Page MenuHomePhabricator

aalekseyev (Arseniy Alekseyev)
User

Projects

User does not belong to any projects.

User Details

User Since
Fri, Oct 1, 12:37 PM (2 w, 3 d)

Recent Activity

Today

aalekseyev added a comment to D11650: rhg: handle null changelog and manifest revisions.

... push a topic to https://foss.heptapod.net/mercurial/mercurial-devel (if you create an account I can give you the rights to do so).

Mon, Oct 18, 1:38 PM
aalekseyev closed D11679: rhg: internally, return a structured representation from hg cat.
Mon, Oct 18, 7:46 AM
aalekseyev committed rHG027ebad952ac: rhg: internally, return a structured representation from hg cat.
rhg: internally, return a structured representation from hg cat
Mon, Oct 18, 7:46 AM
aalekseyev retitled D11679: rhg: internally, return a structured representation from hg cat from rhg: simplify the type of FilelogEntry to rhg: internally, return a structured representation from hg cat.
Mon, Oct 18, 6:36 AM
aalekseyev added a comment to D11679: rhg: internally, return a structured representation from hg cat.

This does not apply on top of default sorry, please send a rebase.

Mon, Oct 18, 6:31 AM
aalekseyev added inline comments to D11679: rhg: internally, return a structured representation from hg cat.
Mon, Oct 18, 6:30 AM
aalekseyev updated the diff for D11679: rhg: internally, return a structured representation from hg cat.
Mon, Oct 18, 6:28 AM
aalekseyev added a comment to D11679: rhg: internally, return a structured representation from hg cat.

Oh also, (sorry I forgot to say anything) the commit message has a typo in "representation".

Mon, Oct 18, 6:28 AM
aalekseyev updated the summary of D11679: rhg: internally, return a structured representation from hg cat.
Mon, Oct 18, 6:25 AM
aalekseyev updated the summary of D11679: rhg: internally, return a structured representation from hg cat.
Mon, Oct 18, 6:20 AM
aalekseyev retitled D11679: rhg: internally, return a structured representation from hg cat from rhg: internally, return a structured representatioon from hg cat to rhg: simplify the type of FilelogEntry.
Mon, Oct 18, 6:19 AM
aalekseyev added inline comments to D11679: rhg: internally, return a structured representation from hg cat.
Mon, Oct 18, 6:10 AM
aalekseyev added a comment to D11679: rhg: internally, return a structured representation from hg cat.

However I wonder if the fragmented allocations have a measurable impact on performance for a case with a large output (in both number of files and number of lines). Do you have a way of testing this easily?

Mon, Oct 18, 6:07 AM
aalekseyev closed D11664: rhg: fix `hg cat` interaction with null revision.
Mon, Oct 18, 5:49 AM
aalekseyev committed rHGf8dc78716ad2: rhg: fix `hg cat` interaction with null revision.
rhg: fix `hg cat` interaction with null revision
Mon, Oct 18, 5:49 AM
aalekseyev added inline comments to D11650: rhg: handle null changelog and manifest revisions.
Mon, Oct 18, 5:27 AM
aalekseyev added a comment to D11650: rhg: handle null changelog and manifest revisions.

This breaks the CI: https://foss.heptapod.net/mercurial/mercurial-devel/-/jobs/254045
It seems that the previous assumption made in the code was that null revs were to be shortcut and this assumption was not changed everywhere. @aalekseyev The easiest way of making sure you're not breaking anything is to push a topic to https://foss.heptapod.net/mercurial/mercurial-devel (if you create an account I can give you the rights to do so).
Right now other tests are failing, so you might see unrelated failures, but I'm looking into those as well. I try to keep the CI is mostly green, but I was on vacation.

Mon, Oct 18, 5:21 AM

Fri, Oct 15

aalekseyev added a comment to D11664: rhg: fix `hg cat` interaction with null revision.

Rebased! (I think)

Fri, Oct 15, 9:41 AM
aalekseyev updated the diff for D11664: rhg: fix `hg cat` interaction with null revision.
Fri, Oct 15, 9:41 AM
aalekseyev updated the diff for D11664: rhg: fix `hg cat` interaction with null revision.
Fri, Oct 15, 9:39 AM
aalekseyev retitled D11679: rhg: internally, return a structured representation from hg cat from rhg: simplify the type of FilelogEntry to rhg: internally, return a structured representatioon from hg cat.
Fri, Oct 15, 9:33 AM
aalekseyev retitled D11679: rhg: internally, return a structured representation from hg cat from rhg: add FilelogEntry.into_data to rhg: simplify the type of FilelogEntry.
Fri, Oct 15, 9:31 AM
aalekseyev retitled D11679: rhg: internally, return a structured representation from hg cat from rhg: simplify the type of FilelogEntry to rhg: add FilelogEntry.into_data.
Fri, Oct 15, 9:30 AM
aalekseyev retitled D11679: rhg: internally, return a structured representation from hg cat from rhg: internally, return a structured representatioon from hg cat to rhg: simplify the type of FilelogEntry.
Fri, Oct 15, 9:30 AM
aalekseyev retitled D11679: rhg: internally, return a structured representation from hg cat from rhg: simplify the type of FilelogEntry to rhg: internally, return a structured representatioon from hg cat.
Fri, Oct 15, 9:27 AM
aalekseyev created D11679: rhg: internally, return a structured representation from hg cat.
Fri, Oct 15, 9:21 AM
aalekseyev added a comment to D11670: rhg: fix formatting error reported by test-check-format-rust.t.

@martinvonz, indeed, thanks. Sorry.

Fri, Oct 15, 5:02 AM
aalekseyev added a comment to D11664: rhg: fix `hg cat` interaction with null revision.

I forgot to mention that this test was broken with a "panic" since 61ce70fd420e (sorry!).

Fri, Oct 15, 4:59 AM

Thu, Oct 14

aalekseyev closed D11659: rhg: do not try to open a nodemap for an inline index.
Thu, Oct 14, 6:36 PM
aalekseyev committed rHG5e77bdc29d56: rhg: do not try to open a nodemap for an inline index.
rhg: do not try to open a nodemap for an inline index
Thu, Oct 14, 6:35 PM
aalekseyev created D11664: rhg: fix `hg cat` interaction with null revision.
Thu, Oct 14, 2:05 PM
aalekseyev added a comment to D11616: rhg: stop manifest traversal when no more files are needed.

Thanks!

Thu, Oct 14, 12:15 PM
aalekseyev closed D11616: rhg: stop manifest traversal when no more files are needed.
Thu, Oct 14, 11:17 AM
aalekseyev closed D11615: rhg: faster hg cat when many files are requested.
Thu, Oct 14, 11:17 AM
aalekseyev committed rHG0cc69017d47f: rhg: stop manifest traversal when no more files are needed.
rhg: stop manifest traversal when no more files are needed
Thu, Oct 14, 11:17 AM
aalekseyev committed rHG6b5773f89183: rhg: faster hg cat when many files are requested.
rhg: faster hg cat when many files are requested
Thu, Oct 14, 11:16 AM
aalekseyev closed D11656: rhg: fix the test.
Thu, Oct 14, 9:09 AM
aalekseyev created D11659: rhg: do not try to open a nodemap for an inline index.
Thu, Oct 14, 9:03 AM
aalekseyev added a comment to D11615: rhg: faster hg cat when many files are requested.

Even though my measurement above talks about concatenation of 200 files, a real production use case we have is concatenating around 10000 files.
The time of that operation is improved from ~3.5 seconds to ~0.4 seconds by this change.

Thu, Oct 14, 5:29 AM

Wed, Oct 13

aalekseyev added a comment to D11616: rhg: stop manifest traversal when no more files are needed.

I'm getting ~630ms on **/moz.build when I run the original version (1b0f8aafedea), by the way. Could it be that your ~650ms measurement corresponds to that?

Wed, Oct 13, 1:41 PM
aalekseyev committed rHG9d0e5629cfbf: rhg: do not fail when the repo is empty.
rhg: do not fail when the repo is empty
Wed, Oct 13, 1:18 PM
aalekseyev added a comment to D11650: rhg: handle null changelog and manifest revisions.

I'll do that.

Wed, Oct 13, 1:15 PM
aalekseyev created D11656: rhg: fix the test.
Wed, Oct 13, 1:14 PM
aalekseyev added a comment to D11650: rhg: handle null changelog and manifest revisions.

This broke a Rust unit test: https://foss.heptapod.net/mercurial/mercurial-devel/-/jobs/253010#L32
You can run these tests with (cd rust && cargo test). Could you send a follow-up patch to fix this?

