest IDE icons
I vote to change them back to original ,too.Icons are too large and almost lay on each other (visually it's hard to find needed control on tab Standard or Additional) Some of them are good looking however. Whit the images itself I don't have a problem. They don't differ much from the original (its this thread which brought it to my attention). But indeed, I think the buttons might be somewhat bigger. I like some of them, and dislike others myself but I LIKE the idea of a distinct lazarus icon set, some of the component buttons are too small though (or perhaps the icons are too big ?) On a related noted, I have been thinking that it would be great to have a set of developer-icons which can be (optionally) installed with lazarus for use by lazarus programmers inside apps. Now I don't feel like starting from nothing, so my suggestion would be to take the icon set from kylix to get a LIST of what icons are shipped with delphi (but NOT use those - licensing) and then take some very good LGPL'd icon set (I love nuvola so it has my vote), take the 16x16 icons from that and do a set of xpm/ico conversions selecting a replacement for everything on the kylix/delphi list. The default location of this set could then be the default start directory for the IDE image-open dialog - making it just that much easier on somebody using it for the first time as they will be able to instantly find icons for their image-lists etc. What do you all think about this ? I am happy to do the work but obviously I won't bother if nobody else thinks it's a good idea. Since it will be fairly static, it may not have to go into the SVN tree, perhaps it should be distributed as a separate zip/tarball - all off of these things are worth thinking about and discussing. Ciao A.J. -- A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Changing SynEdit - actions
- Move all (all non-basic at least) editor commands to a separate unit and implement each command on a separate method or class. I'm inclined to make them TAction descendants to unify shortcut handling, but I'm not set yet... What would we gain? Reuse ? In theory (depending on how that unit is implemented). It should be designed to work on any stringlist (which is what I believe the underlying data-set in synedit is), so it can update a synedit, or a memo or whatever and can be used for other components or heck if called from code directly to do easy advanced stringset parsing. Imagine if we could tell a programmer that the stringset he is trying to modify in a certain way before displaying - well he can access the search| replace menu's own command from his code with all it's features|power. Right now most string parsing seems to be handled by pos, copy and delete - which works but why should people need to write the same code over and over when lazarus already contains it ? A.J. Some might notice I am highly in favour of lazarus having MORE than delphi ever did - to make delphi users SWOON over it :) -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] The Latest IDE icons
On Wednesday 01 March 2006 11:29, Andrew Higgs wrote: A.J. Venter wrote: snip What do you all think about this ? I am happy to do the work but obviously I won't bother if nobody else thinks it's a good idea. Since it will be fairly static, it may not have to go into the SVN tree, perhaps it should be distributed as a separate zip/tarball - all off of these things are worth thinking about and discussing. Ciao A.J. I like the idea. Let me know if there is anyway I can assist. Well then, why not split the job evenly ? I'll do the xpm's, and you can do the ICO version ? The first part is going to be selecting the right icon replacements, I have no doubt I will on occasion be stuck between choices, at those times I would like to put it up for a vote if the list admins have no objections to me sending the ones I am looking at an asking which the users would prefer. A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Changing SynEdit - actions
Right now most string parsing seems to be handled by pos, copy and delete - which works but why should people need to write the same code over and over when lazarus already contains it ? The same reason, why there are so many databases. Touché, anyway your explanations did give me a lot more clarity on the nature of the underlying code, and I can see that making it generic would be a lot of work, perhaps not worth it unless somebody has an itch. Ciao A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Somewhat OT: The massive db-less search
Actually I found a nicer solution :) I integrated with wikiquote.org (which was something I came up with while discussing the problem with you guys). Selecting a phrase search hides the book and submits the search to wikiquote, grabs the results, preparses them and displays the list of matches in an iphtml panel, clicking on any of the links in the list, then submits the link value as a search back into my indexing in BOTH artist and title fields. So if you search for ¨Wherefore art though romeo¨ one of the results will be ¨William Shakespear¨, clicking the link brings up the entry ¨Shakespear´s first folio¨ by William Shakespear from my list, which happens to contain among it´s 35 plays ¨Romeo and Juliet¨. True it requires an internet connection but what it doesn´t require is any real CPU/disk usage - and it has the power of an index maintained by thousands of volunteers :), these days programs being able to integrate cleanly with online information sources is considdered a good thing right ? it was for this that I needed the connectivity check, so that I could disable the quote-search checkbox if there was no internet. Ciao A.J. On Tuesday 28 February 2006 17:51, William Cairns wrote: Have you considered a two pass approach? ie Do the first regular search using Looking and Kid to get a list of the books that might have the full phrase in it. Then only decompress and search in those books for the full phrase. -Original Message- From: A.J. Venter [mailto:[EMAIL PROTECTED] Sent: 28 February 2006 17:44 PM To: lazarus@miraclec.com Subject: Re: [lazarus] Somewhat OT: The massive db-less search On Tuesday 28 February 2006 17:24, Vincent Snijders wrote: If java is an option for you: http://lucene.apache.org/java/docs/ If not, maybe you can port it to fpc. We use this (the .NET port) at work to index all publications of Statistic Netherlands. Searching is fast. Thanks, I am looking now, there is of course a nice catch most search engins do word-list indexing, which is FINE for web-pages, but NOT for searching 12000 books as just about every search would match nearly every book - a book is MUCH more data than a