[PATCH] opener: add self._audit (issue2862)

Idan Kamara idankk86 at gmail.com
Wed Jun 22 11:26:18 CDT 2011


On Wed, Jun 22, 2011 at 12:35 PM, Adrian Buehlmann <adrian at cadifra.com> wrote:
>
> On 2011-06-22 10:54, Idan Kamara wrote:
> > On Wed, Jun 22, 2011 at 11:39 AM, Adrian Buehlmann <adrian at cadifra.com> wrote:
> >> # HG changeset patch
> >> # User Adrian Buehlmann <adrian at cadifra.com>
> >> # Date 1308730424 -7200
> >> # Node ID 348c9f997718f66661190e1640db5b51c12a76e4
> >> # Parent  8deebb5777511369a852e5cb42e69844c40ebf7a
> >> opener: add self._audit (issue2862)
> >
> > It might have been nicer to see a longer commit message
> > describing how this actually fixes the issue. Because for
> > someone (like myself) who didn't investigate this bug report,
> > it's hard to understand what's going on.
>
> Yeah. Perhaps I've been a bit too long in this opener business to do yet
> another rain dance here.
>
> Plus wasting time on messages on patches that potentially get refused
> anyway isn't that funny. So at times I get lazy and just try without yet
> another cliff notes that probably no one wants to read anyway. I didn't
> expect that GSOC students are interested in scrutinizing patches that
> fix issues in contrib which affect Windows only.

I may be a GSoC student but my interests in Mercurial go beyond
my personal project.

>
> So let's have the cliff notes here then.
>
> My patch passes the testsuite on Ubuntu (pyflakes installed) and Pascal
> Quantin reported that it fixes his issue.
>
> The point is, shrink-revlog.py uses opener in rather unorthodox ways and
> it gives a hint that it doesn't want to care about auditing paths by
> setting audit=False when calling.
>
> I used that to disable the check, which is raising the exception that
> causes the traceback documented in the bts entry.

Thanks for the explanation, that's a lot clearer now.

>
> Feel free to take over this bug and patch and rewrite the change message
> or write a better fix.

I was merely curious about this change and thought it could use
a longer commit message (like most commits lately, including
mine which I'm trying to improve), that's all.


More information about the Mercurial-devel mailing list