On May 20, 2012, at 11:47 PM, Robert Spier wrote:

> Please combine this into the previous patch.  Actually, I applied it
> anyway, see below.

Most of the patches that are related, or are very big, are in their own branch, 
and I've pushed the branches to github as well.  Perhaps it would be easier to 
pull them in from there, where the branch associations are intact?  But then, 
there seems to be issues with that too. 

> These patch dumps are really hard to deal with.  They often contain
> multiple overlapping patches all dealing with similar things.  Also, not
> all the patches are making it through.  In the recent set [PATCH 2/13] is
> missing (I double checked the list archive).

The missing ones shouldn't matter. There's some changes I know won't (or 
shouldn't) get merged upstream (like adding extra plugins to 
config.sample/plugins).  When using "git send-email", I just selectively choose 
not to send those. 

> There's definitely a sweet spot between big commits and small commits, that
> we need to find... but with the very limited time I have to apply things...
> I think that larger (but still focused) commits are a little easier.

I've erred on the side of bigger. But then there's often follow-up patches, 
because after pushing them up to github, Travis complains. Or someone makes 
suggestions or comments that I incorporate.  But 'git send-email' doesn't 
preserve the branches, and merging/squashing patches is EVIL in all caps. While 
developing, I push my changes to github for testing via Travis. If I later 
merge them so I can send 'em to you merged, then my repo is b0rked and I can't 
ever push that repo to github. Icky, icky, icky.

> Anyway,
> I've gone through most of the patches and applied most things that just
> worked from mutt with git am.  The ones that didn't get applied most likely
> wouldn't, and I didn't have time to figure out why.  Likely an ordering
> issue.
> 
> I don't want to discourage you from sending patches -- this is the most
> interest anyone has had in qpsmtpd in a while, and we definitely appreciate
> it.

I'll rebase and resubmit. 

Matt

Reply via email to