web-page. So litterally the only in data search that would give more or less usefull results is full-sentence searches - e.g ALL the words you entered IN THE ORDER you entered them DIRECTLY juxtaposed - easier in one sence since a substring search will either find an exact match or none at all, but harder in that wordlist indexing simply will not work. Looking at things like lucene and egothor it seems that they actually want to search the files themselves... all good and well except for a catch - all the files are gz compressed, openbook has on-demand decompression built-in - so users don't even need to know about it, the file just appears to open from a users PoV. Now this is not to say that using the indexes from such a search index will not work - I can index on the uncompressed copy and then just use the data - but somehow I just don't see keyword based searching as being truly usefull here, the data is just too different. Most large document warehouses have fairly diverse data in each document, but this is a disk full of books - most of them fiction, in other words the data you are talking about here is several megabytes per file, highly repetitive (in computing terms) and not very diverse (again in computing terms). A character name will probably get you only a few books, but a search like Here's looking at you kid is supposed to get pretty much only cassablanca, not every book that ever used the words looking and kid (which are the ones in that phrase which typical keyword searches would consider uncommon). Frankly I am ready to tell my boss it cannot be done, doing per-file searching on the DVD is likely to take a few DAYS per result, and I just don't think you can DO this kind of search from metadata. Well maybe if I could stick wikiquotes in there and then compare the results to my available book list - of course wikiqoutes is about 20GB and needs a webserver etc. - so it cannot exactly run from a DVD. Basically unless somebody already knows how to do this, I am happy to admit I am not smart enough to solve THIS one :) A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] GTK2 SetFont Patch - final (for now)
Yes, since Mattias applied the patches to the main tree. At this stage I am waiting for more clarity on where Darek's GTK1 work is going before advancing mine so that the two can be kept in design sync. What I can do for You now. Heh, well what would actually be very usefull is if you send me a list of what you added for which widgets which were not handled by the existing code - then I will have a good idea where to start implementing for mine, I already started picking up some special cases, notably that right now setcolor works for tpanel but setfont does not. SInce gtk1 and gtk2 are so similiar I think it is a good thing if we keep our work synchronized because a lot of what one of us does will influence how the other should do things. What I think might be wise as well is if we start a wiki page (when the wiki is back) documenting which widgets we change and how - this will be very usefull to those working on other widget sets, e.g. the QT and Carbon set guys. While it won´t apply 100% to their sets if we can point out 80% of the special cases for fonts and colours, they only have to do 20% as much work :) What do you think ? A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Repository access via the SVN protocol
Our german FTP machine was compromised recently (that is why the wiki is down). The main server has also been compromised some time ago. Since then, I am reluctant to open too much ports. In reality, the problem is with your ISP, namely that they don't seem to support all WebDAV commands in their proxy. What about letting apache listening to a second port ? What good would that do ? Fooling the proxy ? Yes, mostly transparent proxies are configured for known http ports like 80, 1080 and 8080 and wont affect other ports. Hiya Graem Sorry I didn´t respond earlier, I was a bit tied up with getting my new bike registered :) Anyway, just for the record, I have had no problems accessing either http or svn based subversion sites from sentech, so I presume either your own ISP or telkom did something amazingly stupid on their transparent proxy. Of course we do not live in the kind of free market system where if a company buggers up like that they better appologise and fix it before they lose a customer, we live in a ¨internet-consumer¨ system with a single government sanctioned monopoly provider who doesn´t care two hoots about their customers OR their staff (what kind of a company records multiple-bilions in profit and then fires 7000 employees in the same month ? A very unethical one). Oh well, moaning about it won´t fix it, I just wanted to let you know that sentech is coming through clean. Ciao A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Object pascal a Modern Language
python is a very nice language. Why don't you just write the speed critical parts in C or C++. That's how everyone does it in python. Showing only that python isn't up to the task. Just like all these other interpreted languages: .NET, Java... It's cheating, and they all do it: falling back to compiled languaged. My message to the designers of these languages: Quit fooling around, and do things right from the start. This does bring up an interesting point, why is it that more and more languages are becoming interpreted or bytecompiled ? I mean, surely the python designers could have written platform level compilers for no more difficulty than to write bytecompilers ? Heck they even got jython - compiling to the java vm so they must have had the option. Instead it offers the choices of bytecompiled or scripted. Why is source portability not enough ? Why sacrifice so much speed and power to get binary compatibility (or at least cheat so it looks like that) ? Of course that is not the only power that gets lost, memory management, pointers - there is so much these approaches sacrifices, now I know their designers love to say things like if you need pointers you are doing it wrong but in my mind that is a cop-out to avoid dealing with the fact that they make it impossible for programmers to use a truly powerful construct that can do some pretty amazing things. The simple truth is that even java's pseudo implementations of things like hash tables to get around the fact that in other languages those are done with pointers aren't good - I have USED java's hash table interface, it's horrid ! As I told a SUN employee a few days ago I will stick with having a truly powerful source-portable fully object oriented language that doesn't practise bondage-and-discipline on style for as long as such languages exist - in short, I will use and promote lazarus - not java As an aside, I remain adamant that anybody who writes a free software/open source program in Java is being foolish. What is the POINT of a free software program that REQUIRES a non-free program just to RUN ? Ciao A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Object pascal a Modern Language
