This is the mail archive of the
cygwin-apps@cygwin.com
mailing list for the Cygwin project.
Re: Package review status
- From: Nicholas Wourms <nwourms at netscape dot net>
- To: Pavel Tsekov <ptsekov at gmx dot net>
- Cc: Ryunosuke Satoh <sun594 at hotmail dot com>, cygwin-apps at cygwin dot com
- Date: Tue, 17 Sep 2002 18:50:30 -0400
- Subject: Re: Package review status
- References: <Pine.LNX.4.44.0209172111410.12870-100000@moria>
Pavel Tsekov wrote:
On Wed, 18 Sep 2002, Ryunosuke Satoh wrote:
The binary package looks OK. I'm not a doxygen user and I don't know how
to test it :) So I've just run the executables to see that they are
working.
I've been pretty busy, but I'll definitely look at it more
closely as soon as I can. Things that we should test for is
support for Cygwin's native latex/pdftex/ghostscript.
However I noticed something - the binary is built using the included
libpng. Cygwin has libpng as a package in the net distribution. Isn't it
better to link against that ?
Getting doxygen to link against it is a PITA b/c of the
TrollTech makefile mess that it uses for building. Still it
is possible, IIRC.
You're passing D_WIN32 together with D__CYGWIN__ when building. This is
kind of strange. I suppose D_WIN32 will trigger usage of native Windows
constructs rather than Posix ones. Btw you don't need to pass D__CYGWIN__
because the compiler already defines this symbol for you.. Well, unless it
isn't undefined somewhere in the source. Please, comment on this.
This is *very* concerning and something I did not have a
chance to look at closely when I first made comments on this
package. FWIW, I actually ported doxygen about a month ago.
It was quite a pain, because Dimitry has WIN32 and _WIN32
ifdef's littered throughout the source code. This is a
"bad" thing, because we don't want doxygen mixing the POSIX
api with the w32api [as it can and will result in runtime
conflicts]. I had a very unproductive chat with Dimitry in
an attempt to explain why using WIN32 ifdefs was bad [The
old Cygwin != Mingw issue]. Long story short, I located
every WIN32 ifdef and added !defined(__CYGWIN__) where
necessary. This is *required* to use Cygwin's
teTeX/Ghostscript implimentation. I can't recall, but I
believe this was also necessary to make sure doxygen
implicitly used POSIX path naming conventions. Needless to
say, it wasn't easy. The only reason I was holding off was
because I was planning to release it with the Doxywizard
right after I released QT-2. Since someone else is doing it
now, I will try to work with him on this. When I get some
free time, I'll dig up my patches and see what may (or may
not) need to be done to Ryunosuke's release. Meanwhile, if
anyone knowledgable in Doxygen (Stipe?) could please test
this release in building latex/pdf/postscript documentation,
that would be great. My main concern is to make sure that
people aren't required to have a separate install of
Ghostscript/w32 and fpTeX in addition to any Cygwin native
versions. I hope this doesn't seem too demanding on my part
:-(. Also, if these issues were resolved by Ryunosuke, then
please disregard this commentary.
Cheers,
Nicholas