Re: Qt2 filenames

[Kalle Dalheimer]

On Thursday 14 June 2001 02:36, John Levon wrote:
 On Wed, Jun 13, 2001 at 10:06:24AM +0200, Edwin Leuven wrote:
   FormXXX
   FormXXXDialog - machine generated
   FormXXXDialogImpl
 
  I still hold the opinion that the Form part is redundant. Why not:
   XXX
   XXXDialog - machine generated
   XXXDialogImpl
 
  On the other 2 points I am agnostic...
 
  gr.ed.

 ok, good point. As is Allan's. How about :

 QXXX.[Ch]
 ui/QXXXDialog.[Ch]
 QXXXDialogImpl.[Ch]

 Kalle, does this address your ls concern ? Now QXXX and QXXXDialogImpl
 will be next to each other. As stuff in ui/ isn't human editable, it makes
 sense to keep it out of the way.

At least, I like the idea of dropping the Form prefix a lot (I never 
understood why it was there, probably an xforms thing). It only addresses my 
ls concern partly, as I would prefer to have the .ui files inlined as well 
- whether it is human editable is not so important (at least for me), rather 
whether it needs editing.

Oh, now I see that you want to put the Designer-generated C and h files into 
ui/. Yes, that would be good, especially if the .ui files stay out of there. 
Could be a bit confusing for the casual observer that the .ui files are not 
in the ui/ directory. Perhaps the directory should be called uic/, then.

Kalle

-- 
Matthias Kalle Dalheimer
President  CEO/VD
Klarälvdalens Datakonsult AB
Fax +46-563-540028
Email [EMAIL PROTECTED]

Re: [PATCH] update INSTALL

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| On Wed, Jun 13, 2001 at 04:06:09PM +0200, Jean-Marc Lasgouttes wrote:
| 
|   John == John Levon [EMAIL PROTECTED] writes:
|  
|  John self-documenting
|  
|  I applied it, but I am not sure that GNU m4 (or any m4) is really
|  needed if people use a tarball of LyX... This is probably a bug.
| 
| is this a bug to do with dependencies and timestamps, or a requirement bug
| with the sigc++ stuff ?

Most likely a bug with the dist target.

-- 
Lgb

Re: [PATCH] paste bug (again)

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| On Wed, Jun 13, 2001 at 09:04:33AM +0200, Lars Gullik Bjønnes wrote:
| 
|  Actually I would prefere to see the addition of a String that takes a
|  ostream..., alternatively:
| 
| ho-hum, you were free to mention this after not liking the last attempt
| 
| no problem, I'll have another go.

Hey, I can't think of everything at once can I?

_and_ you are allowed to disagree with me.

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[Jean-Marc Lasgouttes]

 John == John Levon [EMAIL PROTECTED] writes:

John On Thu, Jun 14, 2001 at 01:46:06AM +0100, John Levon wrote:
  Or the const char * _() could return a string. Or it could be
 removed  altogether...
 
 removed sounds good. I'll have a go building with this change.

John dur, of course we use the char * one everywhere.

John And cgrepall '(.*).*+' alone has 69 matches. Argh.

But the char const * should be promoted to string automatically, no?

JMarc

Re: xforms help needed !

[Jean-Marc Lasgouttes]

 Allan == Allan Rae [EMAIL PROTECTED] writes:

Allan This could be another of those problems with nested tabfolders
Allan -- much like the problem with setting shortcuts for them.
  What about using something else? Like a list on the left instead
 of the first-level tabs?

Allan Sure. All we need is time.

Fine.

 And what about the patch you posted earlier for shortcut? Is it
 worth applying?

Allan Only once we've gotten rid of the nested tabfolders or xforms
Allan is fixed so they work. I thinnk I mentioned then that I think
Allan we _might_ be able to make them work by overriding or at least
Allan supplying callbacks for the shortcut handlers in the inner
Allan forms -- tabfolders have a large number of callbacks available
Allan that are _NOT_ listed in the documentation but which are public
Allan in the headers. Anyway, that becomes black magic and I'm not
Allan inclinded to explore that path very far.

So the technique could be applied to the paragraph and document
popups? Some people are very eager to see this done...

JMarc

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

Jean-Marc Lasgouttes [EMAIL PROTECTED] writes:

|  John == John Levon [EMAIL PROTECTED] writes:
| 
| John On Thu, Jun 14, 2001 at 01:46:06AM +0100, John Levon wrote:
|   Or the const char * _() could return a string. Or it could be
|  removed  altogether...
|  
|  removed sounds good. I'll have a go building with this change.
| 
| John dur, of course we use the char * one everywhere.
| 
| John And cgrepall '(.*).*+' alone has 69 matches. Argh.
| 
| But the char const * should be promoted to string automatically, no?

Please look into rewriting the code that has problems to use
stringstream, that will actually be better in the long run than to
fiddle with the _() macros.

Just list the code-lines and I'll do the change...

As mentioned earlier, stringstreams often gives smaller code. Most
likely due to a smaller number of temporaries.

string a = string(hello) + string(world) + string(!);

How many temporaries? At least five.

with string stream this could have been:

ostringstream o;
o  hello  world  !;
string a(o.str());

How many temporaries? one?

Also the ostream can often be passed in, and sometimes is passed in
anyway, so we end up with having only one ostream object shared
between several methods. (just look at the write methods)

-- 
Lgb

Re: [Bug] Accents in math mode

[Andre Poenitz]

  No. It should get fixed during the summer.
 
 Yes, but _which_ summer?

The summer of the year 2742 of course. Sorry for not telling the whole truth.

Andre'

-- 
André Pönitz . [EMAIL PROTECTED]

Re: New bug list

[Juergen Vigna]

On 14-Jun-2001 John Levon wrote:

 No, I don't have a really good solution (yet). One dumb idea is separate Apply
 and Apply to current paragraph buttons. Although it's not the best solution prolly,
 it does indicate in a concise way the problem I see currently.

I don't get what the two apply button should buy us? The way I see it we would
need a Update button which updates the Dialog with the underlying paragraph
data.

Where we disagree is in the automatic update of the Dialog when you change
to another paragraph while the Dialog is open. Obviously it will always open
with the values of the actual paragraph when you open it.

Well it should be easy to make this behaviour configurable (automatic update
or not) and the default has to be the old behaviour, obviously ;)

   Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

The trouble with being poor is that it takes up all your time.

layout/tabular

[Stephan Mietens]

Hi,

opening the tabular dialog box with the row/column tab gives you
two meanings of alt-c: either this means Haligment center, or 
close. That should be non-ambiguous.

Greets,

Stephan

Re: importing tables into lyx

[j.heidemeier]

Thanks I looked in it an I presume it will work.
Well if LyX would embedd an Tcl Interpreter all the Import Export Stuff
could be handled very smoothly.
BTW do you have the description of the new format which looks as if 
inspired by XML ? It would be worth trying to support this, too.


 -- 
Dr. Joachim Heidemeier  Tel. +49-30-8903-2780 
Fachgebiet II 3.2   [EMAIL PROTECTED] 
Umweltbundesamt, Bismarckplatz 1 D-14191 Berlin

On Wed, 13 Jun 2001, Martin Vermeer wrote:

 Date: Wed, 13 Jun 2001 19:32:25 +0300
 From: Martin Vermeer [EMAIL PROTECTED]
 To: [EMAIL PROTECTED]
 Cc: [EMAIL PROTECTED]
 Subject: Re: importing tables into lyx
 
 On Wed, Jun 13, 2001 at 07:35:22AM +0200, joachim heidemeier wrote:
  Mailing-List: contact [EMAIL PROTECTED]; run by ezmlm
  Precedence: bulk
  X-No-Archive: yes
  List-Post: mailto:[EMAIL PROTECTED]
  List-Help: mailto:[EMAIL PROTECTED]
  List-Unsubscribe: 
mailto:[EMAIL PROTECTED]
  Delivered-To: mailing list [EMAIL PROTECTED]
  X-Envelope-From: [EMAIL PROTECTED]
  X-Envelope-To: [EMAIL PROTECTED]
  Date: Wed, 13 Jun 2001 07:35:22 +0200
  From: joachim heidemeier [EMAIL PROTECTED]
  Organization: private-uucp-site
  X-Mailer: Mozilla 4.76 [de] (X11; U; Linux 2.2.18 i586)
  X-Accept-Language: en
  To: [EMAIL PROTECTED]
  Subject: importing tables into lyx
  
  dear lyx developers,
  as I urgently need importing RTF-documents into Lyx I have the problem
  that I'm stuck with table imports (C.f the Thread rereLyX and import of
  MS-Word resp. RFT documents from April 2001).
  I now would like to try the route:
  Table copy from Word-Doc to Exel - csv Export - transformation form
  csv to a simple lyx-file with reasonable default values - import to the
  mail document without the tables.
  For the transformation I would write a tcl script.
  For this purpose I need a description of the table (or longtable)
  coding  in lyx and a parameter list affecting the tables.
  The idea is not to do a very smart copying and preserving all attribus
  but to read a simple table in the document which is then refined within
  lyx.
  If others are interested in the script I'll publish it of course.
  Please direct your answers to my email-account at work
  [EMAIL PROTECTED]
  Thanks in advance.
  -- 
  joachim heidemeier ([EMAIL PROTECTED])
 
 Also have a look at tk#.tcl to be found at 
http://www.netby.dk/Oest/Europa-Alle/vermeer/.
 
 It only handles the old table format. 
 
