From: Gabriel Kerneis <kerneis@pp...> - 2011-10-28 13:58:09
Dear CIL users,
I need some testing from people using cilly (and gcc) on large code bases with
while preparing the release of CIL 1.4 (due next week), if fixed two old bugs
The first one (cilly changes LDFLAGS order) should not introduce any
regression. It was just cilly trying to be too smart and resolving -lfoo into
-l/some/path/libfoo.a for some reason that I do not understand (and doing it
wrong, of course).
The second one, on the other hand, was again cilly trying to be too smart but
for some good reason: it stripped "-Wl," from linker parameters so that ld
understood them. The issue is that it used to do so even when calling gcc,
instead of doing it only for ld (and for ar?).
The new situation is that "-Wl," is preserved all the time; this might break
compilation if ld is involved somehow. Since I have no test case where ld is
used, I cannot try and find a complete fix. So, if you could test those two
patches and send me bug reports, it would help a lot.
Thanks in advance,
Get latest updates about Open Source Projects, Conferences and News.