less power. It's always like that. The language of a human (and maths) is just so much more advanced than any computer language, it depresses me at times. (Assembly language is so primitive that we all stopped talking it by now, except when we absolutely must do it. But it sure is flexible, like an atom - you can use tons of them to build a copy of the earth, given you have a few billion years time and way too much patience for your own good :)) This is a nice argument but it's highly subjective, I find OP to be the computer language CLOSEST to human language, python and java both inherrited too much ascii art from C++ and don't get me STARTED on perl ! One of the things I love about OP (a compiled language) is that it uses mostly WORDS to say things, not symbols, true it's slightly more typing - but it is so much more readable. I always tell my trainee's If I NEED comments to read your code - it is badly written. Of course I also make it absolutely clear that any program of more than 100 lines which doesn't have comments is pretty much guaranteed to ensure they won't get a permanent position with me, but comments should be there to add clarity NOT to tell you what the code DOES ! That should be obvious just from reading the code. This isn't possible in most languages, in OP it's the default - that is why I think it's such a brilliantly designed language, and the fact is this is something Pascal has boasted since Wirth invented the first version of it. Pascal is probably the only language you can read (large sections of) aloud without stuttering. In other words, by your argument those languages have all failed at exactly the supposed REASON they are interpreted/bytecompiled in the first place. Like I said, it's way too subjective an argument. Now don't get me wrong, I actually rather like python, it has a nice design (although the whole whitespace has meaning thing annoys me a bit) and a fairly decent syntax but the simple reality is that I write code which will end up running on thin clients - that means if a computer slows down to any noticeable degree when running 50 copies of the program at once -it's not good enough, and that rules out python for anything complex. Lazarus works beautifully, I have the language that I like for most work, I can quickly design my GUI's in a RAD environment and the programs run FAST. In the end, Java has spent ten years telling programmers that computers are now so fast that customers don't care about speed anymore, and in ten years computers still haven't gotten that fast and frankly they never will because as computers get faster customers will always expect software to get faster at the same rate (and customers being customers they will expect it to be just as fast on a PII). In the real world, a program that runs slowly and uses up a lot of resources annoys customers and they find another program. Throw in the realities of Africa, and you end up utilising things like thin clients so that those PII's or sometimes even 486's can run modern day apps. Of course they are never quite as fast as on a P4 but they look as fast to users which is good enough -at least for compiled programs which do not abuse all the available resources in the system but use only what they need leaving the rest for the other users. In the end, the choice of a language is personal and not the least of your concerns should be what language does most of the people on this project know the best, but discounting OP because of what it wasn't in the 80's is ... well uninformed silliness. A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] setColor part 3
gtk does the right thing by letting the programmer register style properties for a widget class. These names can then be used in themes and gtk will assign the theme color to the style property of the widgets of that class on runtime. I know that it usually depends on the target audience whether it's ok to take shortcuts like hardcoding colors (if your program is to be used by an internal group of 5 people, new people never enter, and none of them has color blindness, sure, hardcode red all you like). _But_ when designing a framework, that is the wrong way to take because they affect all the users there are. What you are saying is correct, but not absolutely so. There are exceptions to every rule and the fact is programmers MUST have the capacity to sometimes enforce certain look and feel changes regardless of themes because themes are always generic (they are meant to apply to all apps) and sometimes a program is doing something which is simply specific to that program and the programmer needs a way of visually explaining a concept to users which simply does not exist in any theming system. Take tappytux for example (the program which was the reason why I started the gtk2 font setting patches). Number one requested feature from users was: make the text entry bar at the bottom HUGE. This was not MY choice, it was THEIR choice. This was not something that could be done for EVERY tedit in the program, but it HAD to be done for THAT tedit - because the nature of the program was such that, that entry bar had to be really big and accentuated. Why ? Because the program is for children, and it's a game - and people who only learned to read a month ago needs to be able to read what they type in there at a single glance. The rest of the program they have TIME to spell out letters, if they have to do it there they will never learn not to have to (at least not from the program which is supposed to have been written to help teach them this) because they will simply go game over before they finish the first word. See what I am getting at ? It's partially intended audience but it's also purpose of the program, sometimes a programmer needs to be able to enforce some things in the look and feel department. Finally there is the simple reality that despite the lack of credit popular culture gives us in this regard nearly all programmers are highly creative people. We LIKE to design interfaces and we like to be a little bit artistic when we do, and there is nothing wrong with that. Since we are the creators, we should be able to design something which is visually pleasing to us ? I am not saying we should force users to use bright-red buttons (and I know that green-red as distinguishing features is a horrible thing to do to users since about 8% of men cannot distinguish them), but I am saying that we SHOULD be able to sometimes say This is the main headline part of my main screen - I would really like it if it was in Mincho Gothic font in blue letters on a cream colored button -because it is not SUPPOSED to be the same as the other buttons - and that I believe is entirely a good thing, and something I encourage in my trainee's. Programmer designs should never detract from usability, but we should still be able to employ a little bit of creativity in interface design - the job of interface design is to take an abstract computing concept of some sort and create a visual analogy that allows non-computer experts to understand what they program does for them and how to achieve their goals with it. This IS a creative process and ingenuity can make all the difference here, especially with custom apps which are often the first to design for a certain concept. To imagine that any theme can be broad enough to handle ALL the problems any program will ever be asked to solve and still be usable as a theme is downright ridiculous. The default should always be to follow the theme, the programmer should think five or six times before overriding it - but when he does, it must be assumed that he had a good reason and his changes should be honoured. Ciao A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] setColor part 3
The default should always be to follow the theme, the programmer should think five or six times before overriding it - but when he does, it must be assumed that he had a good reason and his changes should be honoured. Exactly. Let me try to explain what I mean with an EXTREME example. Say you are coding the control system for a nuclear power station. Now 99% of your screen is showing the current values for things and with some sliders the operators can adjust some of the equipment settings to ensure that everything remains normal. For almost all times this is ALL there is to it. But there on one side is a few critical values, these ones won't just be changed by adjusting one thing, if they are outside tolerance values then that means a trained expert will need to adjust probably 80% of those sliders to very specific settings to deal with what is in fact an emergency. So when one of those goes out of tolerable range you start by activating every alarm in the building, syrens howling, strobe lights flashing the works. But on screen you do NOT want everything highlighted. What you DO want is for the area around the specific value that triggered the problem to start cycling through red-green-blue-orange-purple and some more colours at a fairly high rate so the operator your alarm woke up will INSTANTLY see which value is out of range and can start fixing it immediately. You don't want to pop up a dialog - clicking ok means it takes longer before he starts adjusting the parameters, but you MUST highlight THAT specific element in a way that should NEVER happen otherwise. The element may be a TEdit on a tgroupbox. Now if an emergency exists and you start colour-flashing the tgroupbox to draw attention to the RIGHT tedit but a theme overrides your colour calls so it doesn't flash and the whole town is flattened before the operator finds the right value (he cannot fix the problem until he knows WHAT the problem is) - well the whole theory just went up in smoke (as did the town). In a non computerized nuclear power station, a big LED would have flashed below the dial with the out of sync value, in a computerized version you HAVE to find a way of getting exactly the same kind of notification and theme must not prevent you from doing it, especially since computerising has so many other usefull advantages (a computer could solve simpler cases itself, like automatically increasing cooling if the temperature rises by a few degrees) in other words, a computerized interface is a good idea - but it MUST be able to override the theme when needed. Yes this is probably the most extreme example there is but it is a very real one, and just because most of us are not coding nuclear reactor controls doesn't mean that our reasons for occasionally having to FORCE colors or fonts are not good ones. A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] button icons for lazarus
After some positive feedback on the suggestion last week, I did the work and created a full set of button icons for lazarus. This set provides a replacement for everything in the standard delphi buttons directory under images. Right now, only xpm's exist. Most of the icons were taken from the Nuvola set for KDE (lgpl license) but a few came from CrystalSVG as well. Where I could not find a suitable replacement, I made it myself. Overall they are as consistent as I could get them and should hopefully help give lazarus apps a more consistent look and feel. Most importantly though I hope they will save lazarus programmers some time by providing prebuilt icons for most common button-tasks so that nobody has to sit and create/find icons first. The icon names were taken directly from their delphi counterparts for familiarity reasons (of course it also made it a lot easier for me to see what I had done yet and what was waiting). Andrew Higgs has offered to do ico format versions of them, I personally see no point in redoing the masked bitmap format that delphi used though I am open to the idea if somebody can give a good reason and will volunteer for the labour. However, I do think that before Andrew starts to do ico's I would like everyone interested to take a look at the icons and make suggestions for improvements, if you send me a better version of one of them, I will happily replace it. The bit which would make me very happy if guys on top were so inclined would be to see it included in the main lazarus tree eventually (based on it's delphi counterpart it should go under $LAZARUSDIR/images/buttons) but that should wait until after people had a chance to at least comment on the current set. I am wearing my asbestos undies so if you think I did a horrible job, you are welcome to tell me :) The icons are in my subversion server and can be checked out with: svn co svn://silentcoder.co.za/lazarus/buttons/ Ciao A.J. PS. Andrew, mail me with the username and password you would like, and I will give you write access to the tree. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
[lazarus] TRunInThread - Progress, but new problem
Well since my last mail on this a few weeks ago, I have made quite a bit of
progress in my attempts to create a component that will allow me to call a
tprocess in a thread while retrieving the output live.
The structure I have no compiles, and can be placed on a form - it is all but
right. The only trouble is - when I call the run method, it throws an access
violation, the stack trace places it at line 76 (the line where I call
FProcessThread.resume), I did a check and the first line of the execute
method does not ever get executed.
At this stage the whole thing consists of two units.
processthread.pas implements the TprocessThread class which is just a simple
TThread descendent which runs a command.
TRunInThread is a TComponent descendent which encapsulates TProcessThread with
a component, setting up an instantiation of it, executing it, storing and
updating the output etc.
In theory it should all work so I really do not know why it's throwing a
trace, perhaps somebody else can tell me what to change ?
I attach both the units for your viewing pleasure.
--
there's nothing as inspirational for a hacker as a cat obscuring a bug
by sitting in front of the monitor - Boudewijn Rempt
A.J. Venter
Chief Software Architect
OpenLab International
www.getopenlab.com
www.silentcoder.co.za
+27 82 726 5103
unit processthread;
{$mode objfpc}{$H+}
interface
uses
Classes, SysUtils,process;
Type
TProcessThread = class(tthread)
protected
procedure Execute; override;
public
fcommand : String;
Constructor Create(isSuspended : boolean);
end;
Var
MemStream :TMemoryStream;
Running:Boolean;
implementation
constructor TProcessThread.Create(isSuspended : boolean);
begin
MemStream := TMemoryStream.Create;
inherited Create(isSuspended);
end;
procedure TProcessThread.Execute;
const
READ_BYTES = 2048;
Var
Process :Tprocess;
n: LongInt;
BytesRead: LongInt;
begin
Process := TProcess.create(nil);;
Process.CommandLine := Fcommand;
Running := True;
{Actually run the thing and catch the output}
BytesRead := 0;
Process.Options := [poUsePipes,poNoConsole];
Process.Execute;
while Process.Running do
begin
// make sure we have room
MemStream.SetSize(BytesRead + READ_BYTES);
// try reading it
n := Process.Output.Read((MemStream.Memory + BytesRead)^, READ_BYTES);
if n 0
then begin
Inc(BytesRead, n);
end
else begin
// no data, wait 100 ms
Sleep(100);
end;
end;
// read last part
repeat
// make sure we have room
MemStream.SetSize(BytesRead + READ_BYTES);
// try reading it
n := Process.Output.Read((MemStream.Memory + BytesRead)^, READ_BYTES);
if n 0
then begin
Inc(BytesRead, n);
end;
until n = 0;
MemStream.SetSize(BytesRead);
Process.Free;
Running := False;
end;
end.
unit RunInThread;
{$mode objfpc}{$H+}
interface
uses
LResources,Classes, SysUtils,ProcessThread;
Type
TNewOutPutEvent = procedure(Sender: TObject;NewOutPut:TStringList) of object;
TRunInThread = Class(TComponent)
private
FOutPut : TStringList;
FNewOutPut :TStringlist;
FProcessThread : TProcessThread;
FCommand : String;
FRunning : Boolean;
FOnNewOuput : TNewOutPutEvent;
FOnDone : TNotifyEvent;
FLastSize : Integer;
FLastMSize : Integer;
procedure GetOutPut;
destructor destroy; override;
public
constructor Create(AOwner: TComponent);
procedure run;
published
property Command : String read FCommand write Fcommand;
property OutPut: TStringLIst read FOutput;
property Running: Boolean read FRunning;
property OnNewOutput : TNewOutPutEvent read FOnNewOuput write FOnNewOuput;
property OnDone : TNotifyEvent read FOnDone Write FOnDone;
end;
procedure Register;
implementation
procedure Register;
begin
RegisterComponents('OLPack',[TRunInThread]);
end;
Constructor TRunInThread.Create(AOwner: TComponent);
Begin
FLastSize := 0;
FLastMSize := 0;
FOutPut := TStringList.Create;
FNewOutPut := TStringList.Create;
FProcessThread := TProcessThread.Create(false);
FProcessThread.FCommand := FCommand;
if Assigned(FProcessThread.FatalException) then
raise FProcessThread.FatalException;
inherited create(AOwner);
end;
Destructor TRunInThread.Destroy;
Begin
FProcessThread.Suspend;
FOutPut.Free;
FNewOutPut.Free;
FProcessThread.Terminate;
inherited destroy;
end;
Procedure TRunInThread.GetOutPut;
Begin
FOutPut.LoadFromStream(MemStream);
end;
procedure TRunInThread.Run;
Var I : Integer;
Begin
FProcessThread.Resume;
FRunning := True;
While Running do
Begin
If MemStream.Size FLastMSize then
begin
FLastMSize := MemStream.Size;
GetOutPut
Re: [lazarus] TRunInThread - Progress, but new problem
On Sunday 05 March 2006 08:43, A.J. Venter wrote: Well since my last mail on this a few weeks ago, I have made quite a bit of progress in my attempts to create a component that will allow me to call a tprocess in a thread while retrieving the output live. The structure I have no compiles, and can be placed on a form - it is all but right. The only trouble is - when I call the run method, it throws an access violation, the stack trace places it at line 76 (the line where I call FProcessThread.resume), I did a check and the first line of the execute method does not ever get executed. Mmmm, I think I know now why this is happening, though now I have even LESS of an idea how to prevent it. I instantiated TProcessThread directly in a test program (obviously this left me with no events or such but I was just trying to see if it worked) when I call TProcessThread.Create the program dies with error 232 complaining that there isn't a thread driver. I checked the lpr, it calls cthreads. I checked lazarus.pp - that too calls cthreads, the only thing I can imagine is that I must somehow call cthreads in olpack, putting it in olpack.lpk doesn't work - lazarus recreates the file on compile/save apparently, putting it in runinthread.pas doesn't work (probably because it's not early enough in the uses clause) - leaving me unable to determine where I SHOULD add it - or is this the IDE telling me that what I am trying to do is simply not possible ? TIA A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] setColor part 3