 Martin
 -- 
 Martin Vermeer  [EMAIL PROTECTED]
 Helsinki University of Technology 
 Department of Surveying
 P.O. Box 1200, FIN-02015 HUT, Finland
 :wq
 

Q: tabular

[Stephan Mietens]

Hi,

a short question: is there a way to define a tabular such that
lyx generates tex code looking like this?:

\begin{tabular}{|c||c|}
  ^^
Greets,

Stephan

Q: tabular2

[Stephan Mietens]

Hi again,

 a short question: is there a way to define a tabular such that
 lyx generates tex code looking like this?:

 \begin{tabular}{|c||c|}
   ^^

ok, I got the answer right after sending the email, sorry.
But FYI importing tex with such a table will ignore the ||.

Greets,

Stephan

RE: Q: tabular

[Juergen Vigna]

On 14-Jun-2001 Stephan Mietens wrote:
 a short question: is there a way to define a tabular such that
 lyx generates tex code looking like this?:
 
 \begin{tabular}{|c||c|}

Define the right border-line for the left column and the right border
line for the right column. But this has always been like this and you'll
see it also on screen.

  Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

Reality is just a convenient measure of complexity.
-- Alvy Ray Smith

Re: importing tables into lyx

[Juergen Vigna]

On 14-Jun-2001 j.heidemeier wrote:
 Thanks I looked in it an I presume it will work.
 Well if LyX would embedd an Tcl Interpreter all the Import Export Stuff
 could be handled very smoothly.
 BTW do you have the description of the new format which looks as if 
 inspired by XML ? It would be worth trying to support this, too.

Well no we(I) didn't write a format description, but it should be really
easy to figure out what it does, shouldn't it?

   Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

Of all the tyrannies that affect mankind, tyranny in religion is the worst.
- Thomas Paine

Re: Single Quotes Curl

[Jean-Marc Lasgouttes]

 Stan == Stan Gatchel [EMAIL PROTECTED] writes:

Stan Pairs of double quotes curl in different directions but single
Stan quotes don't. It doesn't look correct.

Double quotes are special quote inset which have intelligence in them.
The single quote is plain old quote. To get 'intelligent' single
quotes, you have to go to LayoutDocumentsQuotes (or LayoutQuotes on
1.1.5) and chnage to single quotes.

Stan Your bug number 19 seems to indicate the problem (I think):

No, this is a different problem (which is fixed in 1.1.6fix2).

JMarc

Re: www-user with CSS

[Jean-Marc Lasgouttes]

 Michael == Michael Koziarski [EMAIL PROTECTED] writes:

Michael Netscape 4.x? I guess the choices are either

Michael 1) get the php to turn off css for your browser 2) break the
Michael css.

Michael I'd prefer 1. We could maintain a list of taglines of broken
Michael browsers and have the css turned off by default
Michael (index.php3?nocss=1). With an option for them to turn it back
Michael on again if they feel the need. What do others think?

If you say there is no clean solution to this in netscape, so be it. I
can live with the nocss version, but it really does not look the
same...

JMArc

Re: LyX 1.1.6fix3 and namespaces, take two

[Jean-Marc Lasgouttes]

 Yves == Yves Bastide [EMAIL PROTECTED] writes:

Yves Here is a second patch for building 1.1.6fix with gcc 3.0. This
Yves one should be less bad :)

I applied it. Thanks.

JMarc

Re: layout/tabular

[Jean-Marc Lasgouttes]

 Stephan == Stephan Mietens [EMAIL PROTECTED] writes:

Stephan Hi, opening the tabular dialog box with the row/column tab
Stephan gives you two meanings of alt-c: either this means
Stephan Haligment center, or close. That should be non-ambiguous.

This is somewhat fixed in 1.2.0cvs, where Close does not have a
shortcut at all. I'm not sure it is the best solution, though (and
what the motivation was).

JMarc

Re: layout/tabular

[Juergen Vigna]

On 14-Jun-2001 Jean-Marc Lasgouttes wrote:

 This is somewhat fixed in 1.2.0cvs, where Close does not have a
 shortcut at all. I'm not sure it is the best solution, though (and
 what the motivation was).

Because the Dialogs can all be closed with ESC, so we really don't need
the C shortcut for the Close button!

   Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

The finest eloquence is that which gets things done.

Re: [PATCH] paste bug (again)

[John Levon]

On Thu, Jun 14, 2001 at 09:00:49AM +0200, Lars Gullik Bjønnes wrote:

 John Levon [EMAIL PROTECTED] writes:
 
 | On Wed, Jun 13, 2001 at 09:04:33AM +0200, Lars Gullik Bjønnes wrote:
 | 
 |  Actually I would prefere to see the addition of a String that takes a
 |  ostream..., alternatively:
 | 
 | ho-hum, you were free to mention this after not liking the last attempt
 | 
 | no problem, I'll have another go.
 
 Hey, I can't think of everything at once can I?

:)

 _and_ you are allowed to disagree with me.

not when you're probably right ;)

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: Qt2 filenames

[John Levon]

On Thu, Jun 14, 2001 at 08:33:54AM +0200, Kalle Dalheimer wrote:

 Oh, now I see that you want to put the Designer-generated C and h files into 
 ui/. Yes, that would be good, especially if the .ui files stay out of there. 
 Could be a bit confusing for the casual observer that the .ui files are not 
 in the ui/ directory. Perhaps the directory should be called uic/, then.

actually I would prefer the .ui files to go in there too.

However your suggestion seems a reasonable compromise. So we are all agreed on :

QTabCreate.[Ch]
QTabCreate.ui
QTabCReateDialogImpl.[Ch]
uic/QTabCreateDialog.[Ch]

yes ?

thanks,
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: New bug list

[John Levon]

On Thu, Jun 14, 2001 at 10:47:03AM +0200, Juergen Vigna wrote:

 
 On 14-Jun-2001 John Levon wrote:
 
  No, I don't have a really good solution (yet). One dumb idea is separate Apply
  and Apply to current paragraph buttons. Although it's not the best solution 
prolly,
  it does indicate in a concise way the problem I see currently.
 
 I don't get what the two apply button should buy us? The way I see it we would
 need a Update button which updates the Dialog with the underlying paragraph
 data.

A naive user has the reasonable expectation that a viewport onto an object's 
parameters
will change only that object's parameters, rather than some other unrelated object's 
parameters.

You see the dialog as a current paragraph parameters dialog, but this is clearly
inconsistent with its current behaviour, as you point out below.

Hopefully this is clearer :/

 Where we disagree is in the automatic update of the Dialog when you change
 to another paragraph while the Dialog is open. Obviously it will always open
 with the values of the actual paragraph when you open it.
 
 Well it should be easy to make this behaviour configurable (automatic update
 or not) and the default has to be the old behaviour, obviously ;)

configuration is the last refuge of bad interface design. hopefully we can avoid
this. I don't know how (you are right, automatic update ruins what you want to do
and is obviously not a good solution)

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 10:24:12AM +0200, Jean-Marc Lasgouttes wrote:

 John And cgrepall '(.*).*+' alone has 69 matches. Argh.
 
 But the char const * should be promoted to string automatically, no?

hmmm... then it is a gcc 3.0 bug ???

I don't know :/

if so, do we deal with it, or do we just add don't use gcc 3.0 in INSTALL ?

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 10:34:54AM +0200, Lars Gullik Bjønnes wrote:

 Please look into rewriting the code that has problems to use
 stringstream, that will actually be better in the long run than to
 fiddle with the _() macros.
 
 Just list the code-lines and I'll do the change...

yes, ok.

if this is a compiler bug, it's a fair enough workaround. though if not,
I would prefer to have a special C_() macro for using c-style strings ...

the core lyx code will always use string _(string) after all.

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| On Thu, Jun 14, 2001 at 10:24:12AM +0200, Jean-Marc Lasgouttes wrote:
| 
|  John And cgrepall '(.*).*+' alone has 69 matches. Argh.
|  
|  But the char const * should be promoted to string automatically, no?
| 
| hmmm... then it is a gcc 3.0 bug ???

can some one refresh my memory... what errors are we seening? On the
1.1.6 branch only right?

| I don't know :/
| 
| if so, do we deal with it, or do we just add don't use gcc 3.0 in
| INSTALL ?

that is not an option.

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 06:13:02PM +0200, Lars Gullik Bjønnes wrote:

 John Levon [EMAIL PROTECTED] writes:
 
 | On Thu, Jun 14, 2001 at 10:24:12AM +0200, Jean-Marc Lasgouttes wrote:
 | 
 |  John And cgrepall '(.*).*+' alone has 69 matches. Argh.
 |  
 |  But the char const * should be promoted to string automatically, no?
 | 
 | hmmm... then it is a gcc 3.0 bug ???
 
 can some one refresh my memory... what errors are we seening? On the
 1.1.6 branch only right?

nope, in current CVS. It is corrupting the -message() we send to the minibuffer.

One example is Saving document in lyxfunc.C - it is a garbled string being created.

http://sourceforge.net/tracker/index.php?func=detailaid=429678group_id=15212atid=115212

this is with ENABLE_NLS

 | if so, do we deal with it, or do we just add don't use gcc 3.0 in
 | INSTALL ?
 
 that is not an option.

good !

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| nope, in current CVS. It is corrupting the -message() we send to
| the minibuffer. 
| 
| One example is Saving document in lyxfunc.C - it is a garbled
| string being created. 

Og in what order is A + B + C + D evaluated?
If it is from left to right...

const char* + char  this is not a legal binding...
but if we change the char to int...
const char* + int then we just create a new pointer from the ints
value position...

