Subject: Re: gdk-pixbuf vs. ImageMagick
From: Paul Rohr (firstname.lastname@example.org)
Date: Thu Apr 26 2001 - 08:29:44 CDT
At 03:05 PM 4/26/01 +0200, Hubert Figuiere wrote:
>According to Paul Rohr <email@example.com>:
>> However, I believe that there could still be as many as three proposals on
>> the table for XP implementations of that to-be-specified API:
>> A. We roll our own on top of libjpeg et al. We do all the work, but can
>> make as efficient a codepath as we like.
>> B. Something like an XP version of gdk-pixbuf. I can't tell if anyone's
>> advocating this, but if they did, the argument for this would seem to be
>> that it's less work than rolling our own, and also more efficient than IM.
>B. is not worth the effort. gdk-pixbuf uses libjpeg and libpng AFAIK. And
>gdk-pixbuf rely on gdk and glib. So using it in XP who require to
>bring gdk and glib to XP, which is not what we want. We'd better
>wrap ourself around libjpeg and libpng.
>> C. We just use ImageMagick or miniIM. This is probably the most
>> but seems to have bloat worries.
>The only advantage IM have over wrapping around individual libraries
>is to allow making one converter/rasterizer for lot of file formats.
>Making A easier to implement.
I understand that this is your argument, and I'm very sympathetic to it,
since it's more-or-less what I've been assuming all along. ;-)
However, I want to make sure that anyone who believes that A is *not* as
good a choice as B or C can make their arguments heard. Leonard does a fine
job of advocacy for C, and I want to make sure that anyone who advocates B
(such as, perhaps, Paolo) also has their say.
So, my original question remains -- *are* there any remaining advocates for
B, or are all the gdk-pixbuf folks just talking about using it as a
This archive was generated by hypermail 2b25 : Thu Apr 26 2001 - 08:22:06 CDT