It's much simpler to implement one custom drawing function for a control, then to map to various themed widgetsets. The LCL took the difficult and more complex approach. Other people can contribute custom drawn controls. I wonder, why there is not already a package providing some custom drawn controls as there are so many for Delphi. Well most of my visual components are actually custom drawn since they are designed to implement stuff that GTK doesn't normally do. Where themes fall short is in going one level up as it were. A quick google will find you at least twenty delphi components for doing more advanced pop-up hints with for example, depending on what you are looking for (everything from bubble-shaped to extensive html with image support is available) - THAT is where custom drawing becomes usefull, for adding that which gives your program an edge. Sure you don't NEED bubble-shaped pop-ups but they are more pleasing to users than square lines, and in fact they are becoming the default in several development environments (on windows anyway). As far as I can tell, there is no way to implement this with a theme - you need some form of custom drawing to do it, and I don't think it can be said that this in any way detracts from usability. Several usability studies have come to the same conclusion, the universal look and feel is a myth - icons and colours only have meaning in the culture from which they originate - in other cultures they could mean completely different things or more commonly nothing at all - the core thing a programmer MUST do for usability is be consistent -always use the SAME file open icon, so users can recognise it even if it DOESN'T mean anything to them. The second most important thing is to ensure that your interface makes the purpose of every element obvious right from the start, the more obvious you can make things the easier it is to get the program. The third rule is - make it look pretty. Rule 1 says - uses themeing almost all the time. Rules 2 and 3 says - override the theme if you have to. Mind you lazarus does not at present provide a clean way of SETTING theme style properties, and I think it will be very hard to add since the nature of this will be platform independent (I highly doubt that the same themeing structure would work on windows that works on GTK and even between GTK1 and GTK2 there are huge differences in the theme files). What would be good is if we can create a set of generic theme-style specifiers for the LCL (much like we translated widgets) which are added as ADDITIONAL properties, not replacing the delphi properties but being added as extras. This way you can set at least those elements which are common to all platforms in the IDE and it will follow a theme file on how to actually implement them, but when there is a good reason, you use the delphi style properties to override the them. A side effect is that if it can also translate to CSS properties it would be a very nice tool one day when CGI-from-the-IDE starts to become visual :) When I say : live its not so simple, I think: nobody know all aspect and environment of every program. One tools never be enough good to use in any case. Sure, but we try. :) Lol, that is why there isn't one programming language to rule them all :) but OP comes pretty close. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] TRunInThread - Progress, but new problem
I checked the lpr, it calls cthreads. I checked lazarus.pp - that too calls cthreads, the only thing I can imagine is that I must somehow call cthreads in olpack, putting it in olpack.lpk doesn't work - lazarus recreates the file on compile/save apparently, putting it in runinthread.pas doesn't work (probably because it's not early enough in the uses clause) - leaving me unable to determine where I SHOULD add it - or is this the IDE telling me that what I am trying to do is simply not possible ? You should add it in the program file (.lpr) of your test program. As the _first_ unit (possibly after cmem, if you use that too). It IS the first unit in the lpr :( Could it be that it somehow just doesn't get compiled intoto the component ? Has anybody ever DONE a threaded component ? A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] button icons for lazarus