So... this code is legal, but bogus. This is a bug.

Fix is still to change to stringstream:

ostringstream s1;
s1  _(Saving document)  ' ' 
MakeDisplayPath(owner-buffer()-fileName()  ...;
owner-message(s1.str().c_str());


Other locations please...

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 06:36:14PM +0200, Lars Gullik Bjønnes wrote:

 So... this code is legal, but bogus. This is a bug.

ok.

 Other locations please...

I will collect them this evening. There are tens of these, by the way.

What fun I'll have :)

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

|  Other locations please...
| 
| I will collect them this evening. There are tens of these, by the way.
| 
| What fun I'll have :)

I am tag-searching through the sources right now and fixing as I go
along. So you might want to wait a bit.

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| I will collect them this evening. There are tens of these, by the way.

I think I have gotten most of the dangerous ones now, give me time to
compile and test + commit.

Then tell me if you see the same. 

-- 
Lgb

Searching for code which handles string like char const *

[Jean-Marc Lasgouttes]

The bug I fixed for spellchecker yesterday was because of some code
which looked for character '\0' to detect end of line. I tried to see
whether I could find more.

The typical think I want to catch is
for (int i = 0; str[i]; ++i)
CursorRight(bview);
or variation around the same stupid pattern.

The best grep pattern I found for now is
  grep '\(if\|while\|for\) *(.*[a-z]*\[[^]]\] *\(.*\)*)' {,*/,*/*/}*.C
which is run in src/. As far as I see, the only interesting hit is
  kbsequence.C:   if (s[i]  (s[i]) = ' ') ++i;
which is bad code (as all of kb_sequence::parse,) but not dangerous.

Of course, this regexp returns too many false positives, and I am
wondering whether it sees all cases. So if some creative regexp hacker
has ideas... 

JMarc

Re: www-user with CSS

[Zvezdan Petkovic]

On Thu, Jun 14, 2001 at 05:05:46PM +0200, Jean-Marc Lasgouttes wrote:
  Michael == Michael Koziarski [EMAIL PROTECTED] writes:
 
 Michael Netscape 4.x? I guess the choices are either
 
 Michael 1) get the php to turn off css for your browser 2) break the
 Michael css.
 
 Michael I'd prefer 1. We could maintain a list of taglines of broken
 Michael browsers and have the css turned off by default
 Michael (index.php3?nocss=1). With an option for them to turn it back
 Michael on again if they feel the need. What do others think?
 
 If you say there is no clean solution to this in netscape, so be it. I
 can live with the nocss version, but it really does not look the
 same...
 
 JMArc

First let me admit that I didn't follow the complete correspondence but
I do not quite get the problem.
 
There _IS_ a way to make a clean CSS for Netscape 4.7x. I've spent a lot
of time to do some tricks but succeeded and my CSS and XHTML _validated_
clean.
  
Take a look at my Web page and corresponding css files under
URL/style/personal.css and URL/style/pictures.css.  Regarding the issue
of black letters you mention, I use in pictures.css:

/*
 * personal.css must be loaded before this file
 */
body, p, ul, ol, dl, address, blockquote, td, caption,
h1, h2, h3, h4, h5, h6, th,
tt, pre, code, kbd, samp {
color: white;
background-color: rgb(53,99,144)
}
   
I admit it's overkill, but it validates as clean CSS and does the job.

In personal.css there are some tricks to make a boxed paragraph with a
different background, paragraph with a border, paragraph with a red line
to the left for notices, small caps, drop caps, transparent images
looking really transparent... Unfortunately, Konqueror doesn't get the
trick for small caps and drop caps yet. But it looks fine in Netscape
and IE.
 
I believe that you deliberately refrain from such tricks but they work
and they _are_ clean CSS.
 
If I'm missing the point please forgive me. :-)

-- 
Zvezdan Petkovic [EMAIL PROTECTED]
http://www.cs.wm.edu/~zvezdan/

Re: Searching for code which handles string like char const *

[Lars Gullik Bjønnes]

Jean-Marc Lasgouttes [EMAIL PROTECTED] writes:

| The bug I fixed for spellchecker yesterday was because of some code
| which looked for character '\0' to detect end of line. I tried to see
| whether I could find more.

I don't think we have a lot of them.
 
| The typical think I want to catch is
| for (int i = 0; str[i]; ++i)
| CursorRight(bview);
| or variation around the same stupid pattern.

mmm...yes...  
| The best grep pattern I found for now is
|   grep '\(if\|while\|for\) *(.*[a-z]*\[[^]]\] *\(.*\)*)' {,*/,*/*/}*.C
| which is run in src/. As far as I see, the only interesting hit is
|   kbsequence.C:   if (s[i]  (s[i]) = ' ') ++i;
| which is bad code (as all of kb_sequence::parse,) but not dangerous.
| 
| Of course, this regexp returns too many false positives, and I am
| wondering whether it sees all cases. So if some creative regexp hacker
| has ideas... 

I tried with a simpler wone:

 \[.\].*\+\+.

-- 
Lgb

Re: importing tables into lyx

[Joachim Heidemeier]

I think it doesn't solve the main problem that reLyX didn't handle
longtables.
I'm working on a tcl-package which supports the new tabular-format.
joachim
-- 
joachim heidemeier ([EMAIL PROTECTED])


On Wed, 13 Jun 2001, Herbert Voss wrote:


 On Wed, 13 Jun 2001, joachim heidemeier wrote:

  as I urgently need importing RTF-documents into Lyx I have the problem
  that I'm stuck with table imports (C.f the Thread rereLyX and import of
  MS-Word resp. RFT documents from April 2001).
  I now would like to try the route:
  Table copy from Word-Doc to Exel - csv Export - transformation form
  csv to a simple lyx-file with reasonable default values - import to the
  mail document without the tables.

 have you ever tried the package excel2latex, available
 at CTAN?

 Herbert

Re: www-user with CSS

[A.]

 I believe that you deliberately refrain from such tricks but they work
 and they _are_ clean CSS.
  
 If I'm missing the point please forgive me. :-)

No you're not missing the point :)

As I said when I started this work I am quite millitant on the issue of
CSS compliance and Netscape4 should be considered broken and have CSS
turned off.  I realise that if we added


td, p, li, div, span, table, h1, h2, h3, h4 h5 . 
{ color: rbg(238,238,238) }

It will work *ok* in netscape 4 but I don't see the point.  I guess I
could be convinced that as a *temporary* measure the changes could be
put in.  However I'd be inclined to remove them at the moment it becomes
difficult to maintain.

I've put the changes in at http://www.koziarski.org/LyX/www-user/ and
www-devel but they've introduced other problems (ignoring requests for
sans-serif fonts etc.)

Does that look better to everyone else?

-- 

Michael Koziarski
Software Developer
Compudigm International Ltd.

perl -e print(pack('h45','d69636861656c604b6f6a796162737b696e236f6d6'))

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 07:09:11PM +0200, Lars Gullik Bjønnes wrote:

 John Levon [EMAIL PROTECTED] writes:
 
 | I will collect them this evening. There are tens of these, by the way.
 
 I think I have gotten most of the dangerous ones now, give me time to
 compile and test + commit.
 
 Then tell me if you see the same. 

certainly the bug I noticed is fixed. I also couldn't find any others
you hadn't dealt with.

thanks
john


-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

[PATCH] slight cleanup in bv_pimpl

[John Levon]

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard


Index: ChangeLog
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/ChangeLog,v
retrieving revision 1.193
diff -u -r1.193 ChangeLog
--- ChangeLog   2001/06/14 17:58:48 1.193
+++ ChangeLog   2001/06/15 00:22:05
@@ -1,3 +1,8 @@
+2001-06-15  John Levon  [EMAIL PROTECTED]
+
+   * BufferView_pimpl.C: fix typo, remove minibuffer message
+ when buffer has loaded
+
 2001-06-14  Lars Gullik Bjønnes  [EMAIL PROTECTED]
 
* lyxfunc.C (Dispatch): use stringstream
Index: BufferView_pimpl.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/BufferView_pimpl.C,v
retrieving revision 1.118
diff -u -r1.118 BufferView_pimpl.C
--- BufferView_pimpl.C  2001/06/14 17:58:48 1.118
+++ BufferView_pimpl.C  2001/06/15 00:22:10
@@ -347,6 +347,9 @@
setState();
AllowInput(bv_);
 
+   /// clear the Formatting Document message 
+   owner_-message();
+ 
/// get rid of the splash screen if it's not gone already
owner_-getDialogs()-destroySplash();
  
@@ -3199,7 +3202,7 @@
break;
 
case LFUN_UNKNOWN_ACTION:
-   owner_-getLyXFunc()-setErrorMessage(N_(Unknow function!));
+   owner_-getLyXFunc()-setErrorMessage(N_(Unknown function!));
break;

default:

tex-defs.h removal

[John Levon]

it's still waiting to be applied.

also lars I still need write on lyxdoc so I can commit the Intro
patch mike ressler approved.

And I would like write on qt2/ too so I can do the big renaming

thanks
john

-- 
Research is not a warm puppy. At least, I hope not.
- David Rydeheard

Re: www-user with CSS

[Zvezdan Petkovic]

On Fri, Jun 15, 2001 at 12:09:45PM +1200, Michael A. Koziarski wrote:
  I believe that you deliberately refrain from such tricks but they work
  and they _are_ clean CSS.
   
