Le dimanche 31 juillet 2011 16:27:57, grischka a écrit : > Thomas Preud'homme wrote: > >> Can you show an example where this happens, that is that > >> > >> const_wanted is false > >> > >> and > >> > >> cur_text_section is NULL > >> > >> at the same time in cond_expr()? > > > > Sure. It's a side effect of the VLA array patches. On Hurd, there is a > > header with something like: > > > > #define FOO_COUNT (cond ? 2 : 42) > > > > type foo_array[FOO_SIZE]; > > > > This lead post_type to call gexpr() which will eventually call > > cond_expr() without const_wanted. This in turn will try to generate some > > code but will segfault due to cur_text_section being NULL (this is how I > > spotted it, it caused a segfault in the testsuite on Hurd) > > I still cannot reproduce this. Can you show a self contained > example to exhibit the bug? Mmmmh good point, I should have tried with mob. I can't reproduce it neither. So I guess master is not as close to mob as I thought. I revert the patch immediately and keep it in Debian (it's only use in Debian experimental) until I found which patch fixed (or hide) the issue (it's not gonna be easy since I heard git is not available on Hurd so no git bisect).
Best regards, Thomas Preud'homme > > --- grischka > > > I can give you the exact content of the header on Hurd if you need. > > > > Best regards, > > > > Thomas Preud'homme
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Tinycc-devel mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/tinycc-devel
