Re: Question on warnings for compiling gcc from svn source

1 msg	RETRACTED: Precompiled headers making for some ...
3 msg	can i cross compile glibc then copy it to the h...
1 msg	Precompiled headers making for some odd trouble...
3 msg	Similarities between gcc and binutils
1 msg	Tracking down a difference in test_results for ...
2 msg	autovectorization and aligned access
3 msg	any companies provide services for building gcc...
2 msg	Solaris 64 bit build environment

Question on warnings for compiling gcc from svn...
\ NightStrike (26 Jun 2007)
. \ David Daney (26 Jun 2007)
. . \ NightStrike (26 Jun 2007)
. . . \ David Daney (26 Jun 2007)
. . . . \ NightStrike (26 Jun 2007)
. . . . . \ David Daney (26 Jun 2007)
. . . . . . \ NightStrike (27 Jun 2007)
. . . . . . . \ David Daney (27 Jun 2007)
. . . . . . \ NightStrike (28 Jun 2007)

4 msg	How to use GCC testsuite..?
3 msg	option to see the Intermediate repersantation (...
4 msg	c++ const aggregate vs constructor
2 msg	varargs
1 msg	Re: relation between gcc/glibc version and linu...
3 msg	The (black?) art of inlining...
3 msg	order of -L options created by gcc for the linker
1 msg	loop peeling in autovectorization
3 msg	Building gcc 4.2.0 cross-compiler (host=i386-ap...
5 msg	Building only C++ libraries
1 msg	Can we use GCC testsuite to test other C compil...

Subject:	Re: Question on warnings for compiling gcc from svn source
Group:	Gcc-help
From:	NightStrike
Date:	28 Jun 2007

On 6/26/07, David Daney <ddaney> wrote:
> NightStrike wrote:
> > On 6/26/07, David Daney <ddaney> wrote:
> >> NightStrike wrote:
> >> > On 6/26/07, David Daney <ddaney> wrote:
> >> >> NightStrike wrote:
> >> >> > I am building a cross compiler, and I'm not good at it, so it
> >> requires
> >> >> > many tries with different options. The outcome is that I spend a
> >> lot
> >> >> > of time looking at output from the various compiles I keep doing. I
> >> >> > am curious.. there are a number of warnings for things that seem
> >> >> > trivial to fix... Does anyone go and fix these things? For
> >> instance,
> >> >> > the file libiberty/argv.c has a variable "int ret" defined at
> >> line 313
> >> >> > that is never used, and this triggers a warning. Sure enough, the
> >> >> > variable is never used (that I can see from the source). Are things
> >> >> > like that monitored and cleaned up?
> >> >> Being open source software, GCC relies on volunteers. You can do a
> >> >> couple of things:
> >> >>
> >> >> 1) Submit a patch to correct the problem following the procedures
> >> on the
> >> >> gcc.gnu.org web site.
> >> >> 2) File a bug report in the GCC bugzilla at the same site.
> >> >>
> >> >> Failing either of those options, someone that sees your message might
> >> >> fix it.
> >> >>
> >> >> In any event thanks for reporting it.
> >> >>
> >> >> David Daney.
> >> >
> >> > Ah, and here I thought that everyone knew about those things. If it's
> >> > policy to get things like that cleaned up, then I'd love to help.
> >> > I'll see if I can figure out how to create and submit a patch (I'm
> >> > assuming I just make the change and run the diff tool with certain
> >> > options..) Hopefully someone will be willing to help me through it :)
> >>
> >> Please read this document:
> >>
> >> http://gcc.gnu.org/contribute.html
> >>
> >> If you follow the suggestions there, things should go fairly smoothly.
> >> For a one line patch, you probably don't need a copyright assignment.
> >
> > Ok, I have the output from "svn diff" using my own svndiff.sh diff
> > script. Do I need a Changelog entry? If so, I'm a little unclear
> > from reading that page of what I'm supposed to do with it. I know I
> > don't post a Changelog diff, so do I just post my addition alongside
> > the patch?
> >
> > Here's the svn diff output:
> >
> > Index: libiberty/argv.c
> > ===================================================================
> > --- libiberty/argv.c (revision 126005)
> > +++ libiberty/argv.c (working copy)
> > @@ -310,7 +310,6 @@ writeargv (char **argv, FILE *f)
> >
> > while (*argv != NULL)
> > {
> > - int ret;
> > const char *arg = *argv;
> >
> > while (*arg != EOS)
> >
> >
> > Now I should put this in a txt file and attach it to a message that I
> > send to gcc-patches, correct?
>
> Essentially.
>
> You need a ChangeLog entry for the patch. The ChangeLog must be
> properly formatted and follow the rules for said entries. You should
> also say how the patch was tested.

I just entered a bug report real quick listed as "trivial":

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32532

It looks like someone already replied to commit it, so I guess I don't
have to follow up with a post to gcc-patches. Comments?