  If I'm missing the point please forgive me. :-)
 
 No you're not missing the point :)
 
 As I said when I started this work I am quite millitant on the issue of
 CSS compliance and Netscape4 should be considered broken and have CSS
 turned off.  I realise that if we added
 
 
 td, p, li, div, span, table, h1, h2, h3, h4 h5 . 
 { color: rbg(238,238,238) }
 
 It will work *ok* in netscape 4 but I don't see the point.  I guess I
 could be convinced that as a *temporary* measure the changes could be
 put in.  However I'd be inclined to remove them at the moment it becomes
 difficult to maintain.
 

I absolutely agree with you. 

However, there's a way to improve the maintainability of CSS. Get rid of
redundancy! Below are a few examples.

But first the remaining issues. This time I looked carefully in your CSS
and XHTML.  I noticed the following:

a) You do not use XHTML correctly in some places and hence the problem
with black letters at the bottom of the page. The thing between two
hr's should be inside p.../p. If you used XHTML strict instead of
transitional the validator would've reported the issue. 

b) The problem of non-serif bulleted items is related to the previous
colour problem and Netscape obtuseness. Read more below and notice use of
ul, ol, and dl below on the line where I define the fonts.

c) I'm said to say this, but your CSS is an example of _non-user_
friendly design. If you are so militant on the issue of compliance why
do you avoid to comply yourself? You use px and pt sizes all over the
place. The first rule of a good Web design (or any design for that
matter) is: Give a choice to user. They will choose the preferred
size in their browser. CSS designer should use the relative sizes in em
units (width of a small letter m in a given font). See examples below
for headings.

Finally, I'm back to the issue of maintainability.

There are basically two approaches to writing CSS:

1. Put everything that belongs to an element together, as in:

body {

}

   You are using that approach.

   Advantage: everything related to an element is in one place.
   Disadvantage: repetition of things that can be together (e.g. we
   want all headers to be bold, why repeat 6 times?)

   Hence,

2. Put everything common together. This is the approach that I and
   some other people use.

   Disadvantage: You have to use search feature of your editor to
   find all parts related to a single element (not too difficult).

   Advantage: You quickly find and change fonts, colours, weight.
   Once all browsers work well with CSS (and there is currently none
   that supports it completely -- give me one that supports dropped
   capitals?), you can remove elements easily. There is no
   redundancy in the code though.
   _And it works in all browsers!_

   Here's an example:

   /*
*  Fonts
*/
   body, p, ul, ol, dl, address, td, caption {
   font-family: Georgia, serif;
   }

   h1, h2, h3, h4, h5, h6, th {
   font-family: Verdana, sans-serif;
   }

   tt, pre, code, kbd, samp {
   font-family: Andale Mono, monospace;
   }
   /* font sizes are defined below */

   /*
*  Page layout
*/
   body, p, ul, ol, dl {
   text-align: justify
   }

   body {
   margin-left: 10%;
   margin-right: 10%
   }
   
   h1 {
   margin-left: -8%;
   text-align: left;
   font-size: 1.728em;
   font-weight: bold
   }

   h2, h3, h4, h5, h6 {
   margin-left: -4%;
   text-align: left
   }

   h2 {
   font-size: 1.44em;
   font-weight: bold
   }

   h3 {
   font-size: 1.2em;
   font-style: italic;
   font-weight: bold
   }

   h4, h5, h6 {
   font-size: 1em;
   }

   h4 {
   font-weight: bold
   }

   h5, h6 {
   font-weight: normal;
   }

   h5 {
   font-style: italic;
   }

   
   
As you can see there is not a single repeated element above. You do
not need ul any more in font definitions because browsers work? Fine,
take it from that line and everything is OK. The other important
settings for ul still remain somewhere in the file:

   ul.square {
   list-style-type: square
   }

***
To conclude, most browsers are dumb in regard to CSS. We have to provide
precise definitions for all elements instead of relying on inheritance.
To increase maintainability, decrease redundancy. Let's hope that one
day the browsers will understand CSS completely. Till that day we have
to 

Re: www-user with CSS

[A.]

 I absolutely agree with you. 
 
 However, there's a way to improve the maintainability of CSS. Get rid of
 redundancy! Below are a few examples.
 
 But first the remaining issues. This time I looked carefully in your CSS
 and XHTML.  I noticed the following:
 
 a) You do not use XHTML correctly in some places and hence the problem
 with black letters at the bottom of the page. The thing between two
 hr's should be inside p.../p. If you used XHTML strict instead of
 transitional the validator would've reported the issue. 

I'm aware of this problem, the footer has been updated on the main site
anyway so this isn't an issue. (I'm just too lazy to run cvs udpate :))

 b) The problem of non-serif bulleted items is related to the previous
 colour problem and Netscape obtuseness. Read more below and notice use of
 ul, ol, and dl below on the line where I define the fonts.

That's a good idea.  I'll do that on the weekend (stupid paid work is
keeping me busy)

 CSS designer should use the relative sizes in em
 units (width of a small letter m in a given font). See examples below
 for headings.

Yep, I realise this.  the css itself was ripped from 3 sites I wrote in
98.  I intend to use em for all the sizes.  Just a matter of converting
them all to rational values (I'm not much of a designer so my choices
would be ... ugly :))


 Finally, I'm back to the issue of maintainability.
 
 There are basically two approaches to writing CSS:

I'll use method 2 over the weekend.


 ***
 To conclude, most browsers are dumb in regard to CSS. We have to provide
 precise definitions for all elements instead of relying on inheritance.
 To increase maintainability, decrease redundancy. Let's hope that one
 day the browsers will understand CSS completely.

This is where my millitancy comes in :).  The web will *never* leave its
current state of font  ... bgcolor= ...  etc. while people continue
to contort themselves to ensure 'works in [insert broken browser]'.
I'll add the suggestions you mentioned above (they're good and I can see
that they will be maintainable) but any reports of broken rendering with
*that* and I'll just suggest an upgrade or turn off css.


 And both are
 pretty dumb. I remember that IE on Win2000 didn't recognise UTF-8 in the
 first xml line of my strict XHTML doc.

I must admit that's an interesting new bug :)

-- 

Michael Koziarski
Software Developer
Compudigm International Ltd.

perl -e print(pack('h45','d69636861656c604b6f6a796162737b696e236f6d6'))

Re: xforms help needed !

[Allan Rae]

On 14 Jun 2001, Jean-Marc Lasgouttes wrote:

[shortcuts for tabfolders]

 So the technique could be applied to the paragraph and document
 popups? Some people are very eager to see this done...

Definitely can be used now for tabfolders that are not nested.
All you need to do is think of suitable shortcut key combination.
(and then add the code of course ;-)

Allan. (ARRae)

Re: Qt2 filenames

[Allan Rae]

On Thu, 14 Jun 2001, Kalle Dalheimer wrote:

 At least, I like the idea of dropping the Form prefix a lot (I never
 understood why it was there, probably an xforms thing).

The xforms port was always using Form* for our code and form_* for
generated code.  Then other ports started and people refused to entertain
their intellects and just copied the naming convention of the xforms port
even though it was totally irrelevent to the port they were doing.

They even ignored protests about fixing the names used in their port on
the grounds that there was some obtuse logic behind it all (which only
they could see).

Finally an injection of new blood and some small flames later sanity
returns to file naming conventions.

Allan. (ARRae)  wipes brow; steps down off soapbox; exits stage right.

Re: [PATCH] ignore empty minibuffer lines

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| If a user enters a empty minibuffer input, it should be ignored IMHO

Ok, this is good.

_but_ should this be done in the minibuffer code? IMHO the minibuffer
should not define policy that should be done by lyxfunc... so the
check for empty command should not be done in minibuffer, but in the
method run by the emit.

-- 
Lgb

Re: [PATCH] slight cleanup in bv_pimpl

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| thanks
| john

Ok, do this.

-- 
Lgb

Re: [PATCH] help out the internet !

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| ... but maybe put the HD manufacturers out of business

As long as we don't remove it from CVS.

What do others say?

-- 
Lgb

Re: tex-defs.h removal

[Lars Gullik Bjønnes]

John Levon [EMAIL PROTECTED] writes:

| it's still waiting to be applied.
| 
| also lars I still need write on lyxdoc so I can commit the Intro
| patch mike ressler approved.
| 
| And I would like write on qt2/ too so I can do the big renaming

Ok, done.

-- 
Lgb

Re: [PATCH] help out the internet !

[Allan Rae]

On 15 Jun 2001, Lars Gullik Bjønnes wrote:

 John Levon [EMAIL PROTECTED] writes:

 | ... but maybe put the HD manufacturers out of business

 As long as we don't remove it from CVS.

 What do others say?

Ooooh,  I don't know.  It definitely needs to stay in CVS.
It would be nice to trim the distro a little if possible and
this doesn't really affect the distro in any way.

So, sure, why not, lets purge a few more things.  Maybe even take out
ChangeLog.1 since that is also large and isn't really going to interest
users -- anyone serious can get CVS'd.

There must be other stuff we can cull too.  I'm sure I had a list here
somewhere...

Allan. (ARRae)

Re: Qt2 filenames

[Kalle Dalheimer]

On Thursday 14 June 2001 02:36, John Levon wrote:
> On Wed, Jun 13, 2001 at 10:06:24AM +0200, Edwin Leuven wrote:
> > > FormXXX
> > > FormXXXDialog <- machine generated
> > > FormXXXDialogImpl
> >
> > I still hold the opinion that the Form part is redundant. Why not:
> > > XXX
> > > XXXDialog <- machine generated
> > > XXXDialogImpl
> >
> > On the other 2 points I am agnostic...
> >
> > gr.ed.
>
> ok, good point. As is Allan's. How about :
>
> QXXX.[Ch]
> ui/QXXXDialog.[Ch]
> QXXXDialogImpl.[Ch]
>
> Kalle, does this address your "ls" concern ? Now QXXX and QXXXDialogImpl
> will be next to each other. As stuff in ui/ isn't human editable, it makes
> sense to keep it out of the way.

At least, I like the idea of dropping the Form prefix a lot (I never 
understood why it was there, probably an xforms thing). It only addresses my 
"ls" concern partly, as I would prefer to have the .ui files inlined as well 
- whether it is human editable is not so important (at least for me), rather 
whether it needs editing.

Oh, now I see that you want to put the Designer-generated C and h files into 
ui/. Yes, that would be good, especially if the .ui files stay out of there. 
Could be a bit confusing for the casual observer that the .ui files are not 
in the ui/ directory. Perhaps the directory should be called uic/, then.

Kalle

-- 
Matthias Kalle Dalheimer
President & CEO/VD
Klarälvdalens Datakonsult AB
Fax +46-563-540028
Email [EMAIL PROTECTED]

Re: [PATCH] update INSTALL

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| On Wed, Jun 13, 2001 at 04:06:09PM +0200, Jean-Marc Lasgouttes wrote:
| 
| > > "John" == John Levon <[EMAIL PROTECTED]> writes:
| > 
| > John> self-documenting
| > 
| > I applied it, but I am not sure that GNU m4 (or any m4) is really
| > needed if people use a tarball of LyX... This is probably a bug.
| 
| is this a bug to do with dependencies and timestamps, or a requirement bug
| with the sigc++ stuff ?

Most likely a bug with the dist target.

-- 
Lgb

Re: [PATCH] paste bug (again)

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| On Wed, Jun 13, 2001 at 09:04:33AM +0200, Lars Gullik Bjønnes wrote:
| 
| > Actually I would prefere to see the addition of a String that takes a
| > ostream..., alternatively:
| 
| ho-hum, you were free to mention this after not liking the last attempt
| 
| no problem, I'll have another go.

Hey, I can't think of everything at once can I?

_and_ you are allowed to disagree with me.

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[Jean-Marc Lasgouttes]

> "John" == John Levon <[EMAIL PROTECTED]> writes:

John> On Thu, Jun 14, 2001 at 01:46:06AM +0100, John Levon wrote:
>> > Or the const char * _() could return a string. Or it could be
>> removed > altogether...
>> 
>> removed sounds good. I'll have a go building with this change.

John> dur, of course we use the char * one everywhere.

John> And cgrepall '(".*").*+' alone has 69 matches. Argh.

But the char const * should be promoted to string automatically, no?

JMarc

Re: xforms help needed !

[Jean-Marc Lasgouttes]

> "Allan" == Allan Rae <[EMAIL PROTECTED]> writes:

Allan> This could be another of those problems with nested tabfolders
Allan> -- much like the problem with setting shortcuts for them.
>>  What about using something else? Like a list on the left instead
>> of the first-level tabs?

Allan> Sure. All we need is time.

Fine.

>> And what about the patch you posted earlier for shortcut? Is it
>> worth applying?

Allan> Only once we've gotten rid of the nested tabfolders or xforms
Allan> is fixed so they work. I thinnk I mentioned then that I think
Allan> we _might_ be able to make them work by overriding or at least
Allan> supplying callbacks for the shortcut handlers in the inner
Allan> forms -- tabfolders have a large number of callbacks available
Allan> that are _NOT_ listed in the documentation but which are public
Allan> in the headers. Anyway, that becomes black magic and I'm not
Allan> inclinded to explore that path very far.

So the technique could be applied to the paragraph and document
popups? Some people are very eager to see this done...

JMarc

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:

| > "John" == John Levon <[EMAIL PROTECTED]> writes:
| 
| John> On Thu, Jun 14, 2001 at 01:46:06AM +0100, John Levon wrote:
| >> > Or the const char * _() could return a string. Or it could be
| >> removed > altogether...
| >> 
| >> removed sounds good. I'll have a go building with this change.
| 
| John> dur, of course we use the char * one everywhere.
| 
| John> And cgrepall '(".*").*+' alone has 69 matches. Argh.
| 
| But the char const * should be promoted to string automatically, no?

Please look into rewriting the code that has problems to use
stringstream, that will actually be better in the long run than to
fiddle with the _() "macros".

Just list the code-lines and I'll do the change...

As mentioned earlier, stringstreams often gives smaller code. Most
likely due to a smaller number of temporaries.

string a = string("hello") + string("world") + string("!");

How many temporaries? At least five.

with string stream this could have been:

ostringstream o;
o << "hello" << "world" << "!";
string a(o.str());

How many temporaries? one?

Also the ostream can often be passed in, and sometimes is passed in
anyway, so we end up with having only one ostream object shared
between several methods. (just look at the write methods)

-- 
Lgb

Re: [Bug] Accents in math mode

[Andre Poenitz]

> > No. It should get fixed during the summer.
> 
> Yes, but _which_ summer?

The summer of the year 2742 of course. Sorry for not telling the whole truth.

Andre'

-- 
André Pönitz . [EMAIL PROTECTED]

Re: New bug list

[Juergen Vigna]

On 14-Jun-2001 John Levon wrote:

> No, I don't have a really good solution (yet). One dumb idea is separate "Apply"
> and "Apply to current paragraph" buttons. Although it's not the best solution prolly,
> it does indicate in a concise way the problem I see currently.

I don't get what the two apply button should buy us? The way I see it we would
need a "Update" button which updates the Dialog with the underlying paragraph
data.

Where we disagree is in the automatic update of the Dialog when you change
to another paragraph while the Dialog is open. Obviously it will always open
with the values of the actual paragraph when you open it.

Well it should be easy to make this behaviour configurable (automatic update
or not) and the default has to be the old behaviour, obviously ;)

   Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

The trouble with being poor is that it takes up all your time.

layout/tabular

[Stephan Mietens]

Hi,

opening the tabular dialog box with the row/column tab gives you
two meanings of "alt-c": either this means Haligment center, or 
close. That should be non-ambiguous.

Greets,

Stephan

Re: importing tables into lyx

[j.heidemeier]

Thanks I looked in it an I presume it will work.
Well if LyX would embedd an Tcl Interpreter all the Import Export Stuff
could be handled very smoothly.
BTW do you have the description of the new format which looks as if 
inspired by XML ? It would be worth trying to support this, too.


 -- 
Dr. Joachim Heidemeier  Tel. +49-30-8903-2780 
Fachgebiet II 3.2   [EMAIL PROTECTED] 
Umweltbundesamt, Bismarckplatz 1 D-14191 Berlin

On Wed, 13 Jun 2001, Martin Vermeer wrote:

> Date: Wed, 13 Jun 2001 19:32:25 +0300
> From: Martin Vermeer <[EMAIL PROTECTED]>
> To: [EMAIL PROTECTED]
> Cc: [EMAIL PROTECTED]
> Subject: Re: importing tables into lyx
> 
> On Wed, Jun 13, 2001 at 07:35:22AM +0200, joachim heidemeier wrote:
> > Mailing-List: contact [EMAIL PROTECTED]; run by ezmlm
> > Precedence: bulk
> > X-No-Archive: yes
> > List-Post: 
> > List-Help: 
> > List-Unsubscribe: 
>
> > Delivered-To: mailing list [EMAIL PROTECTED]
> > X-Envelope-From: [EMAIL PROTECTED]
> > X-Envelope-To: [EMAIL PROTECTED]
> > Date: Wed, 13 Jun 2001 07:35:22 +0200
> > From: joachim heidemeier <[EMAIL PROTECTED]>
> > Organization: private-uucp-site
> > X-Mailer: Mozilla 4.76 [de] (X11; U; Linux 2.2.18 i586)
> > X-Accept-Language: en
> > To: [EMAIL PROTECTED]
> > Subject: importing tables into lyx
> > 
> > dear lyx developers,
> > as I urgently need importing RTF-documents into Lyx I have the problem
> > that I'm stuck with table imports (C.f the Thread "rereLyX and import of
> > MS-Word resp. RFT documents" from April 2001).
> > I now would like to try the route:
> > Table copy from Word-Doc to Exel -> csv Export -> transformation form
> > csv to a simple lyx-file with reasonable default values -> import to the
> > mail document without the tables.
> > For the transformation I would write a tcl script.
> > For this purpose I need a description of the table (or longtable)
> > coding  in lyx and a parameter list affecting the tables.
> > The idea is not to do a very smart copying and preserving all attribus
> > but to read a simple table in the document which is then refined within
> > lyx.
> > If others are interested in the script I'll publish it of course.
> > Please direct your answers to my email-account at work
> > [EMAIL PROTECTED]
> > Thanks in advance.
> > -- 
> > joachim heidemeier ([EMAIL PROTECTED])
> 
> Also have a look at tk#.tcl to be found at 
>http://www.netby.dk/Oest/Europa-Alle/vermeer/.
> 
> It only handles the old table format. 
> 
> Martin
> -- 
> Martin Vermeer  [EMAIL PROTECTED]
> Helsinki University of Technology 
> Department of Surveying
> P.O. Box 1200, FIN-02015 HUT, Finland
> :wq
> 

Q: tabular

[Stephan Mietens]

Hi,

a short question: is there a way to define a tabular such that
lyx generates tex code looking like this?:

\begin{tabular}{|c||c|}
  ^^
Greets,

Stephan

Q: tabular<2>

[Stephan Mietens]

Hi again,

> a short question: is there a way to define a tabular such that
> lyx generates tex code looking like this?:
>
> \begin{tabular}{|c||c|}
>   ^^

ok, I got the answer right after sending the email, sorry.
But FYI importing tex with such a table will ignore the "||".

