On 26 February 2013 11:49, Jacob Nevins
<0jacobnk.fc...@chiark.greenend.org.uk> wrote:
> Marko Lindqvist writes:
>> [...] Should have added comment at least to remind why exactly so. [...]
> Where I work, we have a policy that every commit is accounted to a bug
> like here, but we have an exception for comment-only changes, on the
> grounds of having the barrier to improving comments be as low as
> possible (not requiring the overhead of raising a bug etc). We have a
> special pseudo-job "CommentChange" to associate such commit with. This
> generally seems to work for us and is not abused.
> I've noticed I have a reticence to just comment something I've found
> because of the bug overhead. Should we do something similar for Freeciv?
> Could be restricted to more "trusted" developers (e.g. those with svn
> access?), and extensive changes touching many files which require some
> explanation can of course continue to go in regular tickets.

 The reason I sometimes fail to do minor improvement tickets about
things I've noticed while working on something else is simply that my
source tree contains the main patch I'm working on, and thus cannot be
used to create the minor one. I then sometime forgot the minor thing
before have tree free for it (using git would help, I know). But I'm
not against making comment-only commits easier if that makes life
easier for you. I'd expect ticket still to be raised if there's
anything others may want to comment on (i.e., you are not absolutely
certain about your new comment) and commit message quality should
remain high (usually to contain explanation what was wrong with the
old comment)

 - ML

Freeciv-dev mailing list

Reply via email to