Re: [hatari-devel] MODE SENSE patch

[Uwe Seimet <Uwe.Seimet@xxxxxxxxx>]

Hi,

> Well, if you squash everything together into one patch, it gets more
> complicated for the maintainers of a project - review is more difficult,
> bi-secting is more difficult later for debugging ... thus every non-trivial
> project that I know (and I'd say that Hatari is non-trivial, too) requires
> contributors to split up non-trivial patches. But maybe I'm also just
> contributing to the wrong projects ;-)

There are different approaches, and one size does not fit all. My experience
with many small patches is that some reviewers complain about too many small
PRs and start asking for more changes within one PR. Or, if you have many
small PRs that step by step contribute to a major change, if it takes too
long for these PRs to be reviewed, the overall progress stalls, and you
may get bogged down. I also know cases where even though the PRs addressed
different issues they partially referred to the same code locations. The
result was a hell of merge conflicts.
A lot depends on the nature of the project/change, the persons involved,
how many of them are able to review your changes, and how profound the
reviews are.

> > It took very long for my first patch to
> > be reviewed, and at that time there were already other minor changes in my
> > pipeline.
> 
> It's perfectly fine to juggle multiple patches in a git branch.
> Interactive rebases ("git rebase -i") are your friends! 

Well, there are good and bad jugglers ;-).
 
> > Anyway, please just discard all of my recent patches.
> 
> Fortunately, the hunks in your big patch were pretty much independent, so I
> was at least able to split out the fix for mode sense in there.

Thank you for spending your time on these patches.

Best regards

Uwe