Mail Archives: djgpp-workers/1999/04/27/12:51:21
On 27 Apr 99, at 15:31, Eli Zaretskii wrote:
>
> On Tue, 27 Apr 1999, Andris Pavenis wrote:
>
> > I suggest following way of testing:
> > grep sources to find where DJGPP_MINOR is used
> >
> > add
> > #if defined(__DJGPP__) && !defined(__DJGPP_MINOR__)
> > #error DJGPP_MINOR is not defined
> > #endif
> > before using this macro
>
> Many packages include their private headers which then include other
> headers which then include standard headers. I cannot afford going
> through all the Megabytes of ported packages and do this. Volunteers
> are welcome to do it, as always.
I'm afraid that it will be needed to revisit sources of many packages after current
upgrade anyway. Maybe even it would be needed to rebuild many packages
with DJGPP-2.03 and perhaps egcs-1.1.2
For example I unpacked dvlj26s.zip and kpse32s.zip and make ends with
compilation error due to different reasons (not DJGPP_MINOR related as I added
test for it):
gcc tex-file.c -DHAVE_CONFIG_H -I. -I. -I.. -I./.. -g -O2 -c -o STATIC/tex-file.
o
In file included from ../kpathsea/config.h:65,
from tex-file.c:19:
../kpathsea/lib.h:117: conflicting types for `basename'
c:/djgpp/include/unistd.h:127: previous declaration of `basename'
make.exe[1]: *** [tex-file.lo] Error 1
make.exe[1]: Leaving directory `d:/x/gnu/web2c-7.2/kpathsea'
make.exe: *** [do-kpathsea] Error 2
--------------------------------------------
unistd.gh defines basename to return char *, but other header file expects it
to return const char *
Note: it was first package I tested for TeX.
This is simply an ilustration of problems we can expect and I think
adding test for presence of DJGPP_MINOR can be done at the same
time.
>
> > Also I don't like that way. I don't like to introduce similar things.
>
> Perhaps you could suggest a better solution, then.
>
Andris
- Raw text -