configure complains about liblcms not being there
Hi, Since the last patches to configure.ac and friends (I haven't seen the warnings before) I get: configure: liblcms development files not found. Wine will be built without Color Management support. (mscms.dll) It look like configure is checking for the availability of lcms.h: configure:6948: checking lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 configure:6971: $? = 0 configure:6985: result: yes configure:6989: checking lcms.h presence configure:7004: gcc -m32 -E conftest.c configure:7010: $? = 0 configure:7024: result: yes configure:7057: checking for lcms.h configure:7065: result: yes configure:6948: checking lcms/lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 conftest.c:65:23: error: lcms/lcms.h: No such file or directory Mine is at /usr/include/lcms.h (and it's found). Shouldn't that be enough to satisfy configure? I doubt anyone will have both. My version for lcms: lcms-1.16-3 lcms-devel-1.16-3 Any idea? Cheers, Paul.
Same problem here on a Ubuntu dapper machine. On my Slackware box I got the same lcms test results in the config.log, but I did not get the warning message. On 7/10/07, Paul Vriens <paul.vriens.wine(a)gmail.com> wrote:
Hi,
Since the last patches to configure.ac and friends (I haven't seen the warnings before) I get:
configure: liblcms development files not found. Wine will be built without Color Management support. (mscms.dll)
It look like configure is checking for the availability of lcms.h:
configure:6948: checking lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 configure:6971: $? = 0 configure:6985: result: yes configure:6989: checking lcms.h presence configure:7004: gcc -m32 -E conftest.c configure:7010: $? = 0 configure:7024: result: yes configure:7057: checking for lcms.h configure:7065: result: yes configure:6948: checking lcms/lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 conftest.c:65:23: error: lcms/lcms.h: No such file or directory
Mine is at /usr/include/lcms.h (and it's found). Shouldn't that be enough to satisfy configure? I doubt anyone will have both.
My version for lcms:
lcms-1.16-3 lcms-devel-1.16-3
Any idea?
Cheers,
Paul.
Paul Vriens <paul.vriens.wine(a)gmail.com> writes:
It look like configure is checking for the availability of lcms.h:
configure:6948: checking lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 configure:6971: $? = 0 configure:6985: result: yes configure:6989: checking lcms.h presence configure:7004: gcc -m32 -E conftest.c configure:7010: $? = 0 configure:7024: result: yes configure:7057: checking for lcms.h configure:7065: result: yes configure:6948: checking lcms/lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 conftest.c:65:23: error: lcms/lcms.h: No such file or directory
Mine is at /usr/include/lcms.h (and it's found). Shouldn't that be enough to satisfy configure? I doubt anyone will have both.
Yes it's enough, only one of them needs to be found. What you are probably lacking is the .so symlink, the new code is more strict about this to make sure we always find the correct library soname. -- Alexandre Julliard julliard(a)winehq.org
Alexandre Julliard wrote:
Paul Vriens <paul.vriens.wine(a)gmail.com> writes:
It look like configure is checking for the availability of lcms.h:
configure:6948: checking lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 configure:6971: $? = 0 configure:6985: result: yes configure:6989: checking lcms.h presence configure:7004: gcc -m32 -E conftest.c configure:7010: $? = 0 configure:7024: result: yes configure:7057: checking for lcms.h configure:7065: result: yes configure:6948: checking lcms/lcms.h usability configure:6965: gcc -m32 -c -Wcast-qual conftest.c >&5 conftest.c:65:23: error: lcms/lcms.h: No such file or directory
Mine is at /usr/include/lcms.h (and it's found). Shouldn't that be enough to satisfy configure? I doubt anyone will have both.
Yes it's enough, only one of them needs to be found. What you are probably lacking is the .so symlink, the new code is more strict about this to make sure we always find the correct library soname.
