with buildroot git version from today . same config of bug #635 (https://bugs.busybox.net/attachment.cgi?id=2269) I have to make cd buildroot/output/build/libglib2-2.24.1/ vi gio/Makefile and add -ldl to LIBS LIBS = -lintl -ldl same for gio/tests/Makefile for libxml2-2.7.7 happens the same. cd buildroot/output/build/libxml2-2.7.7/ I had to add -ldl to LIBS on Makefile doc/examples/Makefile example/Makefile
I'm sorry but with exactly your configuration, I'm able to build both libglib2 and libxml2 properly, without any of the modifications you've done. Could you please to a complete rebuild from scratch (removing everything !) and put the build log somewhere ? Adding -ldl manually to the Makefiles is not a proper fix, the Makefiles are generated files, and the fact that -ldl is needed or not should be detected by the ./configure script.
(In reply to comment #1) > I'm sorry but with exactly your configuration, I'm able to build both libglib2 > and libxml2 properly, without any of the modifications you've done. > > Could you please to a complete rebuild from scratch (removing everything !) and > put the build log somewhere ? I reported, just after do a make clean , make clean is that enough ? > Adding -ldl manually to the Makefiles is not a proper fix, I know that ... > the Makefiles are > generated files, and the fact that -ldl is needed or not should be detected by > the ./configure script. ./configure is the file that I should try to patch ?
> I reported, just after do a make clean , make clean is that enough ? 'make clean' is indeed enough. Please put the full build log online somewhere. As I can't reproduce the issue, if you don't provide a full build log so that I can start the investigation, I will unfortunately have no other option than closing this bug because of missing information. You can also try to reproduce by disabling BR2_CONFIG_CACHE (by default it is enabled). I'm fairly strongly suspecting a configure cache issue. > > the Makefiles are > > generated files, and the fact that -ldl is needed or not should be detected by > > the ./configure script. > > ./configure is the file that I should try to patch ? No. configure is also generated. For the error you're describing, there should be nothing to patch. It should just work.
(In reply to comment #3) > > I reported, just after do a make clean , make clean is that enough ? > > 'make clean' is indeed enough. Please put the full build log online somewhere. > As I can't reproduce the issue, if you don't provide a full build log so that I > can start the investigation, ok , I will put it don't worry , I am in holidays and this is only a hobby ... give me 2 days please. > I will unfortunately have no other option than > closing this bug because of missing information. > > You can also try to reproduce by disabling BR2_CONFIG_CACHE (by default it is > enabled). I'm fairly strongly suspecting a configure cache issue. ok I will check that. thanks,
(In reply to comment #4) > (In reply to comment #3) > > > I reported, just after do a make clean , make clean is that enough ? > > > > 'make clean' is indeed enough. Please put the full build log online somewhere. > > As I can't reproduce the issue, if you don't provide a full build log so that I > > can start the investigation, > > ok , I will put it don't worry , I am in holidays and this is only a hobby ... > give me 2 days please. > > > I will unfortunately have no other option than > > closing this bug because of missing information. > > > > You can also try to reproduce by disabling BR2_CONFIG_CACHE (by default it is > > enabled). I'm fairly strongly suspecting a configure cache issue. > > ok I will check that. After disable BR2_CONFIG_CACHE, I don't have this issue anymore ...
closing now that we have disabled shared config.cache by default.