Re: Small fix of column number in parser

[Jordi Gutiérrez Hermoso]

On 5 April 2011 04:17, Daniel Kraft <address@hidden> wrote:

> attached is a hg changeset that fixes wrong column numbers.  If the
> attached test.m is executed without the patch, one gets:
>
> error: bar
> error: called from:
> error:   /tmp/test.m at line 2, column -1
>
> With the patch, the column number is correctly reported as 1.

I tested your patch and it worked, so I pushed it to the default
branch:

    http://hg.savannah.gnu.org/hgweb/octave/rev/fd367312095a

> Do you prefer this simple patch or should I work on some "clean-up"
> of this situation, too -- like making skip_white_space a member of
> stdio_stream_reader and unifying column- and line-number tracking in
> one place.

I think for a first patch, a minimal solution is fine. Leave the
refactoring for later.

> PS: Since this is my first attempt at both working with Mercurial
> and submitting a patch for octave, please let me know if there's
> something I should do different (and in general, comments welcome).

You used mq; please submit finished patches (commits) next time, and
use a meaningful commit message. You also used our old-style ChangeLog
commit format, which we recently changed. Put the same information
that you put in the ChangeLog in the commit message. Here are the
guidelines for doing so:

    
http://octave.1599824.n4.nabble.com/policy-for-pushing-changesets-tp3443512p3445355.html

>  In particular, is there some tag I should add to messages sent with
> patches?  E.g., on the GCC lists, we have "[patch]" in the subjects
> of messages containing diff's.

Yes, please use our patch tracker. Easier to have them all in one
place instead of scattered in our respected inboxen:

    https://savannah.gnu.org/patch/?group=octave

Thanks,
- Jordi G. H.