Greets,

Stephan

RE: Q: tabular

[Juergen Vigna]

On 14-Jun-2001 Stephan Mietens wrote:
> a short question: is there a way to define a tabular such that
> lyx generates tex code looking like this?:
> 
> \begin{tabular}{|c||c|}

Define the right border-line for the left column and the right border
line for the right column. But this has always been like this and you'll
see it also on screen.

  Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

Reality is just a convenient measure of complexity.
-- Alvy Ray Smith

Re: importing tables into lyx

[Juergen Vigna]

On 14-Jun-2001 j.heidemeier wrote:
> Thanks I looked in it an I presume it will work.
> Well if LyX would embedd an Tcl Interpreter all the Import Export Stuff
> could be handled very smoothly.
> BTW do you have the description of the new format which looks as if 
> inspired by XML ? It would be worth trying to support this, too.

Well no we(I) didn't write a format description, but it should be really
easy to figure out what it does, shouldn't it?

   Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

"Of all the tyrannies that affect mankind, tyranny in religion is the worst."
- Thomas Paine

Re: Single Quotes Curl

[Jean-Marc Lasgouttes]

> "Stan" == Stan Gatchel <[EMAIL PROTECTED]> writes:

Stan> Pairs of double quotes curl in different directions but single
Stan> quotes don't. It doesn't look correct.

Double quotes are special quote inset which have intelligence in them.
The single quote is plain old quote. To get 'intelligent' single
quotes, you have to go to Layout>Documents>Quotes (or Layout>Quotes on
1.1.5) and chnage to single quotes.

Stan> Your bug number 19 seems to indicate the problem (I think):

No, this is a different problem (which is fixed in 1.1.6fix2).

JMarc

Re: www-user with CSS

[Jean-Marc Lasgouttes]

> "Michael" == Michael Koziarski <[EMAIL PROTECTED]> writes:

Michael> Netscape 4.x? I guess the choices are either

Michael> 1) get the php to turn off css for your browser 2) break the
Michael> css.

Michael> I'd prefer 1. We could maintain a list of taglines of broken
Michael> browsers and have the css turned off by default
Michael> (index.php3?nocss=1). With an option for them to turn it back
Michael> on again if they feel the need. What do others think?

If you say there is no clean solution to this in netscape, so be it. I
can live with the nocss version, but it really does not look the
same...

JMArc

Re: LyX 1.1.6fix3 and namespaces, take two

[Jean-Marc Lasgouttes]

> "Yves" == Yves Bastide <[EMAIL PROTECTED]> writes:

Yves> Here is a second patch for building 1.1.6fix with gcc 3.0. This
Yves> one should be less bad :)

I applied it. Thanks.

JMarc

Re: layout/tabular

[Jean-Marc Lasgouttes]

> "Stephan" == Stephan Mietens <[EMAIL PROTECTED]> writes:

Stephan> Hi, opening the tabular dialog box with the row/column tab
Stephan> gives you two meanings of "alt-c": either this means
Stephan> Haligment center, or close. That should be non-ambiguous.

This is somewhat fixed in 1.2.0cvs, where Close does not have a
shortcut at all. I'm not sure it is the best solution, though (and
what the motivation was).

JMarc

Re: layout/tabular

[Juergen Vigna]

On 14-Jun-2001 Jean-Marc Lasgouttes wrote:

> This is somewhat fixed in 1.2.0cvs, where Close does not have a
> shortcut at all. I'm not sure it is the best solution, though (and
> what the motivation was).

Because the Dialogs can all be closed with ESC, so we really don't need
the C shortcut for the Close button!

   Jürgen

--
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._
Dr. Jürgen VignaE-Mail:  [EMAIL PROTECTED]
Italienallee 13/N   Tel/Fax: +39-0471-450260 / +39-0471-450253
I-39100 Bozen   Web: http://www.sad.it/~jug
-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._

The finest eloquence is that which gets things done.

Re: [PATCH] paste bug (again)

[John Levon]

On Thu, Jun 14, 2001 at 09:00:49AM +0200, Lars Gullik Bjønnes wrote:

> John Levon <[EMAIL PROTECTED]> writes:
> 
> | On Wed, Jun 13, 2001 at 09:04:33AM +0200, Lars Gullik Bjønnes wrote:
> | 
> | > Actually I would prefere to see the addition of a String that takes a
> | > ostream..., alternatively:
> | 
> | ho-hum, you were free to mention this after not liking the last attempt
> | 
> | no problem, I'll have another go.
> 
> Hey, I can't think of everything at once can I?

:)

> _and_ you are allowed to disagree with me.

not when you're probably right ;)

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: Qt2 filenames

[John Levon]

On Thu, Jun 14, 2001 at 08:33:54AM +0200, Kalle Dalheimer wrote:

> Oh, now I see that you want to put the Designer-generated C and h files into 
> ui/. Yes, that would be good, especially if the .ui files stay out of there. 
> Could be a bit confusing for the casual observer that the .ui files are not 
> in the ui/ directory. Perhaps the directory should be called uic/, then.

actually I would prefer the .ui files to go in there too.

However your suggestion seems a reasonable compromise. So we are all agreed on :

QTabCreate.[Ch]
QTabCreate.ui
QTabCReateDialogImpl.[Ch]
uic/QTabCreateDialog.[Ch]

yes ?

thanks,
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: New bug list

[John Levon]

On Thu, Jun 14, 2001 at 10:47:03AM +0200, Juergen Vigna wrote:

> 
> On 14-Jun-2001 John Levon wrote:
> 
> > No, I don't have a really good solution (yet). One dumb idea is separate "Apply"
> > and "Apply to current paragraph" buttons. Although it's not the best solution 
>prolly,
> > it does indicate in a concise way the problem I see currently.
> 
> I don't get what the two apply button should buy us? The way I see it we would
> need a "Update" button which updates the Dialog with the underlying paragraph
> data.

A naive user has the reasonable expectation that a "viewport" onto an object's 
parameters
will change only that object's parameters, rather than some other unrelated object's 
parameters.

You see the dialog as a "current paragraph parameters" dialog, but this is clearly
inconsistent with its current behaviour, as you point out below.

Hopefully this is clearer :/

> Where we disagree is in the automatic update of the Dialog when you change
> to another paragraph while the Dialog is open. Obviously it will always open
> with the values of the actual paragraph when you open it.
> 
> Well it should be easy to make this behaviour configurable (automatic update
> or not) and the default has to be the old behaviour, obviously ;)

configuration is the last refuge of bad interface design. hopefully we can avoid
this. I don't know how (you are right, automatic update ruins what you want to do
and is obviously not a good solution)

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 10:24:12AM +0200, Jean-Marc Lasgouttes wrote:

> John> And cgrepall '(".*").*+' alone has 69 matches. Argh.
> 
> But the char const * should be promoted to string automatically, no?

hmmm... then it is a gcc 3.0 bug ???

I don't know :/

if so, do we deal with it, or do we just add "don't use gcc 3.0" in INSTALL ?

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 10:34:54AM +0200, Lars Gullik Bjønnes wrote:

> Please look into rewriting the code that has problems to use
> stringstream, that will actually be better in the long run than to
> fiddle with the _() "macros".
> 
> Just list the code-lines and I'll do the change...

yes, ok.

if this is a compiler bug, it's a fair enough workaround. though if not,
I would prefer to have a special C_() macro for using c-style strings ...

the core lyx code will always use string _(string) after all.

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| On Thu, Jun 14, 2001 at 10:24:12AM +0200, Jean-Marc Lasgouttes wrote:
| 
| > John> And cgrepall '(".*").*+' alone has 69 matches. Argh.
| > 
| > But the char const * should be promoted to string automatically, no?
| 
| hmmm... then it is a gcc 3.0 bug ???

can some one refresh my memory... what errors are we seening? On the
1.1.6 branch only right?

| I don't know :/
| 
| if so, do we deal with it, or do we just add "don't use gcc 3.0" in
| INSTALL ?

that is not an option.

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 06:13:02PM +0200, Lars Gullik Bjønnes wrote:

> John Levon <[EMAIL PROTECTED]> writes:
> 
> | On Thu, Jun 14, 2001 at 10:24:12AM +0200, Jean-Marc Lasgouttes wrote:
> | 
> | > John> And cgrepall '(".*").*+' alone has 69 matches. Argh.
> | > 
> | > But the char const * should be promoted to string automatically, no?
> | 
> | hmmm... then it is a gcc 3.0 bug ???
> 
> can some one refresh my memory... what errors are we seening? On the
> 1.1.6 branch only right?

nope, in current CVS. It is corrupting the ->message() we send to the minibuffer.

One example is "Saving document" in lyxfunc.C - it is a garbled string being created.

http://sourceforge.net/tracker/index.php?func=detail=429678_id=15212=115212

this is with ENABLE_NLS

> | if so, do we deal with it, or do we just add "don't use gcc 3.0" in
> | INSTALL ?
> 
> that is not an option.

good !

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| nope, in current CVS. It is corrupting the ->message() we send to
| the minibuffer. 
| 
| One example is "Saving document" in lyxfunc.C - it is a garbled
| string being created. 

