Some other problems ...
Paul Millar
paulm at astro.gla.ac.uk
Thu Oct 31 12:21:42 CST 2002
Hi everyone,
Just a couple of problems I've noticed. They look like they're due to
some broken build dependencies.
1. I have a copy of wine on my laptop which I do "cvs update" every few
days. After updating today, I did the usual "./configure
--enable-opengl && make depend && make" and left it to compile. But,
when I tried the result, I got a client/server mismatch for
SERVER_PROTOCOL_VERSION. One was claimed to be v88 whilst the other
was v89. Doing a "make distclean" then repeating the build "fixed"
the problem.
Patch http://cvs.winehq.com/patch.py?id=6064 (committed last Tuesday)
bumped the SERVER_PROTOCOL_VERSION up to 89. So it looks like some of
the binaries weren't updated during the build. I tried altering the
version (manually) and rebuilding, but wine rebuilt correctly.
2. WRT hit a problem last night where "make test" failed. The
log included messages like:
gcc -c -I. -I. -I../../../include -I../../../include -g -O2 -Wall
-mpreferred-stack-boundary=2 -fPIC -DSTRICT -DNONAMELESSUNION
-DNONAMELESSSTRUCT -D_REENTRANT -o registry.o registry.c
LD_LIBRARY_PATH="../../../library:../../../unicode:$LD_LIBRARY_PATH"
../../../tools/winebuild/winebuild -fPIC -DSTRICT -DNONAMELESSUNION
-DNONAMELESSSTRUCT -o advapi32_test.exe.spec.c -exe advapi32_test.exe
-mcui registry.o testlist.o -L../../../dlls -ladvapi32 -lkernel32
-lntdll
gcc -shared -Wl,-Bsymbolic advapi32_test.exe.spec.o registry.o testlist.o
-o advapi32_test.exe.so -L../../../library -lwine -lm
advapi32_test.exe.spec.o: In function `__wine_exe_main':
/home/paulm/WINE-cvs/wine/dlls/advapi32/tests/advapi32_test.exe.spec.c:116:
undefined reference to `main'
registry.o: In function `test_enum_value':
/home/paulm/WINE-cvs/wine/dlls/advapi32/tests/registry.c:69: undefined
reference to `winetest_set_ok_location'
/home/paulm/WINE-cvs/wine/dlls/advapi32/tests/registry.c:69: undefined
reference to `winetest_ok'
/home/paulm/WINE-cvs/wine/dlls/advapi32/tests/registry.c:78: undefined
reference to `winetest_set_ok_location'
[... etc ...]
The later build had six tests that succeded. These were the new scanf
tests introduced into that build by patch
http://cvs.winehq.com/patch.py?id=6107
I've just done a "make distclean" followed by a normal rebuild and
test. Its currently running through the regression tests, but so far
it seems to be working.
These are vague reports of unrepeatable problems ... never a good thing!
I'm quite busy right now (real work and everything), so I don't have time
to investigate this myself. Just posting this hoping the cause its obvious
to someone else :^)
----
Paul Millar
More information about the wine-devel
mailing list