This is the mail archive of the
mailing list for the GCC project.
Re: better common debug info
Actually, the -fdebug-kludge option isn't really the place to "fix"
debugging. That code should go away.
See ffecom_member_phase2_ for where I once thought the new code should
go, long time ago. Whether I'm still right, I don't know, but you
might recognize the code. :)
I hesitate to get involved in this at this time, though, because I
know there are all sorts of grody corner issues, many of which I
probably won't remember until I take the time to look at this stuff.
I need to at least submit test cases for them, and can perhaps do that
soon regarding whatever I *can* remember. (E.g. SUBROUTINE X(A);
COMMON /FOO/ N; REAL A(N); and variations thereof.)
Also, I tend to feel I should focus more on dealing with my pending-bug
backlog, which is getting reduced to the point where I should
start focusing on reviving my Alpha and looking at g77 Alpha bugs
OTOH, solving the longstanding "COMMON/EQUIVALENCE -g problem" in g77
would make lots of people very very happy. I think Andy Vaught started
looking into this on the g77-alpha list a couple of months' back -- if
you like, I'll pull up his emails and forward the pertinent ones to
tq vm, (burley)