Build fails on CVS?

[Jerry Pisk]

I can comment on #2. That function was added in GD 2.0.16 to remove a
race condition when using GD in a multithreaded application. I don't
know what check is appropriate, I'm hoping there's a defined version
we can test against (something like #if (LIBGD_VERSION > 0x2000A)), if
not we will need to find a better test.

Jerry

On 4/13/05, Bill Binko <bill at binko.net> wrote:
> On Wed, 13 Apr 2005, Frank Warmerdam wrote:
> 
> > Bill,
> >
> > I suspect this was an oversite on Assefa's part when doing the
> > SVG stuff.  While the FastCGI stuff needs the stdargs.h, there
> > is no reason it couldn't be included in mapsvg.c as well, if it
> > is required.  If Assefa doesn't pick up on this pretty soon, I
> > would suggest you submit a bug report, possibly including
> > a patch for the change.
> >
> 
> I have gotten cvs HEAD to build here.  Here are the issues I fixed:
> 
> 1) the noted SVG problem - adding #include <stdarg.h> worked (I'm a dummy)
> 
> 2) in maputil.c, there was a test for USE_GD_FT which essentially tests
> whether libgd is > 2.0.12.  It then calls a funciton
> (gdFontCacheSetup())that is not in my libgd2.0.15.  I changed the test to
> look for gd_HAS_GETBITMAPFONTS which tests for libgd2.0.22 or higher.
> Perhaps Sean can take a look at what test makes sense there.
> 
> 3) My gradient patch didn't apply (mainly due to the C/C++ comment fixes),
> so I patched it by hand.
> 
> 4) My multiple !BOX! substitution patch didn't apply because of pramsey's
> work in allowing longer query strings.  I redid it to use his new method
> and still allow multiple substitutions.
> 
> This basically gives me four patches: one each against:
> mappostgis.c (for bug 1306)
> mapsvg.c (for #1 above)
> maputil.c (for #2 above)
> 
> and a larger patch for bug 1305 (the gradient stuff).
> map.h
> mapdraw.c
> mapfile.c
> mapfile.h
> maplayer.c
> maplexer.l
> 
> Frank (or anyone else): how should I get this in?  I guess I could create
> bugs for #1 and #2 above and attach them?  Then attach patches to 1305 and
> 1306?  Is there an easier way?
> 
> Bill
>