This is a dir listing: [root(a)penguin usr]# ll lib*/*lcms* lrwxrwxrwx 1 root root 17 2007-06-19 16:51 lib64/liblcms.so -> liblcms.so.1.0.16 lrwxrwxrwx 1 root root 17 2007-06-18 17:35 lib64/liblcms.so.1 -> liblcms.so.1.0.16 -rwxr-xr-x 1 root root 206032 2007-02-08 12:21 lib64/liblcms.so.1.0.16 lrwxrwxrwx 1 root root 17 2007-06-19 16:51 lib/liblcms.so -> liblcms.so.1.0.16 lrwxrwxrwx 1 root root 17 2007-06-18 17:35 lib/liblcms.so.1 -> liblcms.so.1.0.16 -rwxr-xr-x 1 root root 206720 2007-02-08 12:20 lib/liblcms.so.1.0.16 It's the first time I actually looked deeper into configure.ac/configure so bear with me. This part look suspicious in configure.ac: 832 WINE_NOTICE_IF([test "$ac_cv_lib_lcms_cmsOpenProfileFromFile" = "yes"],[liblcms development files not found. 833 Wine will be built without Color Management support. (mscms.dll)]) Shouldn't that be != "yes" ? Cheers, Paul.
Paul Vriens <paul.vriens.wine(a)gmail.com> writes:
It's the first time I actually looked deeper into configure.ac/configure so bear with me. This part look suspicious in configure.ac:
832 WINE_NOTICE_IF([test "$ac_cv_lib_lcms_cmsOpenProfileFromFile" = "yes"],[liblcms development files not found. 833 Wine will be built without Color Management support. (mscms.dll)])
Shouldn't that be != "yes" ?
Yes of course, thanks for catching this. -- Alexandre Julliard julliard(a)winehq.org
On Wednesday 11 July 2007, Alexandre Julliard wrote:
It's the first time I actually looked deeper into configure.ac/configure so bear with me. This part look suspicious in configure.ac:
832 WINE_NOTICE_IF([test "$ac_cv_lib_lcms_cmsOpenProfileFromFile" = "yes"],[liblcms development files not found. 833 Wine will be built without Color Management support. (mscms.dll)])
Shouldn't that be != "yes" ?
Yes of course, thanks for catching this.
libicu detection fails here on Fedora development. configure used to check libsicu and libicu, now it only checks libsicu. The patch below makes the check succeed here. -Hans diff --git a/configure.ac b/configure.ac index 227e410..3134d66 100644 --- a/configure.ac +++ b/configure.ac @@ -796,7 +796,7 @@ dnl **** Check for the ICU library **** if test "$ac_cv_header_unicode_ubidi_h" = "yes" then saved_libs="$LIBS" - icu_libs="-lsicuuc -lsicudata -lstdc++ -lgcc_s" + icu_libs="-licuuc -licudata -lstdc++ -lgcc_s" if test -n "$ICU_LIB_DIR" then icu_libs="-L$ICU_LIB_DIR $icu_libs"
Hans Leidekker wrote:
On Wednesday 11 July 2007, Alexandre Julliard wrote:
It's the first time I actually looked deeper into configure.ac/configure so bear with me. This part look suspicious in configure.ac:
832 WINE_NOTICE_IF([test "$ac_cv_lib_lcms_cmsOpenProfileFromFile" = "yes"],[liblcms development files not found. 833 Wine will be built without Color Management support. (mscms.dll)])
Shouldn't that be != "yes" ? Yes of course, thanks for catching this.
libicu detection fails here on Fedora development. configure used to check libsicu and libicu, now it only checks libsicu.
The patch below makes the check succeed here.
-Hans
diff --git a/configure.ac b/configure.ac index 227e410..3134d66 100644 --- a/configure.ac +++ b/configure.ac @@ -796,7 +796,7 @@ dnl **** Check for the ICU library **** if test "$ac_cv_header_unicode_ubidi_h" = "yes" then saved_libs="$LIBS" - icu_libs="-lsicuuc -lsicudata -lstdc++ -lgcc_s" + icu_libs="-licuuc -licudata -lstdc++ -lgcc_s" if test -n "$ICU_LIB_DIR" then icu_libs="-L$ICU_LIB_DIR $icu_libs"
It appears that was deliberately: http://source.winehq.org/git/wine.git?a=commit;h=bb168aaba84fe0584fb0be9a0e0... I'm having the same problem on my F7 btw. Cheers, Paul.
participants (4)
-
Alexandre Julliard -
Hans Leidekker -
Lei Zhang -
Paul Vriens