jkesselm commented on PR #124:
URL: https://github.com/apache/xalan-java/pull/124#issuecomment-1815659399

   > Please discuss this policy with your fellow Xalan developers.
   
   Interesting thought. I can bring it up for discussion. It would mean 
changing a practice that dates back to when we first put this code onto Git, I 
think, so there's inertia as well as personal taste/experience in the mix. 
   
   Some people are violently offended if we _don't_ squish, though I don't 
think any of the current committers is that dogmatic. 
   
   Personally, I lean toward merging small changes frequently, so a squish is 
usually minor and just eliminates some of the typo/thinko corrections or 
intermediate checkpoints that aren't intended to be shared.  This Maven 
switchover is a major exception since, like an architectural change, it has a 
large all-or-nothing element to it -- or at least I didn't see a good way to 
make it incremental at the time.
   
   The early history of the Maven cut-over is squished largely because we got a 
demand to clearly separate the files which were just being moved from the files 
that were being changed/replaced. I decided to comply rather than argue about 
it, and I had to collapse it to achieve that resequencing. I can understand the 
ask, though I think the need was overstated, especially as the arguably easier 
review enabled by that separation doesn't actually seem to have produced much 
additional review feedback. 
   
   I see the point about fast-forward. That may be the most convincing argument.
   
   In any case: Normally I'd only squish upon merge into the main stream. Which 
should normally be after regression testing has been passed, unlike the 
individual commits. 
   
   The more serious continuous integration system we had running for Websphere 
Liberty still worked in terms of PRs rather than commits, since that was the 
level at which automatically backing out interacting or otherwise suspicious 
changes was practical. Backing out individual commit points would have left the 
system between stable versions, since non-squashed commits are *not* promised 
to be complete enough or isolated enough to accept individually.
   
   Thanks for raising it; I need to think about it more.
   
   Y'know, that sort of thing really ought to be going onto the Xalan 
development mailing list rather than here, since it isn't especially unique to 
this changeset. That's also the easiest way to get it in front of all the 
developers.
   
   (You do realize that if you continue being active and helpful, you're at 
risk of us trying to pull you into the Xalan project, right?)


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to