Wed, Oct 13, 12:58 PM
aalekseyev closed D11651: rhg: do not fail when the repo is empty.
Wed, Oct 13, 11:44 AM
aalekseyev committed rHGeb753dd284a9: rhg: do not fail when the repo is empty.
rhg: do not fail when the repo is empty
Wed, Oct 13, 11:44 AM

Tue, Oct 12

aalekseyev added a comment to D11650: rhg: handle null changelog and manifest revisions.

@martinvonz, thanks!
I pushed the follow-up changes in:
https://phab.mercurial-scm.org/D11651

Tue, Oct 12, 4:09 PM
aalekseyev created D11651: rhg: do not fail when the repo is empty.
Tue, Oct 12, 4:07 PM
aalekseyev closed D11650: rhg: handle null changelog and manifest revisions.
Tue, Oct 12, 4:03 PM
aalekseyev committed rHG61ce70fd420e: rhg: handle null changelog and manifest revisions.
rhg: handle null changelog and manifest revisions
Tue, Oct 12, 4:03 PM
aalekseyev added inline comments to D11650: rhg: handle null changelog and manifest revisions.
Tue, Oct 12, 3:20 PM
aalekseyev created D11650: rhg: handle null changelog and manifest revisions.
Tue, Oct 12, 2:44 PM
aalekseyev added a comment to D11616: rhg: stop manifest traversal when no more files are needed.

I wonder why we see so different results. Are those the times reported by hyperfine in your case?

Tue, Oct 12, 8:34 AM

Mon, Oct 11

aalekseyev added a comment to D11616: rhg: stop manifest traversal when no more files are needed.

...
Mine has ~260k files. I don't know how many mozilla repo has, but I'll re-run my benchmarks on that tomorrow to see if I'm getting the same results.

Mon, Oct 11, 1:18 PM
aalekseyev added a comment to D11616: rhg: stop manifest traversal when no more files are needed.

I tried just now in the mozilla-unified repo. The differences are not measurable there either (see below).

Mon, Oct 11, 1:08 PM
aalekseyev added a comment to D11616: rhg: stop manifest traversal when no more files are needed.

I think this can be done more simply. I've sent D11622 to show what I mean. Does that make sense? How does it perform in your test repo?

Mon, Oct 11, 9:35 AM
aalekseyev updated the diff for D11616: rhg: stop manifest traversal when no more files are needed.
Mon, Oct 11, 6:51 AM
aalekseyev updated the diff for D11615: rhg: faster hg cat when many files are requested.
Mon, Oct 11, 6:51 AM
aalekseyev added inline comments to D11615: rhg: faster hg cat when many files are requested.
Mon, Oct 11, 6:40 AM

Fri, Oct 8

aalekseyev closed D11617: rhg: in rhg cat cli, fix the long name of the --rev flag.
Fri, Oct 8, 9:34 PM
aalekseyev committed rHG08c8cd2527bc: rhg: in rhg cat cli, fix the long name of the --rev flag.
rhg: in rhg cat cli, fix the long name of the --rev flag
Fri, Oct 8, 9:34 PM

Tue, Oct 5

aalekseyev retitled D11617: rhg: in rhg cat cli, fix the long name of the --rev flag from rhg: in rhg cat cli, fix the long name of the --rev flag\n\nAlso tweak the help for the anonymous argument. to rhg: in rhg cat cli, fix the long name of the --rev flag.
Tue, Oct 5, 11:15 AM
aalekseyev created D11617: rhg: in rhg cat cli, fix the long name of the --rev flag.
Tue, Oct 5, 11:09 AM
aalekseyev updated the diff for D11616: rhg: stop manifest traversal when no more files are needed.
Tue, Oct 5, 10:50 AM
aalekseyev updated the summary of D11616: rhg: stop manifest traversal when no more files are needed.
Tue, Oct 5, 10:44 AM
aalekseyev updated the summary of D11615: rhg: faster hg cat when many files are requested.
Tue, Oct 5, 10:44 AM
aalekseyev created D11615: rhg: faster hg cat when many files are requested.
Tue, Oct 5, 10:16 AM
aalekseyev created D11616: rhg: stop manifest traversal when no more files are needed.
Tue, Oct 5, 10:16 AM