Ralf Wildenhues wrote:
I suggest this patch to fix the export test on MinGW. It did not fail on Cygwin due to auto-import, but on MinGW it did for the data objects.
That is odd, because mingw supports auto-import too. (However, it might be "off" by default, since libraries created that way wouldn't be usable by msvc, which is one of the goals of mingw, at least for non-name-mangled languages.) So, yeah, being a little more careful about declspec() niceties for mingw is probably a good thing.
WDYT? I guess stresstest.at needs a similar fix, but there the same code is currently used to link against a shared and a static version of the library, so that would need to be fixed as well.
The only thing that's then still worrying me is that on Cygwin, the mdemo and mdemo_static programs sometimes throw segmentation faults on my system. Not all the time though.
Well, it's failing all the time for me, but I'm not sure it's a segfault. What does "Hangup" mean, when reported by the shell after executing the app:
$ tests/mdemo/mdemo_static tests/mdemo/foo1.la tests/mdemo/libfoo2.la Welcome to GNU libtool mdemo! module name: foo1 module filename: foo1.a module reference count: 1 ** This is foolib 1 ** hello returned: 57616 hello is ok! cos (0.0) = 1 sub() called foo1 is ok! module name: libfoo2 module filename: libfoo2.a module reference count: 1 ** This is foolib 2 ** hello returned: 57616 hello is ok! sin (0.0) = 0 sub() called foo2 is ok! myfunc returned: 57616 myfunc is ok! Hangup $ echo $? 129
And then on both Cygwin and MinGW test 16 (Link order of deplibs) still fails. I guess I need to weaken it even further...
I see this, too. I'll post a more thorough test report on the full test suite on cygwin, tomorrow evening.