Well I wasn't trying to extend the IDE, but to give a tool for USERS of the IDE, specifically to give replacements for the standard icons that delphi shipped with. These are not really meant for use INSIDE the IDE, but are meant to be available to people coding WITH the IDE. For compatibility I replaced specifically the icons that are in delphi (well Kylix OE actually), with the same names, and the same sizes (delphi's icon's are ALL 16x16) That is not completely correct: The component palette icons are 24x24, this is even a requirement which is documented. 23x23 Actually ideintf/formeditingintf.pas defines 25x25. Minus the border of at least 1 gives 23x23. Where is 24x24 documented? Well like I said I was never TRYING to do buttons for use by the IDE or for the component pallette, I was doing icons for use by programs, e.g. for toolbars and menu's - basically to provide a set of instant access default buttons to put in timagelists. Now having said that, there is no reason the same icons cannot be done at ANY size, or based on ANY icon set, but since there wasn't anything at ALL for this, I asked and people said it would be a good idea so I went ahead and pulled an allnighter and did it. Now perhaps what is coming out of this is that we should in fact expand the set and possibly even start a full-scale subproject doing lazarus icons both for use by lazarus programs and for the IDE itself in all the appropriate sizes and styles ? The icons delivered with Delphi are 16x16, but that is because it is essentially still the same set of icons as delivered with Borland Pascal for windows 3.1 While this wouldn't surprize me at all there is more to it, considdering what these icons were typically meant for - 16x16 is STILL the standard size, I'm sitting in front of KDE right now and all the toolbar icons are 16x16 as are all the icons in menu's - because bigger than that and it would dwarve the text that goes next to it. While there may be good reasons for providing larger icons as well - for the goal I was trying to achieve, 16x16 remains the standard size on every desktop I know about and is likely to remain so for a long time yet. GUI and graphics requirements have increased since 3.1, so I would opt for a standard size of 22x22 ? Possibly, though yet again - we are not talking about stuff to put on your Tglyphbuttons here but about things to put on toolbars and inside TMainMenu's - 16x16 as an available option at least must be considered, unless we will code automatic scaling into all the onpaints... Basically what I am trying to do is to complete lazarus a little - by adding some of the things delphi users have come to expect, like a readily available set of standard icons they can assign to typical tasks like File|Open. Ciao A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] button icons for lazarus
On Sunday 05 March 2006 14:13, Darius Blaszijk wrote: We should have fixed sizes (16x16, 24x24, etc.) not scaled sizes which are looking ugly. It makes no sense to use more than two sizes inside an app. 16x16 - menuitems, treeviewitems, listviewitems, small toolbars buttons 22x22 or 24x24 - large toolbar buttons I'm with you on that one, but I spent about 5 hours doing these (quite a lot considering most of them were just conversions) the hard part was finding all the names, so I am not doing more today, but I reckon adding a 22x22 set should be faster. I want to do it the right way though - recreating them from the original SVG scaled images, not scaling the existing xpm's which will just make them ugly. A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: {SPAM?} Re: [lazarus] System Information
On Sunday 05 March 2006 22:29, Panagiotis Sidiropoulos wrote: I found Linux command hdparm at http://www.webmasterworld.com/forum40/957.htm. This command provides a part of information I need, as follows. Is it possible to retrieve this information from inside a Lazarus app but without root previledges? You should be able to get almost all of it from /proc normal users have read-only privileges there, it's basically a case of reading files to get the information you want. A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Using MySQL
On Monday 06 March 2006 13:15, Andreas Berger wrote: A few years back I went to a trade show where I met a representative of MySQL. Since MySQL is not free for comercial applications I asked him if I could use Delphi (at the time) that supports MySQL to write an app that uses MySQL and distribute it. He answered: only if you supply the souce to the app as well. I then asked if I could distribute the app and let the client download and install MySQL seperatly. He said no since the component used by Delphi was GPL and therefore forced me to distribute my sourcecode anyway. My question: does anyone know how this issue stands with Lazarus? Same rules, lazarus is lgpl but the mysql libraries it uses are GPL´d - you don´t need anything from the lazarus guys to create a closed source mysql based app - but you are linking against a GPL´d library - and that requires the permission of the copyright holder - e.g. mysql inc. So to put it another way, the catch is with libmysqlclient.so which is gpl´d and that means any app linked against it MUST be gpl´d as well, you need a commercial license from mysql if you want to write a non-gpl´d program using it. The lazarus mysql components are NOT affected since they do not link against the library - they merely translate the function names and then use an external call to implement them - your program which uses that unit/components however DOES get linked against it. Only question is whether this means the lazarus IDE gets linked against libmysqlclient when the mysql support for the IDE is compiled. Even then at worst we may have to say the IDE is gpld and the LCL is lgpld - it can´t have any wider impact than that and I don´t see how that can affect anything, after all nobody is building a proprietory derivative of the IDE - we are building apps with the LCL. Ciao A.J. -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Executing another application from within a lazarus app
On Tuesday 07 March 2006 14:49, Sam Washkansky wrote:
Thanks a lot I would appreciate that.
