| Submitter | Katsunori FUJIWARA |
|---|---|
| Date | June 12, 2017, 11:12 a.m. |
| Message ID | <uziddxxw8.wl%foozy@lares.dti.ne.jp> |
| Download | mbox | patch |
| Permalink | /patch/21338/ |
| State | Not Applicable |
| Headers | show |
Comments
At Mon, 12 Jun 2017 20:12:39 +0900, FUJIWARA Katsunori wrote: > > At Mon, 12 Jun 2017 19:31:55 +0900, > FUJIWARA Katsunori wrote: > > > > At Mon, 12 Jun 2017 19:04:28 +0900, > > FUJIWARA Katsunori wrote: > > > > > > At Sun, 11 Jun 2017 22:24:20 -0700, > > > Siddharth Agarwal wrote: > > > > > > > > On 6/11/17 10:21 PM, Martin von Zweigbergk via Mercurial-devel wrote: > > > > > Doesn't look like it's guaranteed to be unset. Running "run-tests.py > > > > > -j50 --runs-per-test=50 test-dirstate-race.t" seems to confirm that (2 > > > > > of 5 failed for me). How do you rely on it in an upcoming patch? > > > > > > > > > > > > Interesting, I'm not seeing that on my Linux box. Could you drop this > > > > for now? I think I can work around this problem. > > > > > > Unfortunately, this can occur, if mtime of "a" is N, and _getfsnow() > > > returns N+1 or more (= "a" is written out at almost N+1 but N in > > > "int"). > > > > > > IMHO, "hg update" with --config debug.dirstate.delaywrite=2 should > > > avoid this problem. > > > > Oops, sorry, this is incorrect, because this is the way to ENSURE > > mtime in dirstate :-< > > > > IMHO, additional "touch -t YYYYmmddHHMM 'a'" and "hg status" with > > tests/fakedirstatewritetime.py in previous $TESTTMP/dirstaterace.sh > > can ensure "unset"-ness of mtime in dirstate in this > > test-dirstate-race.t case. > > > > FYI, changes below can always reproduce Martin's situation, in which > files are changed at N, but dirstate is written out N+1 or later in > sec. > > And uncommenting two lines after "enabling below ..." line should fix > this issue. > > After confirming that fakedirstatewritetime.py works as expected, > "--config debug.dirstate.delaywrite=2" can be omitted for efficiency :-) > > ======================================== > diff --git a/tests/test-dirstate-race.t b/tests/test-dirstate-race.t > --- a/tests/test-dirstate-race.t > +++ b/tests/test-dirstate-race.t > @@ -134,8 +134,21 @@ condition (see issue5584 for detail). > > # To reproduce similar situation easily and certainly, #2 and #3 > > # are swapped. "hg cat" below ensures #2 on "hg status" side. > > > - > hg update -q -C 0 > + > # "--config debug.dirstate.delaywrite=2" emulates that files are > + > # changed at N, but dirstate is written out N+1 or later in sec > + > hg update -q -C 0 --config debug.dirstate.delaywrite=2 > > hg cat -r 1 b > b > + > > + > #### enabling below can ensure that mtime of 'a' in dirstate is unset > + > # touch -t 200001010000 a > + > # hg status --config extensions.fakedirstatewritetime=$TESTDIR/fakedirstatewritetime.py > + > EOF > + > + $ cat >> .hg/hgrc <<EOF > + > [fakedirstatewritetime] > + > # emulate invoking dirstate.write() via repo.status() > + > # at 2000-01-01 00:00 > + > fakenow = 200001010000 > > EOF > > "hg status" below should excludes "e", of which exec flag is set, for > ======================================== Sorry for consecutive posting, but one more :-) "hg status" with fakedirstatewritetime above causes "? b" output before "skip updating dirstate: identity mismatch" message of outer "hg status". To avoid this meaningless output: - add "-m" option or so to "hg status" with fakedirstatewritetime, or - move "hg cat -r 1 b > b" after "hg status" with fakedirstatewritetime (the former seems reasonable, in order to put commands together by each purpose) > > > > > > > > _______________________________________________ > > > > Mercurial-devel mailing list > > > > Mercurial-devel@mercurial-scm.org > > > > https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel > > > > > > -- > > > ---------------------------------------------------------------------- > > > [FUJIWARA Katsunori] foozy@lares.dti.ne.jp > > > _______________________________________________ > > > Mercurial-devel mailing list > > > Mercurial-devel@mercurial-scm.org > > > https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel > > > > -- > > ---------------------------------------------------------------------- > > [FUJIWARA Katsunori] foozy@lares.dti.ne.jp > > > > -- > ---------------------------------------------------------------------- > [FUJIWARA Katsunori] foozy@lares.dti.ne.jp >
