D542: effectflag: document effect flag

[martinvonz (Martin von Zweigbergk)]

martinvonz added a comment.


  In https://phab.mercurial-scm.org/D542#12108, @quark wrote:
  
  > In https://phab.mercurial-scm.org/D542#11973, @lothiraldan wrote:
  >
  > > @quark Thank you for reposting your comment here, it's helpful to not lose track 
  > >  on all these discussions.
  > >
  > > We've had this experiment in Evolve for a little while now and the users using it
  > >  are giving positive feedback. The next step is to upstream it in core but turned off by
  > >  default like the others experiments and gather data and feedback.
  >
  >
  > I think the upstream has a higher barrier about implementation details. A lot of fb extensions need to be partially or completely rewritten before going upstream. I'm not sure if evolve is somehow an exception. But it does not feel quite right especially because your approach has known hard-to-solve defect and there are known better alternatives.
  >
  > > [...]
  > >  I propose that we start a separate experiment, so users could use it
  > >  separately from effect-flag. I would be happy to collaborate on a first
  > >  version of this experiment before the sprint and use the sprint to further
  > >  discuss the design, what do you think ?
  >
  > I'm not going to the sprint this time unfortunately. But if you have //concrete// examples where effect-flag is better in theory, I'd be happy to discuss them here.
  >
  >  ----
  >
  > Since you have admitted the defect without giving a solution, and did not response to the alternative I proposed, I'm still -1 on this series. I have marked it explicitly hopefully to make other reviewers' life easier.
  >
  > I think what I proposed is not hard to implement (I can complete it in 3 work days), so if you can help implement it instead, that'll be great and I'll be happy to provide technical help for any difficulty you encounter.
  
  
  I've found the information from the effectflag that's used in "hg obslog" useful. As you've probably seen, I've also sent a patch for recording the operation by default (implemented by Durham once upon a time, I think). I don't mind seeing turning both of these features on by default and we can let time tell which of them users find most useful and keep only that, or we can decide to keep both. I was about to queue this series, but it uses the old [evolution] config section, so at least that will have to change before I can accept it.

REPOSITORY
  rHG Mercurial

REVISION DETAIL
  https://phab.mercurial-scm.org/D542

To: lothiraldan, #hg-reviewers, quark
Cc: martinvonz, quark, mercurial-devel