Cheers
Sam
Here you go, I´m not going to post the whole huge unit now, just the two
procedures, you can use them as you see fit, they are pretty much just
adaptations of the examples in the tutorial anyway (note that they are
overloaded):
procedure execute (command:string; var output:Tstrings);
{Run command and store the output from it in output
Notes:
1) do not create output prior to assigning it
2) you must free output yourself
}
const
READ_BYTES = 2048;
Var
Process :Tprocess;
MemStream : TMemoryStream;
n: LongInt;
BytesRead: LongInt;
begin
Process := TProcess.create(nil);;
Process.CommandLine := command;
{Actually run the thing and catch the output}
MemStream := TMemoryStream.Create;
outPut := TStringList.Create;
BytesRead := 0;
Process.Options := [poUsePipes,poNoConsole];
Process.Execute;
while Process.Running do
begin
// make sure we have room
MemStream.SetSize(BytesRead + READ_BYTES);
// try reading it
n := Process.Output.Read((MemStream.Memory + BytesRead)^, READ_BYTES);
if n 0
then begin
Inc(BytesRead, n);
end
else begin
// no data, wait 100 ms
Sleep(100);
end;
end;
// read last part
repeat
// make sure we have room
MemStream.SetSize(BytesRead + READ_BYTES);
// try reading it
n := Process.Output.Read((MemStream.Memory + BytesRead)^, READ_BYTES);
if n 0
then begin
Inc(BytesRead, n);
end;
until n = 0;
MemStream.SetSize(BytesRead);
OutPut.LoadFromStream(MemStream);
MemStream.Free;
Process.Free;
end;
procedure execute (command:string);
{Execute command, do not wait for it to exit, do
not capture output.}
Var
Process :Tprocess;
begin
Process := TProcess.create(nil);;
Process.CommandLine := command;
Process.Options := [poNoConsole];
Process.Execute;
Process.Free;
end;
--
80% Of a hardware engineer's job is application of the uncertainty principle.
80% of a software engineer's job is pretending this isn't so.
A.J. Venter
Chief Software Architect
OpenLab International
http://www.getopenlab.com | +27 82 726 5103 (South Africa)
http://www.silentcoder.co.za| +55 118 162 2079 (Brazil)
_
To unsubscribe: mail [EMAIL PROTECTED] with
unsubscribe as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
Re: Mozilla Plugin Panel component
I will write some docs regarding what is working and what isn't. I will definitely try to help with filling in some of the linux bits at least. That would be much appreciated! What part of the component are you interested in using? The Adobe Reader or the Flash Player? Well being mozilla, presumably there is a web-browsing option ? :p There is a delphi component somewhere that maps to gecko, I have no idea if it would work under linux/lazarus though. GTKSocket and GTKPlug is the way to handle window embedding on linux however, and so it seems that my generic embedding components (which are only a design so far, haven't had time to write them) may well be of use if tied in with yours ? A.J. -- there's nothing as inspirational for a hacker as a cat obscuring a bug by sitting in front of the monitor - Boudewijn Rempt A.J. Venter Chief Software Architect OpenLab International www.getopenlab.com www.silentcoder.co.za +27 82 726 5103 _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
Re: [lazarus] Lazarus advocacy : Best App wrtitten in Lazarus
I like this idea, and it ties in with an idea I had, I would like to do a series of ¨people behind lazarus¨ interviews, similiar to the ¨people behind KDE¨ series from kde dot news, by itself though it wouldn´t work but perhaps combined with a site like this it would be a nice section ? If others like it, I am happy to handle this part - and if Mattias doesn´t mind I would like to start with him :) Ciao A.J. On Tuesday 14 March 2006 19:11, Alexsander Rosa wrote: Maybe somethink like spreadfirefox.com: uselazarus.com or something like that, with a CMS inplace so the mantainers could easily edit/add pages. Alex 2006/3/14, Alexander Todorov [EMAIL PROTECTED]: On 3/14/06, Vincent Snijders [EMAIL PROTECTED] wrote: I think it would be good to have such a page. I wanted to start it here: http://wiki.lazarus.freepascal.org/index.php/Lazarus_Application_Galler y A (not-complete) list of candidates can be found here: http://wiki.lazarus.freepascal.org/index.php/Projects_using_Lazarus I was inspired by: http://www.freepascal.org/gallery.html Some of those are lcl applications or written with Lazarus, so they could be added. Vincent. IMO the page must be more glamorous than a simple wiki page. Maybe a separate page. With a lot of screen shots and description of the used programming techniques, classes, DB servers, etc ... All stuff that will make you use Lazarus / FPC for development. The FPC page is a nice start. _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives -- Atenciosamente, Alexsander da Rosa -- 80% Of a hardware engineer's job is application of the uncertainty principle. 80% of a software engineer's job is pretending this isn't so. A.J. Venter Chief Software Architect OpenLab International http://www.getopenlab.com | +27 82 726 5103 (South Africa) http://www.silentcoder.co.za| +55 118 162 2079 (Brazil) _ To unsubscribe: mail [EMAIL PROTECTED] with unsubscribe as the Subject archives at http://www.lazarus.freepascal.org/mailarchives
