This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
Dan Kegel <dank@kegel.com> wrote:
http://tsukuba.m17n.org/linux-sh/ml/linux-sh/2003-11/msg00016.html describes a bug which shows up on sh4 in shared libraries that do a structure copy. A minimal test case is ready, as is a fix, and I'd like to submit patches for both. But I'm having trouble figuring out how to fit the regression test into the test framework. The testsuite seems to be made up of lots of standalone programs; how do you add a test to the testsuite that consists of two c files, one that produces a small shared library, and one that calls it?
Although I have no idea for the testsuite problem, doesn't
http://gcc.gnu.org/ml/gcc-patches/2003-12/msg00156.html
which was applied to 3.3.3 solve the original sh4 problem?
Yes, I think that's the patch you sent me. That together with http://gcc.gnu.org/ml/gcc-patches/2002-11/msg01641.html probably fix his problem. I just now verified that those two patches when applied to gcc-3.3.2 don't introduce any new regressions on sh4, and am about to verify they fix his problem. At some point soon I'll also test gcc-3.3.3 and gcc-3.4 snapshots, time permitting. (This is all a spare time project now; I no longer work for the company that originally had me supporting gcc on sh4.)
But it'd be nice to know how to submit a testcase that involves a shared library :-)
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |