pulkit (Pulkit Goyal)
Spy

Projects

User does not belong to any projects.

User Details

User Since
Jun 28 2017, 8:54 PM (55 w, 1 d)

Recent Activity

Wed, Jul 18

pulkit added a comment to D3901: histedit: add history-editing-backup config option.

And, should we drop the --no-backup option?

IIUC,
dropped -> user won't be able to use that option
deprecated -> can use, but this is not preferred option to use
Am I right above?

Wed, Jul 18, 3:16 PM
pulkit added a comment to D3901: histedit: add history-editing-backup config option.
In D3901#61668, @yuja wrote:

+coreconfigitem('ui', 'historyediting_backup',
+ default=True,
+)

coreconfigitem('ui', 'interactive',
    default=None,
)

diff --git a/hgext/histedit.py b/hgext/histedit.py

    • a/hgext/histedit.py +++ b/hgext/histedit.py @@ -1111,7 +1111,8 @@ fm.startitem() goal = _getgoal(opts) revs = opts.get('rev', [])
  • nobackup = opts.get('no_backup') + nobackup = (opts.get('no_backup') or + not ui.configbool('ui', 'historyediting_backup'))

history-editing-backup per new rule.

https://www.mercurial-scm.org/wiki/UIGuideline#config

Can you add # experimental config: ui.history-editing-backup to silence
check-config? It's probably too late to add full support for this option
and make it documented.

And, should we drop the --no-backup option? @pulkit what do you think?

Wed, Jul 18, 3:14 PM

Mon, Jul 16

pulkit accepted D3950: shelve: pick the most recent shelve if none specified for --patch/--stat.
Mon, Jul 16, 5:59 PM
pulkit accepted D3949: shelve: improve help text for --patch and --stat.
Mon, Jul 16, 5:58 PM

Wed, Jul 11

pulkit planned changes to D3670: cmdutil: don't return statedetectionpredicate function from _getrepostate().
Wed, Jul 11, 5:36 PM
pulkit planned changes to D3669: cmdutil: have statefile names in STATES instead of functions.

I won't have time before the upcoming freeze to get this done.

Wed, Jul 11, 5:35 PM
pulkit updated subscribers of D3672: retractboundary: add dryrun parameter.
Wed, Jul 11, 5:24 PM
pulkit added a comment to D3919: grep: restore pre-9ef10437bb88 behavior, enable wdir search by tweakdefaults.

I think it would make sense to defer the behavior change until we test some more. That said, I do want us to plan to make an intentionally breaking change with the out of the box grep experience *without* tweakdefaults enabled. My reasoning is more or less this: approximately nobody uses hg grep today[0] because its behavior doesn't match what users expect out of the box. I'm willing to call the current behavior a *bug*, and prominently announce that with 4.8 we expect to change the default behavior of hg grep to match user expectations. That's in line with the decision outlined in [1] which is old enough that mpm was part of the decision. I strongly suspect hglib will be the only meaningful breakage, and we can patch hglib to pass the -r 0:tip flag that'll give consistent behavior across all hg versions.

How do we feel about that?

Wed, Jul 11, 2:43 PM
pulkit accepted D3926: scmutil: fix shortesthexnodeidprefix on Python 3 for 0-prefixed nodes.
Wed, Jul 11, 2:01 PM
pulkit accepted D3925: tests: add missing b prefix in test-context.py.
Wed, Jul 11, 2:00 PM
pulkit accepted D3924: manifest: just duplicate the definition of items as iteritems.
Wed, Jul 11, 2:00 PM
pulkit accepted D3923: tests: properly sort imports in test-nointerrupt.t.
Wed, Jul 11, 1:59 PM
pulkit accepted D3922: revlog: adjust doctest examples to be portable to Python 3.
Wed, Jul 11, 1:58 PM
pulkit accepted D3921: py3: whitelist 3 more passing tests caught with the ratchet.
Wed, Jul 11, 1:57 PM
pulkit added a comment to D3919: grep: restore pre-9ef10437bb88 behavior, enable wdir search by tweakdefaults.

So what is the way around this, does this mean grep can't be changed at all?
Also can't we update hglib to work accordingly.

Wed, Jul 11, 1:54 PM
pulkit accepted D3920: grep: hide cryptic revision number by default.
Wed, Jul 11, 1:42 PM
pulkit accepted D3919: grep: restore pre-9ef10437bb88 behavior, enable wdir search by tweakdefaults.
Wed, Jul 11, 1:39 PM
pulkit accepted D3918: grep: add config knob to enable/disable the default wdir search.
Wed, Jul 11, 1:36 PM
pulkit accepted D3917: grep: search working directory files by default if --all-files is specified.
Wed, Jul 11, 1:35 PM
pulkit accepted D3916: grep: reject --diff --all-files.
Wed, Jul 11, 1:34 PM
pulkit accepted D3915: grep: rename --allfiles to --all-files.
Wed, Jul 11, 1:33 PM

Mon, Jul 9

pulkit planned changes to D3866: histedit: use cbor to write histedit-state file.

I am not confident to land some state format changes at this point in the cycle. Let's do it early in next cycle.

Mon, Jul 9, 2:31 PM
pulkit closed D2874: remotenames: synchronise remotenames after push also.
Mon, Jul 9, 10:45 AM
pulkit committed rHG4d5fb4062f0b: remotenames: synchronise remotenames after push also.
remotenames: synchronise remotenames after push also
Mon, Jul 9, 10:45 AM

Sun, Jul 8

pulkit added inline comments to D3896: copies: handle case when both merge cset are not descendant of merge base.
Sun, Jul 8, 5:14 AM
pulkit created D3896: copies: handle case when both merge cset are not descendant of merge base.
Sun, Jul 8, 5:12 AM

Thu, Jul 5

pulkit added a comment to D3887: rebase: add --no-backup option.
In D3887#60752, @yuja wrote:

The idea was to add --no-backup option in all those commands which alter
history. For this one, let say if someone wants to abort a rebase in between
then a --no-backup option should be available (same as histedit)

If he don't want any backup bundle on history editing, it might be better
to add a config knob to turn off the feature than forcing him to always
specify --no-backup.

Thu, Jul 5, 2:41 PM

Wed, Jul 4

pulkit accepted D3884: tests: use interfaceutil in simplestorerepo.
Wed, Jul 4, 3:30 PM
pulkit accepted D3872: histedit: add --no-backup option (issue5825).
Wed, Jul 4, 3:30 PM
pulkit accepted D3764: rebase: improve output of --dry-run.
Wed, Jul 4, 3:16 PM

Mon, Jul 2

pulkit added a comment to D3873: rebase: refactor logic to read rebasestate in a separate function.
In D3873#60501, @yuja wrote:

Queued, thanks.

f = repo.vfs("rebasestate")
for i, l in enumerate(f.read().splitlines()):
    if i == 0:
  • originalwd = repo[l].rev() + data['originalwd'] = repo[l].rev()

FWIW, I don't have any better idea to keep old hg not crash with new state
file, other than:

a. use separate file (e.g. "rebasestate2") and leave "rebasestate" in old
   format (or make it an empty file to trigger error.)
Mon, Jul 2, 2:18 PM
pulkit closed D3876: rebase: check whether the rebasestate exists or not a bit early.
Mon, Jul 2, 8:43 AM
pulkit closed D3875: rebase: use staeobj to check whether interrupted rebase exists.
Mon, Jul 2, 8:43 AM
pulkit closed D3874: rebase: add a stateobj variable to rebaseruntime class.
Mon, Jul 2, 8:43 AM
pulkit committed rHGcf24f678adda: rebase: check whether the rebasestate exists or not a bit early.
rebase: check whether the rebasestate exists or not a bit early
Mon, Jul 2, 8:43 AM
pulkit committed rHG4df9cef86740: rebase: use staeobj to check whether interrupted rebase exists.
rebase: use staeobj to check whether interrupted rebase exists
Mon, Jul 2, 8:43 AM
pulkit committed rHG7c853edcf4ed: rebase: add a stateobj variable to rebaseruntime class.
rebase: add a stateobj variable to rebaseruntime class
Mon, Jul 2, 8:43 AM
pulkit closed D3873: rebase: refactor logic to read rebasestate in a separate function.
Mon, Jul 2, 8:43 AM
pulkit committed rHG19076e2d62e7: rebase: refactor logic to read rebasestate in a separate function.
rebase: refactor logic to read rebasestate in a separate function
Mon, Jul 2, 8:43 AM

Sun, Jul 1

pulkit added inline comments to D3669: cmdutil: have statefile names in STATES instead of functions.
Sun, Jul 1, 5:27 PM
pulkit added inline comments to D3809: synthrepo: use progress helper.
Sun, Jul 1, 5:24 PM
pulkit added a comment to D3639: remotenames: add paths argument to remotenames revset.
In D3639#60247, @yuja wrote:
> > +@revsetpredicate('remotenames([path, ...])')
>
> My proposal was `remotenames([pattern])`, just like bookmark(), tag(),
>  branch(), etc.
>  If we want a convenient way to specify path prefix, we can
>  add it to the stringmatcher (e.g. 'remotenames("path:server2")'.)


No, I don't want to specify a path prefix. remotenames() will take a path only. Do you want to say that we should allow passing full remotenames instead of just path?

Yes, because a remote branch is merely a branch pulled from peer for example,
and we'll probably want to select it by -r remotebranch("foo/bar") just like
-r branch("foo/bar") for normal branches. Do I get it wrong? Why is it so
important to filter by a peer path, not by a branch name itself?

Sun, Jul 1, 5:13 PM
pulkit added a comment to D3866: histedit: use cbor to write histedit-state file.
In D3866#60421, @yuja wrote:

Queued the first 4 patches, thanks.

$ HGEDITOR="sh $TESTTMP/editplan.sh" hg histedit --edit-plan
$ cat .hg/histedit-state
  • v1
  • 055a42cdd88768532f9cf79daa407fc8d138de9b
  • 3c6a8ed2ebe862cc949d2caa30775dd6f16fb799
  • False
  • 3
  • drop
  • e860deea161a2f77de56603b340ebbb4536308ae
  • drop
  • 652413bf663ef2a641cab26574e46d5f5a64a55a
  • drop
  • 3c6a8ed2ebe862cc949d2caa30775dd6f16fb799
  • 0
  • strip-backup/177f92b77385-0ebe6a8f-histedit.hg + 2

I know anything other than "v1" will make old hg crash, but I think the
version syntax should stay "v%d" (i.e. "v2" in this case.)

Sun, Jul 1, 5:11 PM
pulkit created D3876: rebase: check whether the rebasestate exists or not a bit early.
Sun, Jul 1, 4:33 PM
pulkit added a dependent revision for D3875: rebase: use staeobj to check whether interrupted rebase exists: D3876: rebase: check whether the rebasestate exists or not a bit early.
Sun, Jul 1, 4:33 PM
pulkit created D3875: rebase: use staeobj to check whether interrupted rebase exists.
Sun, Jul 1, 4:33 PM
pulkit added a dependent revision for D3874: rebase: add a stateobj variable to rebaseruntime class: D3875: rebase: use staeobj to check whether interrupted rebase exists.
Sun, Jul 1, 4:33 PM
pulkit created D3874: rebase: add a stateobj variable to rebaseruntime class.
Sun, Jul 1, 4:33 PM
pulkit added a dependent revision for D3873: rebase: refactor logic to read rebasestate in a separate function: D3874: rebase: add a stateobj variable to rebaseruntime class.
Sun, Jul 1, 4:32 PM
pulkit created D3873: rebase: refactor logic to read rebasestate in a separate function.
Sun, Jul 1, 4:32 PM
pulkit accepted D3868: manifest: make cachesize a private attribute.
Sun, Jul 1, 1:52 PM
pulkit accepted D3867: repository: document that file() return value conforms to interface.
Sun, Jul 1, 1:50 PM
pulkit closed D3865: histedit: use self.stateobj to check whether interrupted histedit exists.
Sun, Jul 1, 8:37 AM
pulkit closed D3864: histedit: add a stateobj variable to histeditstate class.
Sun, Jul 1, 8:37 AM
pulkit committed rHG03e7ec8180f0: histedit: use self.stateobj to check whether interrupted histedit exists.
histedit: use self.stateobj to check whether interrupted histedit exists
Sun, Jul 1, 8:37 AM
pulkit committed rHG18f348e035fb: histedit: add a stateobj variable to histeditstate class.
histedit: add a stateobj variable to histeditstate class
Sun, Jul 1, 8:37 AM
pulkit closed D3863: histedit: factor out logic of processing state data in separate fn.
Sun, Jul 1, 8:37 AM
pulkit closed D3862: histedit: use hg.updaterepo() to avoid ui.{push|pop}buffer() hack.
Sun, Jul 1, 8:37 AM
pulkit committed rHG3beb0ea083df: histedit: use hg.updaterepo() to avoid ui.{push|pop}buffer() hack.
histedit: use hg.updaterepo() to avoid ui.{push|pop}buffer() hack
Sun, Jul 1, 8:37 AM
pulkit committed rHGc6a2ce82e60b: histedit: factor out logic of processing state data in separate fn.
histedit: factor out logic of processing state data in separate fn
Sun, Jul 1, 8:37 AM

Sat, Jun 30

pulkit abandoned D2946: histedit: factor out the logic to read oldstate file in separate fn.
Sat, Jun 30, 2:21 PM
pulkit created D3866: histedit: use cbor to write histedit-state file.
Sat, Jun 30, 11:10 AM
pulkit added a dependent revision for D3865: histedit: use self.stateobj to check whether interrupted histedit exists: D3866: histedit: use cbor to write histedit-state file.
Sat, Jun 30, 11:10 AM
pulkit created D3865: histedit: use self.stateobj to check whether interrupted histedit exists.
Sat, Jun 30, 11:10 AM
pulkit added a dependent revision for D3864: histedit: add a stateobj variable to histeditstate class: D3865: histedit: use self.stateobj to check whether interrupted histedit exists.
Sat, Jun 30, 11:10 AM
pulkit created D3864: histedit: add a stateobj variable to histeditstate class.
Sat, Jun 30, 11:09 AM
pulkit added a dependent revision for D3863: histedit: factor out logic of processing state data in separate fn: D3864: histedit: add a stateobj variable to histeditstate class.
Sat, Jun 30, 11:09 AM
pulkit created D3863: histedit: factor out logic of processing state data in separate fn.
Sat, Jun 30, 11:09 AM
pulkit added a dependent revision for D3862: histedit: use hg.updaterepo() to avoid ui.{push|pop}buffer() hack: D3863: histedit: factor out logic of processing state data in separate fn.
Sat, Jun 30, 11:09 AM
pulkit created D3862: histedit: use hg.updaterepo() to avoid ui.{push|pop}buffer() hack.
Sat, Jun 30, 11:09 AM

Fri, Jun 29

pulkit added a comment to D3693: shelve: directly handle the initial parent alignment.

I am very much happy with the cleanup but I don't feel confident enough to push this. Queued D3692.

Fri, Jun 29, 3:18 PM
pulkit accepted D3694: shelve: use more accurate description in conflict marker.
Fri, Jun 29, 3:10 PM
pulkit accepted D3693: shelve: directly handle the initial parent alignment.
Fri, Jun 29, 3:09 PM
pulkit accepted D3692: merge: add a 'keepconflictparent' argument to graft.

Looks like this will help fixing issue5927 too.

Fri, Jun 29, 3:07 PM
pulkit closed D3853: py3: convert opts keys to bytes using pycompat.byteskwargs().
Fri, Jun 29, 8:31 AM
pulkit committed rHG2394cd58b81f: py3: convert opts keys to bytes using pycompat.byteskwargs().
py3: convert opts keys to bytes using pycompat.byteskwargs()
Fri, Jun 29, 8:31 AM

Thu, Jun 28

pulkit added a comment to D3639: remotenames: add paths argument to remotenames revset.
In D3639#60108, @yuja wrote:

+@revsetpredicate('remotenames([path, ...])')