If Sid did want an unset timestamp in the dirstate (it sounded like he didn't really need it), he could also use debugrebuilddirstate, right? I think that always leaves the timestamp unset. On Jun 12, 2017 4:52 AM, "FUJIWARA Katsunori" <foozy@lares.dti.ne.jp> wrote: At Mon, 12 Jun 2017 20:12:39 +0900, FUJIWARA Katsunori wrote: > > At Mon, 12 Jun 2017 19:31:55 +0900, > FUJIWARA Katsunori wrote: > > > > At Mon, 12 Jun 2017 19:04:28 +0900, > > FUJIWARA Katsunori wrote: > > > > > > At Sun, 11 Jun 2017 22:24:20 -0700, > > > Siddharth Agarwal wrote: > > > > > > > > On 6/11/17 10:21 PM, Martin von Zweigbergk via Mercurial-devel wrote: > > > > > Doesn't look like it's guaranteed to be unset. Running "run-tests.py > > > > > -j50 --runs-per-test=50 test-dirstate-race.t" seems to confirm that (2 > > > > > of 5 failed for me). How do you rely on it in an upcoming patch? > > > > > > > > > > > > Interesting, I'm not seeing that on my Linux box. Could you drop this > > > > for now? I think I can work around this problem. > > > > > > Unfortunately, this can occur, if mtime of "a" is N, and _getfsnow() > > > returns N+1 or more (= "a" is written out at almost N+1 but N in > > > "int"). > > > > > > IMHO, "hg update" with --config debug.dirstate.delaywrite=2 should > > > avoid this problem. > > > > Oops, sorry, this is incorrect, because this is the way to ENSURE > > mtime in dirstate :-< > > > > IMHO, additional "touch -t YYYYmmddHHMM 'a'" and "hg status" with > > tests/fakedirstatewritetime.py in previous $TESTTMP/dirstaterace.sh > > can ensure "unset"-ness of mtime in dirstate in this > > test-dirstate-race.t case. > > > > FYI, changes below can always reproduce Martin's situation, in which > files are changed at N, but dirstate is written out N+1 or later in > sec. > > And uncommenting two lines after "enabling below ..." line should fix > this issue. > > After confirming that fakedirstatewritetime.py works as expected, > "--config debug.dirstate.delaywrite=2" can be omitted for efficiency :-) > > ======================================== > diff --git a/tests/test-dirstate-race.t b/tests/test-dirstate-race.t > --- a/tests/test-dirstate-race.t > +++ b/tests/test-dirstate-race.t > @@ -134,8 +134,21 @@ condition (see issue5584 for detail). > > # To reproduce similar situation easily and certainly, #2 and #3 > > # are swapped. "hg cat" below ensures #2 on "hg status" side. > > > - > hg update -q -C 0 > + > # "--config debug.dirstate.delaywrite=2" emulates that files are > + > # changed at N, but dirstate is written out N+1 or later in sec > + > hg update -q -C 0 --config debug.dirstate.delaywrite=2 > > hg cat -r 1 b > b > + > > + > #### enabling below can ensure that mtime of 'a' in dirstate is unset > + > # touch -t 200001010000 a > + > # hg status --config extensions.fakedirstatewritetime=$ TESTDIR/fakedirstatewritetime.py > + > EOF > + > + $ cat >> .hg/hgrc <<EOF > + > [fakedirstatewritetime] > + > # emulate invoking dirstate.write() via repo.status() > + > # at 2000-01-01 00:00 > + > fakenow = 200001010000 > > EOF > > "hg status" below should excludes "e", of which exec flag is set, for > ======================================== Sorry for consecutive posting, but one more :-) "hg status" with fakedirstatewritetime above causes "? b" output before "skip updating dirstate: identity mismatch" message of outer "hg status". To avoid this meaningless output: - add "-m" option or so to "hg status" with fakedirstatewritetime, or - move "hg cat -r 1 b > b" after "hg status" with fakedirstatewritetime (the former seems reasonable, in order to put commands together by each purpose) > > > > > > > > _______________________________________________ > > > > Mercurial-devel mailing list > > > > Mercurial-devel@mercurial-scm.org > > > > https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel > > > > > > -- > > > ---------------------------------------------------------------------- > > > [FUJIWARA Katsunori] foozy@lares.dti.ne.jp > > > _______________________________________________ > > > Mercurial-devel mailing list > > > Mercurial-devel@mercurial-scm.org > > > https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel > > > > -- > > ---------------------------------------------------------------------- > > [FUJIWARA Katsunori] foozy@lares.dti.ne.jp > > > > -- > ---------------------------------------------------------------------- > [FUJIWARA Katsunori] foozy@lares.dti.ne.jp > -- ---------------------------------------------------------------------- [FUJIWARA Katsunori] foozy@lares.dti.ne.jp
Patch
======================================== diff --git a/tests/test-dirstate-race.t b/tests/test-dirstate-race.t --- a/tests/test-dirstate-race.t +++ b/tests/test-dirstate-race.t @@ -134,8 +134,21 @@ condition (see issue5584 for detail). > # To reproduce similar situation easily and certainly, #2 and #3 > # are swapped. "hg cat" below ensures #2 on "hg status" side. > - > hg update -q -C 0 + > # "--config debug.dirstate.delaywrite=2" emulates that files are + > # changed at N, but dirstate is written out N+1 or later in sec + > hg update -q -C 0 --config debug.dirstate.delaywrite=2 > hg cat -r 1 b > b + > + > #### enabling below can ensure that mtime of 'a' in dirstate is unset + > # touch -t 200001010000 a + > # hg status --config extensions.fakedirstatewritetime=$TESTDIR/fakedirstatewritetime.py + > EOF + + $ cat >> .hg/hgrc <<EOF + > [fakedirstatewritetime] + > # emulate invoking dirstate.write() via repo.status() + > # at 2000-01-01 00:00 + > fakenow = 200001010000 > EOF "hg status" below should excludes "e", of which exec flag is set, for