Re: Time for a HP/PA hackathon?

[mec.gnu@mindspring.com (Michael Elizabeth Chastain)]

> The 64-bit off_t, enabling the TUI, the HP/UX frame update, or the BFD 
> file I/O rewrite?

The 64-bit off_t is the part that bothers me.

It's using a new host ABI that gdb hasn't used before, which means it's
likely to have problems on some hosts.

I don't feel as nervous about the TUI as you do.  If it doesn't work on
some host, then we can just disable it on that host.

The HP/UX frame update doesn't bother me at all.  We have a test suite.
I run it.  It's covered, although the coverage is limited to HP's
compilers (no gcc) and HP-UX 11.11 (no 10.20 or 11.00).

I don't know anything about the BFD file I/O rewrite.

We aren't getting enough host-side coverage in gdb-testers.  It's a
QA issue rather than a development issue.  How does anyone know if the
new off_t code or the most recent readline code works on Cygwin or
AIX or Irix or Solaris?

> I do wonder if the next binutils branch should be held off until after 
> GDB's been released - so that it can learn from GDB's mistakes :-)

It's easier for me if gdb doesn't branch until gcc 3.3.3 has released,
and a lot easier if binutils doesn't branch until gdb 6.1 has released
(preferably until after gdb gdb-6_1-branch has died).

Michael C