My proposal was remotenames([pattern]), just like bookmark(), tag(),
branch(), etc.
If we want a convenient way to specify path prefix, we can
add it to the stringmatcher (e.g. 'remotenames("path:server2")'.)

Thu, Jun 28, 4:37 PM
pulkit accepted D3825: grep: deprecates `--all` flag.
Thu, Jun 28, 2:01 PM
pulkit created D3853: py3: convert opts keys to bytes using pycompat.byteskwargs().
Thu, Jun 28, 1:38 PM
pulkit added inline comments to D3851: tests: pass "rev" argument to commands.update() as string.
Thu, Jun 28, 5:56 AM
pulkit accepted D3851: tests: pass "rev" argument to commands.update() as string.
Thu, Jun 28, 5:42 AM

Tue, Jun 26

pulkit created D3844: remotenames: add support to accept literal/re pattern in revsets.
Tue, Jun 26, 5:46 PM
pulkit added a dependent revision for D3639: remotenames: add paths argument to remotenames revset: D3844: remotenames: add support to accept literal/re pattern in revsets.
Tue, Jun 26, 5:46 PM
pulkit updated the diff for D3639: remotenames: add paths argument to remotenames revset.
Tue, Jun 26, 5:46 PM
pulkit updated the diff for D2874: remotenames: synchronise remotenames after push also.
Tue, Jun 26, 5:46 PM
pulkit accepted D3842: cleanup: migrate from re.escape to stringutil.reescape.
Tue, Jun 26, 1:28 PM
pulkit accepted D3841: stringutil: add a new function to do minimal regex escaping.
Tue, Jun 26, 1:26 PM
pulkit accepted D2409: graft: add no-commit mode (issue5631).

Macro landed:

Tue, Jun 26, 1:25 PM
pulkit closed D3836: patchbomb: use email.mime.base instead of email.MIMEBase.
Tue, Jun 26, 8:10 AM
pulkit committed rHGd17d1ee1d602: patchbomb: use email.mime.base instead of email.MIMEBase.
patchbomb: use email.mime.base instead of email.MIMEBase
Tue, Jun 26, 8:10 AM
pulkit committed rHG4c358bdaada8: py3: add b'' prefixes in tests/test-bundle2-pushback.t.
py3: add b'' prefixes in tests/test-bundle2-pushback.t
Tue, Jun 26, 8:10 AM
pulkit committed rHG91228d9ae7c8: patchbomb: use email.mime.multipart instead of email.MIMEMultipart.
patchbomb: use email.mime.multipart instead of email.MIMEMultipart
Tue, Jun 26, 8:10 AM
pulkit closed D3835: patchbomb: use email.mime.multipart instead of email.MIMEMultipart.
Tue, Jun 26, 8:10 AM
pulkit closed D3834: py3: add b'' prefixes in tests/test-bundle2-pushback.t.
Tue, Jun 26, 8:10 AM
pulkit committed rHG9ef9884e5d50: py3: make tests/test-diff-antipatience.t work with python 3.
py3: make tests/test-diff-antipatience.t work with python 3
Tue, Jun 26, 8:10 AM
pulkit closed D3833: py3: make tests/test-diff-antipatience.t work with python 3.
Tue, Jun 26, 8:10 AM
pulkit added a comment to D2409: graft: add no-commit mode (issue5631).

The patch mostly looks good to me. Added some minor nits. It will be great if you can add test where we graft multiple revs with --no-commit flag.

Tue, Jun 26, 4:04 AM
pulkit accepted D3838: terse: pass "clean" and "unknown" booleans by name for clarity.
Tue, Jun 26, 3:51 AM
pulkit accepted D3837: terse: add tests of running from subdirectory.
Tue, Jun 26, 3:50 AM

Mon, Jun 25

pulkit created D3836: patchbomb: use email.mime.base instead of email.MIMEBase.
Mon, Jun 25, 4:45 PM