Differences between revisions 1 and 36 (spanning 35 versions)
Revision 1 as of 2014-04-12 20:30:48
Size: 338
Comment:
Revision 36 as of 2016-03-25 17:03:39
Size: 4826
Comment:
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
#pragma section-numbers 2

<<Include(A:style)>>
<<Include(A:dev)>>
Line 3: Line 8:
This page explain the Mercurial Patch Review Process and how (anyone) can help. This page explains the Mercurial patch review process and how (anyone) can help.
Line 5: Line 10:
 * all review happen one MailingLists#The_Mercurial-Devel_list <<TableOfContents>>
Line 7: Line 12:
 * Contributor follow the ContributingChanges and send they patch to the list (hopefully using the PatchbombExtension) == Generic Fact ==
Line 9: Line 14:
 * Everybody is welcome to do review.  * All reviews happen on MailingLists#The_Mercurial-Devel_list

 * Contributors follow the ContributingChanges and send their patch(es) to the list (hopefully using the PatchbombExtension)

 * Reviews are just email replies to the emailed patch

 * ``Everyone`` is welcome to do review.

== Simple Review Checklist ==

 * The patch should conform to the ContributingChanges bullet list.

   Quick reminder of important things:

   * commit message format,

   * Patch does one and only one thing,

   * Change is tested

   * Documentation augmented and updated

   * (all the other things in the list)

 * You understand the change

 * The change seems correct

 * The change seems efficient

If any concerns raised, reply to the email asking questions.

If everything sounds good, reply to the email too. Just state it looks good to you in your reply. To make your pre-review even more useful, don't forget to go to [[http://patchwork.serpentine.com/project/hg/list/|Patchwork]] and mark the patches as "Pre-Reviewed".

== Things we commonly miss ==

(we should probably move these recommendation in other page and just link to them)

 * Good 'topic'. (The part is the one before colon in the first line `topic: short desc`). It is unvaluable to sort thing out when scanning through commit, especially when building a release changelog. A common mistake is to pick a very general work like "commands".

 * Config section. All debatable/temporary/unsure-we-want-this config should go to the `[experimental]` config section. For other option, try to avoid adding a new section if we can't foresee more than one option in it and another pre-existing section would be a good fit.

 * Deprecation warning, if a major internal API get killed, encourage the preservation of the old version for 1 version with a deprecation warning (if it is easy to implement). This makes third party extensions maintainer life easier.

== Accepters Review Checklist ==

Some people are blessed to accept patches and push them to a repo where Matt Mackall ultimately pulls from.

Here is another check list for them

 * Run check code on all patches

 * Run the whole test suites

 * Reply to the list saying that you took care of the patch

 * you can get the patches files directly from http://hgpatches.durin42.com/patches/<node> Appropriate hg alias would be:

  {{{
[alias]
getpatch=import --partial --obsolete http://hgpatches.durin42.com/patches/$1
}}}

 * Make sure you created obsolescence marker between the node in the patch and the one you created, e.g.

  {{{#!bash
hg import --partial --obsolete <patches>:
}}}


 * use the [[https://bitbucket.org/marmoute/mutable-history/src/98b5ac44a25913e546f4588dba5b7a1db39c82d4/hgext/drophack.py?at=default|drophack extension]] if you need to drop a changeset you queued

 * Rebase your queue on top of main's ```@```

 * Move ```@``` with the changeset you pushed.

 * Update [[http://patchwork.serpentine.com/project/hg/list/|Patchwork]] once you have pushed


== Patchwork States ==

||'''New''' || ||
||'''Pre-Reviewed''' ||non-reviewer have "lgtm", but still needs someone to look at it ||
||'''Under Review''' || ||
||'''2nd Review Requested'''||reviewer looked at it, but second pair of eyes requested ||
||'''Accepted''' || ||
||'''Changes Requested''' ||changes requested by reviewer, needs new version ||
||'''Rejected''' || ||
||'''RFC''' ||an RFC patch, needs more reviews (?) ||
||'''Superseded''' ||new version available ||
||'''Not Applicable''' ||not a patch (?) ||
||'''Deferred''' ||? ||