Og in what order is A + B + C + D evaluated?
If it is from left to right...

const char* + char  this is not a legal binding...
but if we change the char to int...
const char* + int then we just create a new pointer from the ints
value position...

So... this code is legal, but bogus. This is a bug.

Fix is still to change to stringstream:

ostringstream s1;
s1 << _("Saving document") << ' ' <<
MakeDisplayPath(owner->buffer()->fileName() << "...";
owner->message(s1.str().c_str());


Other locations please...

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 06:36:14PM +0200, Lars Gullik Bjønnes wrote:

> So... this code is legal, but bogus. This is a bug.

ok.

> Other locations please...

I will collect them this evening. There are tens of these, by the way.

What fun I'll have :)

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| > Other locations please...
| 
| I will collect them this evening. There are tens of these, by the way.
| 
| What fun I'll have :)

I am tag-searching through the sources right now and fixing as I go
along. So you might want to wait a bit.

-- 
Lgb

Re: Problem with gettext _() (bug 429678)

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| I will collect them this evening. There are tens of these, by the way.

I think I have gotten most of the dangerous ones now, give me time to
compile and test + commit.

Then tell me if you see the same. 

-- 
Lgb

Searching for code which handles string like char const *

[Jean-Marc Lasgouttes]

The bug I fixed for spellchecker yesterday was because of some code
which looked for character '\0' to detect end of line. I tried to see
whether I could find more.

The typical think I want to catch is
for (int i = 0; str[i]; ++i)
CursorRight(bview);
or variation around the same stupid pattern.

The best grep pattern I found for now is
  grep '\(if\|while\|for\) *(.*[a-z]*\[[^]]\] *\(&&.*\)*)' {,*/,*/*/}*.C
which is run in src/. As far as I see, the only interesting hit is
  kbsequence.C:   if (s[i] && (s[i]) <= ' ') ++i;
which is bad code (as all of kb_sequence::parse,) but not dangerous.

Of course, this regexp returns too many false positives, and I am
wondering whether it sees all cases. So if some creative regexp hacker
has ideas... 

JMarc

Re: www-user with CSS

[Zvezdan Petkovic]

On Thu, Jun 14, 2001 at 05:05:46PM +0200, Jean-Marc Lasgouttes wrote:
> > "Michael" == Michael Koziarski <[EMAIL PROTECTED]> writes:
> 
> Michael> Netscape 4.x? I guess the choices are either
> 
> Michael> 1) get the php to turn off css for your browser 2) break the
> Michael> css.
> 
> Michael> I'd prefer 1. We could maintain a list of taglines of broken
> Michael> browsers and have the css turned off by default
> Michael> (index.php3?nocss=1). With an option for them to turn it back
> Michael> on again if they feel the need. What do others think?
> 
> If you say there is no clean solution to this in netscape, so be it. I
> can live with the nocss version, but it really does not look the
> same...
> 
> JMArc

First let me admit that I didn't follow the complete correspondence but
I do not quite get the problem.
 
There _IS_ a way to make a clean CSS for Netscape 4.7x. I've spent a lot
of time to do some tricks but succeeded and my CSS and XHTML _validated_
clean.
  
Take a look at my Web page and corresponding css files under
URL/style/personal.css and URL/style/pictures.css.  Regarding the issue
of black letters you mention, I use in pictures.css:

/*
 * personal.css must be loaded before this file
 */
body, p, ul, ol, dl, address, blockquote, td, caption,
h1, h2, h3, h4, h5, h6, th,
tt, pre, code, kbd, samp {
color: white;
background-color: rgb(53,99,144)
}
   
I admit it's overkill, but it validates as clean CSS and does the job.

In personal.css there are some tricks to make a boxed paragraph with a
different background, paragraph with a border, paragraph with a red line
to the left for notices, small caps, drop caps, transparent images
looking really transparent... Unfortunately, Konqueror doesn't get the
trick for small caps and drop caps yet. But it looks fine in Netscape
and IE.
 
I believe that you deliberately refrain from such tricks but they work
and they _are_ clean CSS.
 
If I'm missing the point please forgive me. :-)

-- 
Zvezdan Petkovic <[EMAIL PROTECTED]>
http://www.cs.wm.edu/~zvezdan/

Re: Searching for code which handles string like char const *

[Lars Gullik Bjønnes]

Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:

| The bug I fixed for spellchecker yesterday was because of some code
| which looked for character '\0' to detect end of line. I tried to see
| whether I could find more.

I don't think we have a lot of them.
 
| The typical think I want to catch is
| for (int i = 0; str[i]; ++i)
| CursorRight(bview);
| or variation around the same stupid pattern.

mmm...yes...  
| The best grep pattern I found for now is
|   grep '\(if\|while\|for\) *(.*[a-z]*\[[^]]\] *\(&&.*\)*)' {,*/,*/*/}*.C
| which is run in src/. As far as I see, the only interesting hit is
|   kbsequence.C:   if (s[i] && (s[i]) <= ' ') ++i;
| which is bad code (as all of kb_sequence::parse,) but not dangerous.
| 
| Of course, this regexp returns too many false positives, and I am
| wondering whether it sees all cases. So if some creative regexp hacker
| has ideas... 

I tried with a simpler wone:

 \[.\].*\+\+.

-- 
Lgb

Re: importing tables into lyx

[Joachim Heidemeier]

I think it doesn't solve the main problem that reLyX didn't handle
longtables.
I'm working on a tcl-package which supports the new tabular-format.
joachim
-- 
joachim heidemeier ([EMAIL PROTECTED])


On Wed, 13 Jun 2001, Herbert Voss wrote:

>
> On Wed, 13 Jun 2001, joachim heidemeier wrote:
>
> > as I urgently need importing RTF-documents into Lyx I have the problem
> > that I'm stuck with table imports (C.f the Thread "rereLyX and import of
> > MS-Word resp. RFT documents" from April 2001).
> > I now would like to try the route:
> > Table copy from Word-Doc to Exel -> csv Export -> transformation form
> > csv to a simple lyx-file with reasonable default values -> import to the
> > mail document without the tables.
>
> have you ever tried the package excel2latex, available
> at CTAN?
>
> Herbert
>
>
>

Re: www-user with CSS

[A.]

> I believe that you deliberately refrain from such tricks but they work
> and they _are_ clean CSS.
>  
> If I'm missing the point please forgive me. :-)

No you're not missing the point :)

As I said when I started this work I am quite millitant on the issue of
CSS compliance and Netscape4 should be considered broken and have CSS
turned off.  I realise that if we added

"""
td, p, li, div, span, table, h1, h2, h3, h4 h5 . 
{ color: rbg(238,238,238) }
"""
It will work *ok* in netscape 4 but I don't see the point.  I guess I
could be convinced that as a *temporary* measure the changes could be
put in.  However I'd be inclined to remove them at the moment it becomes
difficult to maintain.

I've put the changes in at http://www.koziarski.org/LyX/www-user/ and
www-devel but they've introduced other problems (ignoring requests for
sans-serif fonts etc.)

Does that look better to everyone else?

-- 

Michael Koziarski
Software Developer
Compudigm International Ltd.

perl -e "print(pack('h45','d69636861656c604b6f6a796162737b696e236f6d6'))"

Re: Problem with gettext _() (bug 429678)

[John Levon]

On Thu, Jun 14, 2001 at 07:09:11PM +0200, Lars Gullik Bjønnes wrote:

> John Levon <[EMAIL PROTECTED]> writes:
> 
> | I will collect them this evening. There are tens of these, by the way.
> 
> I think I have gotten most of the dangerous ones now, give me time to
> compile and test + commit.
> 
> Then tell me if you see the same. 

certainly the bug I noticed is fixed. I also couldn't find any others
you hadn't dealt with.

thanks
john


-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

[PATCH] slight cleanup in bv_pimpl

[John Levon]

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard


Index: ChangeLog
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/ChangeLog,v
retrieving revision 1.193
diff -u -r1.193 ChangeLog
--- ChangeLog   2001/06/14 17:58:48 1.193
+++ ChangeLog   2001/06/15 00:22:05
@@ -1,3 +1,8 @@
+2001-06-15  John Levon  <[EMAIL PROTECTED]>
+
+   * BufferView_pimpl.C: fix typo, remove minibuffer message
+ when buffer has loaded
+
 2001-06-14  Lars Gullik Bjønnes  <[EMAIL PROTECTED]>
 
* lyxfunc.C (Dispatch): use stringstream
Index: BufferView_pimpl.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/BufferView_pimpl.C,v
retrieving revision 1.118
diff -u -r1.118 BufferView_pimpl.C
--- BufferView_pimpl.C  2001/06/14 17:58:48 1.118
+++ BufferView_pimpl.C  2001/06/15 00:22:10
@@ -347,6 +347,9 @@
setState();
AllowInput(bv_);
 
+   /// clear the "Formatting Document" message 
+   owner_->message("");
+ 
/// get rid of the splash screen if it's not gone already
owner_->getDialogs()->destroySplash();
  
@@ -3199,7 +3202,7 @@
break;
 
case LFUN_UNKNOWN_ACTION:
-   owner_->getLyXFunc()->setErrorMessage(N_("Unknow function!"));
+   owner_->getLyXFunc()->setErrorMessage(N_("Unknown function!"));
break;

default:

tex-defs.h removal

[John Levon]

it's still waiting to be applied.

also lars I still need write on lyxdoc so I can commit the Intro
patch mike ressler approved.

And I would like write on qt2/ too so I can do the big renaming

thanks
john

-- 
"Research is not a warm puppy. At least, I hope not."
- David Rydeheard

Re: www-user with CSS

[Zvezdan Petkovic]

On Fri, Jun 15, 2001 at 12:09:45PM +1200, Michael A. Koziarski wrote:
> > I believe that you deliberately refrain from such tricks but they work
> > and they _are_ clean CSS.
> >  
> > If I'm missing the point please forgive me. :-)
> 
> No you're not missing the point :)
> 
> As I said when I started this work I am quite millitant on the issue of
> CSS compliance and Netscape4 should be considered broken and have CSS
> turned off.  I realise that if we added
> 
> """
> td, p, li, div, span, table, h1, h2, h3, h4 h5 . 
> { color: rbg(238,238,238) }
> """
> It will work *ok* in netscape 4 but I don't see the point.  I guess I
> could be convinced that as a *temporary* measure the changes could be
> put in.  However I'd be inclined to remove them at the moment it becomes
> difficult to maintain.
> 

I absolutely agree with you. 

However, there's a way to improve the maintainability of CSS. Get rid of
redundancy! Below are a few examples.

But first the remaining issues. This time I looked carefully in your CSS
and XHTML.  I noticed the following:

a) You do not use XHTML correctly in some places and hence the problem
with black letters at the bottom of the page. The thing between two
's should be inside  If you used XHTML strict instead of
transitional the validator would've reported the issue. 

b) The problem of non-serif bulleted items is related to the previous
colour problem and Netscape obtuseness. Read more below and notice use of
ul, ol, and dl below on the line where I define the fonts.

c) I'm said to say this, but your CSS is an example of _non-user_
friendly design. If you are so militant on the issue of compliance why
do you avoid to comply yourself? You use px and pt sizes all over the
place. The first rule of a good Web design (or any design for that
matter) is: "Give a choice to user." They will choose the preferred
size in their browser. CSS designer should use the relative sizes in em
units (width of a small letter m in a given font). See examples below
for headings.

Finally, I'm back to the issue of maintainability.

There are basically two approaches to writing CSS:

1. Put everything that belongs to an element together, as in:

body {

}

   You are using that approach.

   Advantage: everything related to an element is in one place.
   Disadvantage: repetition of things that can be together (e.g. we
   want all headers to be bold, why repeat 6 times?)

   Hence,

2. Put everything common together. This is the approach that I and
   some other people use.

   Disadvantage: You have to use search feature of your editor to
   find all parts related to a single element (not too difficult).

   Advantage: You quickly find and change fonts, colours, weight.
   Once all browsers work well with CSS (and there is currently none
   that supports it completely -- give me one that supports dropped
   capitals?), you can remove elements easily. There is no
   redundancy in the code though.
   _And it works in all browsers!_

   Here's an example:

   /*
*  Fonts
*/
   body, p, ul, ol, dl, address, td, caption {
   font-family: "Georgia", serif;
   }

   h1, h2, h3, h4, h5, h6, th {
   font-family: "Verdana", sans-serif;
   }

   tt, pre, code, kbd, samp {
   font-family: "Andale Mono", monospace;
   }
   /* font sizes are defined below */

   /*
*  Page layout
*/
   body, p, ul, ol, dl {
   text-align: justify
   }

   body {
   margin-left: 10%;
   margin-right: 10%
   }
   
   h1 {
   margin-left: -8%;
   text-align: left;
   font-size: 1.728em;
   font-weight: bold
   }

   h2, h3, h4, h5, h6 {
   margin-left: -4%;
   text-align: left
   }

   h2 {
   font-size: 1.44em;
   font-weight: bold
   }

   h3 {
   font-size: 1.2em;
   font-style: italic;
   font-weight: bold
   }

   h4, h5, h6 {
   font-size: 1em;
   }

   h4 {
   font-weight: bold
   }

   h5, h6 {
   font-weight: normal;
   }

   h5 {
   font-style: italic;
   }

   
   
As you can see there is not a single repeated element above. You do
not need ul any more in font definitions because browsers work? Fine,
take it from that line and everything is OK. The other important
settings for ul still remain somewhere in the file:

   ul.square {
   list-style-type: square
   }

***
To conclude, most browsers are dumb in regard to CSS. We have to provide
precise definitions for all elements instead of relying on inheritance.
To increase maintainability, decrease redundancy. Let's hope that one
day the browsers will understand CSS 

Re: www-user with CSS

[A.]

> I absolutely agree with you. 
> 
> However, there's a way to improve the maintainability of CSS. Get rid of
> redundancy! Below are a few examples.
> 
> But first the remaining issues. This time I looked carefully in your CSS
> and XHTML.  I noticed the following:
> 
> a) You do not use XHTML correctly in some places and hence the problem
> with black letters at the bottom of the page. The thing between two
> 's should be inside  If you used XHTML strict instead of
> transitional the validator would've reported the issue. 

I'm aware of this problem, the footer has been updated on the main site
anyway so this isn't an issue. (I'm just too lazy to run cvs udpate :))

> b) The problem of non-serif bulleted items is related to the previous
> colour problem and Netscape obtuseness. Read more below and notice use of
> ul, ol, and dl below on the line where I define the fonts.

That's a good idea.  I'll do that on the weekend (stupid paid work is
keeping me busy)

> CSS designer should use the relative sizes in em
> units (width of a small letter m in a given font). See examples below
> for headings.

Yep, I realise this.  the css itself was ripped from 3 sites I wrote in
98.  I intend to use em for all the sizes.  Just a matter of converting
them all to rational values (I'm not much of a designer so my choices
would be ... ugly :))


> Finally, I'm back to the issue of maintainability.
> 
> There are basically two approaches to writing CSS:

I'll use method 2 over the weekend.


> ***
> To conclude, most browsers are dumb in regard to CSS. We have to provide
> precise definitions for all elements instead of relying on inheritance.
> To increase maintainability, decrease redundancy. Let's hope that one
> day the browsers will understand CSS completely.

This is where my millitancy comes in :).  The web will *never* leave its
current state of  < ... bgcolor= ... > etc. while people continue
to contort themselves to ensure 'works in [insert broken browser]'.
I'll add the suggestions you mentioned above (they're good and I can see
that they will be maintainable) but any reports of broken rendering with
*that* and I'll just suggest an upgrade or turn off css.


> And both are
> pretty dumb. I remember that IE on Win2000 didn't recognise UTF-8 in the
> first xml line of my strict XHTML doc.

I must admit that's an interesting new bug :)

-- 

Michael Koziarski
Software Developer
Compudigm International Ltd.

perl -e "print(pack('h45','d69636861656c604b6f6a796162737b696e236f6d6'))"

Re: xforms help needed !

[Allan Rae]

On 14 Jun 2001, Jean-Marc Lasgouttes wrote:

[shortcuts for tabfolders]

> So the technique could be applied to the paragraph and document
> popups? Some people are very eager to see this done...

Definitely can be used now for tabfolders that are not nested.
All you need to do is think of suitable shortcut key combination.
(and then add the code of course ;-)

Allan. (ARRae)

Re: Qt2 filenames

[Allan Rae]

On Thu, 14 Jun 2001, Kalle Dalheimer wrote:

> At least, I like the idea of dropping the Form prefix a lot (I never
> understood why it was there, probably an xforms thing).

The xforms port was always using Form* for our code and form_* for
generated code.  Then other ports started and people refused to entertain
their intellects and just copied the naming convention of the xforms port
even though it was totally irrelevent to the port they were doing.

They even ignored protests about fixing the names used in their port on
the grounds that there was some obtuse logic behind it all (which only
they could see).

Finally an injection of new blood and some small flames later sanity
returns to file naming conventions.

Allan. (ARRae)  wipes brow; steps down off soapbox; exits stage right.

Re: [PATCH] ignore empty minibuffer lines

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| If a user enters a empty minibuffer input, it should be ignored IMHO

Ok, this is good.

_but_ should this be done in the minibuffer code? IMHO the minibuffer
should not define policy that should be done by lyxfunc... so the
check for empty command should not be done in minibuffer, but in the
method run by the emit.

-- 
Lgb

Re: [PATCH] slight cleanup in bv_pimpl

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| thanks
| john

Ok, do this.

-- 
Lgb

Re: [PATCH] help out the internet !

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| ... but maybe put the HD manufacturers out of business

As long as we don't remove it from CVS.

What do others say?

-- 
Lgb

Re: tex-defs.h removal

[Lars Gullik Bjønnes]

John Levon <[EMAIL PROTECTED]> writes:

| it's still waiting to be applied.
| 
| also lars I still need write on lyxdoc so I can commit the Intro
| patch mike ressler approved.
| 
| And I would like write on qt2/ too so I can do the big renaming

Ok, done.

-- 
Lgb

Re: [PATCH] help out the internet !

[Allan Rae]

On 15 Jun 2001, Lars Gullik Bjønnes wrote:

> John Levon <[EMAIL PROTECTED]> writes:
>
> | ... but maybe put the HD manufacturers out of business
>
> As long as we don't remove it from CVS.
>
> What do others say?

Ooooh,  I don't know.  It definitely needs to stay in CVS.
It would be nice to trim the distro a little if possible and
this doesn't really affect the distro in any way.

So, sure, why not, lets purge a few more things.  Maybe even take out
ChangeLog.1 since that is also large and isn't really going to interest
users -- anyone serious can get CVS'd.

There must be other stuff we can cull too.  I'm sure I had a list here
somewhere...

Allan. (ARRae)