Re: [PATCH 01/17] Enable building with -Wshadow

[Andrey Smirnov <andrew.smirnov@gmail.com>]

> I don't think we want -ansi, though I am not sure.
> I think it would be good to separate out these patches so we can see.
>

I myself would rather had -std=c99 or even -std=gnu99, for that matter,
but
http://sourceware.org/gdb/current/onlinedocs/gdbint/Coding-Standards.html#Coding-Standards
said "ISO C90" and "ISO C90" it was. Anyway, as far as I can recall,
there were no fixes to gdb directories related to that flag and the only
two files required fixing were libiberty/_doprnt.c and bfd/peXXigen.c

> However -- I think the best overall approach would be to separate
> "pretty obvious" patches from ones that are not as obvious.  I know they
> are all just renamings, but I think some forms of shadowing (like purely
> within a single function) are much easier to review than others.  Also I
> expect some may be contentious, perhaps even the whole project will be.

Oh, boy, this is going to be one hell of a "git rebase --interactive"!
Well it looks like it is going to the "Santa Barbara" of patches,
nonetheless, OK, I'll regroup them by function and complexity, but I'll
group the ones fixing clashes with stdlib names, by function or variable
name to reduce number of patches, though.

>
> Andrey> - Does the part of MAINTAINERS file about patches to libiberty,
> Andrey>   libdecnumber and intl needing to be sent to GCC still holds true or
> Andrey>   should I send them to this list?  Bfd and binutils?
>
> Yes, libdecnumber and libiberty are canonically maintained in GCC.
> I don't know if they want -Wshadow or not.
>
> BFD patches should go to binutils.  I also don't know if they want -Wshadow.
>

 Well, since contributing to them require copyright assignment too, I'll
 contact them after it is done.

P.S. I stand corrected there was one change in gdb tree in
gdb/gdbserver/tracepoint.c
regarding "-ansi" option, but that is all, and I'll be separating it
to standalone patch anyway.