== Review Tooling ==

 * [[http://patchwork.serpentine.com/project/hg/list/|Patchwork]] ([[http://hgpatches.appspot.com/?days=30|backlog plot]])

 * [[http://hgpatches.durin42.com/|The Patches Bot]] ([[http://hg.durin42.com/patchbot/|sources]])

 * [[http://42.netv6.net/reviewtools/|Collection of script]]

 * Various data collection [[http://review.octopoid.net/]] (STALED)

 * Matt Mackall Inbox [[http://www.selenic.com/inbox|Metrix (nb email, nb patches, oldest email (in day))]] and [[http://www.selenic.com/inflight|Content]].

== The Committers Group ==

Current list with push access to the [[https://www.mercurial-scm.org/repo/hg-committed/|hg-committed repository]]

 * Kevin Bullock
 * Pierre-Yves David
 * Augie Fackler
 * Matt Mackall
 * Yuya Nishihara
 * Bryan O'Sullivan
 * Martin von Zweigbergk

----
CategoryDeveloper

{i} This page does not meet our wiki style guidelines. Please help improve this page by cleaning up its formatting.

Note:

This page is primarily intended for developers of Mercurial.

Patch Review Process

This page explains the Mercurial patch review process and how (anyone) can help.

1. Generic Fact

2. Simple Review Checklist

  • The patch should conform to the ContributingChanges bullet list.

    • Quick reminder of important things:
    • commit message format,
    • Patch does one and only one thing,
    • Change is tested
    • Documentation augmented and updated
    • (all the other things in the list)
  • You understand the change
  • The change seems correct
  • The change seems efficient

If any concerns raised, reply to the email asking questions.

If everything sounds good, reply to the email too. Just state it looks good to you in your reply. To make your pre-review even more useful, don't forget to go to Patchwork and mark the patches as "Pre-Reviewed".

3. Things we commonly miss

(we should probably move these recommendation in other page and just link to them)

  • Good 'topic'. (The part is the one before colon in the first line topic: short desc). It is unvaluable to sort thing out when scanning through commit, especially when building a release changelog. A common mistake is to pick a very general work like "commands".

  • Config section. All debatable/temporary/unsure-we-want-this config should go to the [experimental] config section. For other option, try to avoid adding a new section if we can't foresee more than one option in it and another pre-existing section would be a good fit.

  • Deprecation warning, if a major internal API get killed, encourage the preservation of the old version for 1 version with a deprecation warning (if it is easy to implement). This makes third party extensions maintainer life easier.

4. Accepters Review Checklist

Some people are blessed to accept patches and push them to a repo where Matt Mackall ultimately pulls from.

Here is another check list for them

  • Run check code on all patches
  • Run the whole test suites
  • Reply to the list saying that you took care of the patch
  • you can get the patches files directly from http://hgpatches.durin42.com/patches/<node> Appropriate hg alias would be:

    • [alias]
      getpatch=import --partial --obsolete http://hgpatches.durin42.com/patches/$1
  • Make sure you created obsolescence marker between the node in the patch and the one you created, e.g.
    • hg import --partial --obsolete <patches>:
  • use the drophack extension if you need to drop a changeset you queued

  • Rebase your queue on top of main's @

  • Move @ with the changeset you pushed.

  • Update Patchwork once you have pushed

5. Patchwork States

New

Pre-Reviewed

non-reviewer have "lgtm", but still needs someone to look at it

Under Review

2nd Review Requested

reviewer looked at it, but second pair of eyes requested

Accepted

Changes Requested

changes requested by reviewer, needs new version

Rejected

RFC

an RFC patch, needs more reviews (?)

Superseded

new version available

Not Applicable

not a patch (?)

Deferred

?

6. Review Tooling

7. The Committers Group

Current list with push access to the hg-committed repository

  • Kevin Bullock
  • Pierre-Yves David
  • Augie Fackler
  • Matt Mackall
  • Yuya Nishihara
  • Bryan O'Sullivan
  • Martin von Zweigbergk


CategoryDeveloper

ReviewProcess (last edited 2022-10-12 15:53:35 by AugieFackler)