On Fri, 19 May 2006 15:19:34 -0400, Philippe Valembois - Phil
<[EMAIL PROTECTED]> wrote:
Le Friday 19 May 2006 17:20, Youness Alaoui a écrit :
<snip>
The question is, though, how will it work on windows and mac ? the libs
are portable right ? (if no, maybe that was the reason he patched them,
then we could use the cximage libs only for those platforms..)
Libs are portable (the only trouble we had with this extension is with
GIFs
because it isn't a lib)...
Yes, I know about GIF.. but why would the guy include the other libs with
cximage, and I'm sure (almost) that he wrote somewhere, (maybe forums?)
that the libs were modified and should not be replaced because it won't
work if you do.. why noone wants to do a diff ? you lazy bastards :p lol
btw, if you look at his website, read the forums, or the codeproject's
website in the comments, you'll find a few patches and unicode fixes
submitted by other users, maybe we could patch our working copy of
cximage
in order to fix the unicode problem we had (will optimize opening).
Why not... but I can't test...
yes you can, just open a file with accented chars in it, you do have an fr
keyb, right? :p
or ask viv (or anyone) to send you a file with greek (or whatever) chars...
p.s.: I think this should go into 0.96 for sure, so let it be added to
the
TODO, BUT it will be a post RC1, and hopefully a pre RC2...
if not (if not enough manpower), we'll have to go into an RC3 before the
official 0.96 release...
I agree with you...
youpiiii :D
P.s.2: mabye OT, TkCximage is really slow, do you think we should either
use Img instead (assuming it has EVERYTHING we need, like anigifs,
resize,
colorize, etc...), or maybe we should really optimize it. If we decide
to
keep TkcxImage (which would be my choice for sure), how about, as a
side-project, completly unrelated to amsn, how about making tkcximage, a
real tkcximage library, which means, not an extension using cximage to
load/save files, but wrap into Tk commands ALL the cximage methods,
making
TkCximage supercede TkImg by making it not only an image reading
extension
but also an image manipulation extension. I'm sure the tk community
would
be thrilled in having an extension allowing them to create a tkphotoshop
app... :p
Well.... I am angry :p
You don't read my mails !
(hum don't take my anger seriously... I can understand you don't read all
mails but you could read mine :p)
looool, no, I do read ALL mails (ouch :( ) and I did read yours (no
favors:P) and be angry if you want, who cares? :p
Well, as I already said, I think optimization will be useless... We will
get
some microseconds at a place and some other at another place but I think
it's
a slow library because it is bad designed and I proposed to rewrite it
from
scratch with a better design (because for now it isn't done efficiently :
look at how animated GIFs are handled)
Well, you might be right, especially if we 'optimize'... I mean, we should
profile, then optimize the faulty code, not try to optimize all.. I
learned that without a good profiler, you'll just waste so much time
optimizing something that will save you 0.0001 ms and you leave another
which was taking minutes.. we must know WHAT to optimize before we do it,
and that's just as you said, if we don't, we'll get a few nano seconds if
any... Yes, the design is bad, or so, I may want to think, I didn't look
too much at the internals, but I guess it is, for one simple reason, it's
very COMPLEX. It has an internal way of storing the data that eases all
those effects operations, like blurring, sharpening, rotate, skew, dither,
emboss, edge, jitter, etc......
Rewriting all from scratch.. what can I say.. hummm.. ARE YOU CRAZY ???
lol, you may want to do it, but then you'll use C instead of C++, will you
reuse the code ? will you what ? you can't rewrite all that stuff, if you
do, you'll be rewriting TkImg.. in that case, let's just use TkImg! or
maybe modify Tkimg instead to suit our needs, no ? we created TkCximage
because it could use CxImage, have 'almost' no work involved in creating
the extension and that it would take less effort to write it than to try
to fix TkImg makefiles to get at least one person to compile TkImg
correctly.. but if you rewrite it, then we'll be loosing time... If there
is an internal way of storing the data, and that takes a lot of
unnecessary CPU to store it that way, then you might want to change only
the way its stored, and add methods to access it in the most efficient way
and keep the way it was accessed before by CxImage compatible in order to
not break anything else... in short, have a new way to store it, have a
method to access the RGBA easily, and modify the getData from getdata()
{ return data;} into getdata() {return NewFormatToOldFormatdata(data); }
I had some design in head and I should publish it in another place than
in
this mail which will be lost :p
Briefly it was a system were decoder is separated from the data and
where all
frames are in a container....
I will try to write this this evening or maybe tomorrow on the wiki (my
first
page :d)...
lol, congrats on first page.. but make it good :P don't waste time writing
something that you think could never get implemented, ok ?
KaKaRoTo
Phil
-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job
easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache
Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642
_______________________________________________
Amsn-devel mailing list
Amsn-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/amsn-devel
--
KaKaRoTo
-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642
_______________________________________________
Amsn-devel mailing list
Amsn-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/amsn-devel
