[prev in list] [next in list] [prev in thread] [next in thread]
List: kde-devel
Subject: Re: gcc-2.95 vs. KDE_1_1_BRANCH
From: Franz Sirl <Franz.Sirl-kernel () lauterbach ! com>
Date: 1999-07-15 19:52:33
[Download RAW message or body]
Am Don, 15 Jul 1999 schrieb Harri Porten:
>Franz Sirl wrote:
>>
>> below are the remaining problems compiling KDE_1_1_BRANCH (from
>> openprojects as of yesterday) with gcc-2.95 (to be released in the next
>> days).
>
>I had checked some of the packages before. That's why I was surprised
>that you encountered so many errors. My egcs snapshot was about a month
>old so I had to upgrade to reproduce the problems. Did you try
>kdesupport ? The situation is really bad there. *Lots* of warnings (and
>errors which I believe to have fixed today). However the 19990712
>behaves somehow strangely. Declarations like fdopen() and gethostname()
>seem to be missing although the relevant headers are included. Known
>problem ?
For which package? If it's kdelibs, you may need a patched autoconf-2.13, I've
attached the patch I use (sorry, I did forget to mention it).
I didn't test kdesupport yet, cause the old 1.1.1 package didn't fail for me
and I didn't think that it changed a lot, I'll try it now.
>> ---
>> kdegames:
>> make -f Makefile.cvs fails with:
>> *** Building Makefile templates (step one)
>> po/Makefile.am:2: required directory po/sl does not exist
>
>In case your cvs update isn't complete or the dir is really missing
>simply take `po' out of kdegame's Makfile.
Ah, ok. I'm still quite new to KDE cvs builds. I use the following sequence to
build tar.gz files fitting our distribution building:
cvsup kde
cd kdepackage
ln -s ../kde-common/admin .
make -f Makefile.cvs
cd ..
tar -czhf kdepackage.tar.gz kdepackage
>> ---
>> kdenetwork:
>> 1.
>> g++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/include -I/usr/include/qt
>> -I/usr/X11R6/include -I. -I. -I.. -O2 -fsigned-char -DNO_DEBUG -c status.cpp
>> status.cpp: In method KBiffStatus::KBiffStatus(const QString &, const
>> KBiffStatusList &)':
>> status.cpp:42: initialization to KBiffStatusList *' from const
>> KBiffStatusList *' discards qualifiers
>
>This fix I found required *removing* one const so I'll mail it to Kurt
>first.
OK.
>> 3.
>> g++ -DHAVE_CONFIG_H -I. -I. -I../.. -I/usr/include -I/usr/include/qt
>> -I/usr/X11R6/include -DHAVE_KDE -O2 -fsigned-char -DNO_DEBUG -c find_user.cpp
>> find_user.cpp: In function int find_X_process(char *, char *)':
>> find_user.cpp:159: converting int (*)(dirent *)' to int (*)(const dirent *)'
>> is a contravariance violation
>
>Fixed.
Thanks.
>> ---
>> kdemultimedia:
>> g++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/include -I/usr/include/qt
>> -I/usr/X11R6/include -O2 -fsigned-char -DNO_DEBUG -c kmix.cpp
>> kmix.cpp: In method void KMix::createWidgets()':
>> kmix.cpp:312: no matches converting function contextMenu' to type class
>> QPopupMenu * (*)(QObject *, QObject *)'
>> kmix.h:83: candidates are: class QPopupMenu * KMix::contextMenu(QObject *,
>> QObject *)
>
>I've mailed this to Christian Esken. I'm not sure what place needs to be
>fixed.
:-)
>> ---
>> kdetoys:
>> make[2]: Entering directory /home/fsirl/kdetoys/BUILD/kdetoys/mouse'
>> g++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/include -I/usr/include/qt
>> -I/usr/X11R6/include -O2 -fsigned-char -c kodo.cpp
>> In file included from kodo.h:11,
>> from kodo.cpp:112:
>> gridbag.h:93: ANSI C++ forbids initialization of member RELATIVE'
>> gridbag.h:93: making RELATIVE' static
>
>I wonder how this compiled before. I've fixed this and some `const'
>errors on my disc but I can't get kodo to work (broken pixmaps?) so I
>won't commit my patches yet. Can anyone confirm that the program works
>*now* ?
Hmm, til now I haven't found time to actually download and use the stuff I
built :-).
BTW, is Matthias away? I've sent some big-endian and sound fixes to him, but
haven't heard anything back yet.
Franz.
["autoconf-2.13-gcc295.patch" (text/x-c)]
Index: acspecific.m4
===================================================================
RCS file: /cvs/autoconf/autoconf/acspecific.m4,v
retrieving revision 1.174
diff -u -p -r1.174 acspecific.m4
--- acspecific.m4 1999/04/22 21:38:27 1.174
+++ acspecific.m4 1999/05/23 00:32:27
@@ -1087,6 +1087,11 @@ AC_CACHE_CHECK(for working mmap, ac_cv_f
#ifdef __cplusplus
extern "C" { void *malloc(unsigned); }
+extern "C" { int rand(void); }
+extern "C" { int umask(int); }
+# ifdef HAVE_UNISTD_H
+# include <unistd.h>
+# endif
#else
char *malloc();
#endif
@@ -1103,7 +1108,7 @@ main()
/*
* First, make a file with some known garbage in it.
*/
- data = malloc(pagesize);
+ data = (char *)malloc(pagesize);
if (!data)
exit(1);
for (i = 0; i < pagesize; ++i)
@@ -1124,7 +1129,7 @@ main()
fd = open("conftestmmap", O_RDWR);
if (fd < 0)
exit(1);
- data2 = malloc(2 * pagesize);
+ data2 = (char *)malloc(2 * pagesize);
if (!data2)
exit(1);
data2 += (pagesize - ((int) data2 & (pagesize - 1))) & (pagesize - 1);
@@ -1142,7 +1147,7 @@ main()
*/
for (i = 0; i < pagesize; ++i)
*(data2 + i) = *(data2 + i) + 1;
- data3 = malloc(pagesize);
+ data3 = (char *)malloc(pagesize);
if (!data3)
exit(1);
if (read(fd, data3, pagesize) != pagesize)
@@ -1399,7 +1404,7 @@ AC_DEFUN(AC_FUNC_ALLOCA,
# The Ultrix 4.2 mips builtin alloca declared by alloca.h only works
# for constant arguments. Useless!
AC_CACHE_CHECK([for working alloca.h], ac_cv_header_alloca_h,
-[AC_TRY_LINK([#include <alloca.h>], [char *p = alloca(2 * sizeof(int));],
+[AC_TRY_LINK([#include <alloca.h>], [char *p = (char *)alloca(2 * sizeof(int));],
ac_cv_header_alloca_h=yes, ac_cv_header_alloca_h=no)])
if test $ac_cv_header_alloca_h = yes; then
AC_DEFINE(HAVE_ALLOCA_H)
@@ -2044,7 +2049,7 @@ AC_DEFUN(AC_C_INLINE,
[AC_CACHE_CHECK([for inline], ac_cv_c_inline,
[ac_cv_c_inline=no
for ac_kw in inline __inline__ __inline; do
- AC_TRY_COMPILE(, [} $ac_kw foo() {], [ac_cv_c_inline=$ac_kw; break])
+ AC_TRY_COMPILE(, [} $ac_kw int foo() {], [ac_cv_c_inline=$ac_kw; break])
done
])
case "$ac_cv_c_inline" in
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic