Re: [PATCH] tests: avoid spurious failure when gcj is not installed

[Stefano Lattarini]

On 02/25/2012 03:14 PM, Jim Meyering wrote:
> Stefano Lattarini wrote:
>> One ludicrously minor nit: we should put references to bug reports,
>> names of people to thanks, or old commits that introduced a regression
>> *before* the list of touched files, and always separated by a leading
>> and a trailing blank line; like this:
> 
> Adjusted and pushed.
> 
> I'll try to remember.
>
No worry anyway; I labelled it as a "ludicrously minor" nit for
a reason ;-)

But I should definitely improve HACKING and have it document the
standards and best practice for commit logs (since the GCS are sadly
weak and out-of-date in this regard).

> In other projects, I don't have a hard and fast policy for that,
> but more often put it after the ChangeLog entries with no separating
> blank line.  My thinking is that where the flaw was introduced is a
> relatively minor detail, which doesn't deserve its own paragraph.
> That is especially so here, since it was the preceding commit.
>
Yeah; the issue here is that I had other local commits between the
breaking commit and your fixing one, and I advanced my request
without thinking that I'd have finally rebased my local changes
anyway, so your change would have stayed an immediate follow-up.
So in hindsight I should have asked you to label it as a "fixup".
Oh well.

Regards,
  Stefano