| Submitter | Durham Goode |
|---|---|
| Date | Dec. 5, 2014, 10:54 p.m. |
| Message ID | <96d916622019f448bd01.1417820070@dev2000.prn2.facebook.com> |
| Download | mbox | patch |
| Permalink | /patch/7015/ |
| State | Superseded |
| Headers | show |
Comments
On 12/05/2014 02:54 PM, Durham Goode wrote: > # HG changeset patch > # User Durham Goode <durham@fb.com> > # Date 1417818452 28800 > # Fri Dec 05 14:27:32 2014 -0800 > # Node ID 96d916622019f448bd012884ee4fd78fe8dc63ba > # Parent 406dfc63a1ad71213dccc1a45de99a3c5d5ac460 > log: fix log -f slow path to actually follow history > > log -f was using ':.' instead of '::.' so it wasn't actually following history. > This fixes it and adds a test. It's been broken for years, so I guess it wasn't > that big of a deal... good catch. Does this open us to linkrev related bug when using it on a file?
On 12/5/14, 2:56 PM, "Pierre-Yves David" <pierre-yves.david@ens-lyon.org> wrote: > > >On 12/05/2014 02:54 PM, Durham Goode wrote: >> # HG changeset patch >> # User Durham Goode <durham@fb.com> >> # Date 1417818452 28800 >> # Fri Dec 05 14:27:32 2014 -0800 >> # Node ID 96d916622019f448bd012884ee4fd78fe8dc63ba >> # Parent 406dfc63a1ad71213dccc1a45de99a3c5d5ac460 >> log: fix log -f slow path to actually follow history >> >> log -f was using ':.' instead of '::.' so it wasn't actually following >>history. >> This fixes it and adds a test. It's been broken for years, so I guess >>it wasn't >> that big of a deal... > >good catch. > >Does this open us to linkrev related bug when using it on a file? Good question. I updated the test to check for that, and it doesn¹t seem to affect file log traversals when the linkrev commit is not an ancestor. I¹ll resend with the new test.
On 12/5/14 3:19 PM, Durham Goode wrote: > > On 12/5/14, 2:56 PM, "Pierre-Yves David" <pierre-yves.david@ens-lyon.org> > wrote: > >> >> On 12/05/2014 02:54 PM, Durham Goode wrote: >>> # HG changeset patch >>> # User Durham Goode <durham@fb.com> >>> # Date 1417818452 28800 >>> # Fri Dec 05 14:27:32 2014 -0800 >>> # Node ID 96d916622019f448bd012884ee4fd78fe8dc63ba >>> # Parent 406dfc63a1ad71213dccc1a45de99a3c5d5ac460 >>> log: fix log -f slow path to actually follow history >>> >>> log -f was using ':.' instead of '::.' so it wasn't actually following >>> history. >>> This fixes it and adds a test. It's been broken for years, so I guess >>> it wasn't >>> that big of a deal... >> good catch. >> >> Does this open us to linkrev related bug when using it on a file? > Good question. I updated the test to check for that, and it doesn¹t seem > to affect file log traversals when the linkrev commit is not an ancestor. > I¹ll resend with the new test. > > Actually, I lied. It does affect it :( My test was just wrong. Sigh. K I'll figure out a fix and resend.
On 12/05/2014 02:54 PM, Durham Goode wrote: > # HG changeset patch > # User Durham Goode <durham@fb.com> > # Date 1417818452 28800 > # Fri Dec 05 14:27:32 2014 -0800 > # Node ID 96d916622019f448bd012884ee4fd78fe8dc63ba > # Parent 406dfc63a1ad71213dccc1a45de99a3c5d5ac460 > log: fix log -f slow path to actually follow history > > log -f was using ':.' instead of '::.' so it wasn't actually following history. > This fixes it and adds a test. It's been broken for years, so I guess it wasn't > that big of a deal... Still a bugfix, so should be on stable? > > diff --git a/mercurial/cmdutil.py b/mercurial/cmdutil.py > --- a/mercurial/cmdutil.py > +++ b/mercurial/cmdutil.py > @@ -1836,7 +1836,7 @@ def getgraphlogrevs(repo, pats, opts): > possiblyunsorted = True > else: > if follow and len(repo) > 0: > - revs = repo.revs('reverse(:.)') > + revs = repo.revs('reverse(::.)') > else: > revs = revset.spanset(repo) > revs.reverse() > @@ -1880,7 +1880,7 @@ def getlogrevs(repo, pats, opts): > if opts.get('rev'): > revs = scmutil.revrange(repo, opts['rev']) > elif follow: > - revs = repo.revs('reverse(:.)') > + revs = repo.revs('reverse(::.)') > else: > revs = revset.spanset(repo) > revs.reverse() > diff --git a/tests/test-log.t b/tests/test-log.t > --- a/tests/test-log.t > +++ b/tests/test-log.t > @@ -1543,3 +1543,19 @@ issue3772: hg log -r :null showing revis > > > $ cd .. > + > +hg log -f dir across branches > + > + $ hg init acrossbranches > + $ cd acrossbranches > + $ mkdir d > + $ echo a > d/a && hg ci -Aqm a > + $ echo b > d/b && hg ci -Aqm b > + $ hg up -q 0 > + $ echo c > d/c && hg ci -Aqm c > + $ hg log -f d -T '{desc}' -G > + @ c > + | > + o a > + > + $ cd .. > _______________________________________________ > Mercurial-devel mailing list > Mercurial-devel@selenic.com > http://selenic.com/mailman/listinfo/mercurial-devel
On 12/5/14 3:22 PM, Durham Goode wrote: > > On 12/5/14 3:19 PM, Durham Goode wrote: >> >> On 12/5/14, 2:56 PM, "Pierre-Yves David" >> <pierre-yves.david@ens-lyon.org> >> wrote: >> >>> >>> On 12/05/2014 02:54 PM, Durham Goode wrote: >>>> # HG changeset patch >>>> # User Durham Goode <durham@fb.com> >>>> # Date 1417818452 28800 >>>> # Fri Dec 05 14:27:32 2014 -0800 >>>> # Node ID 96d916622019f448bd012884ee4fd78fe8dc63ba >>>> # Parent 406dfc63a1ad71213dccc1a45de99a3c5d5ac460 >>>> log: fix log -f slow path to actually follow history >>>> >>>> log -f was using ':.' instead of '::.' so it wasn't actually following >>>> history. >>>> This fixes it and adds a test. It's been broken for years, so I guess >>>> it wasn't >>>> that big of a deal... >>> good catch. >>> >>> Does this open us to linkrev related bug when using it on a file? >> Good question. I updated the test to check for that, and it doesn¹t >> seem >> to affect file log traversals when the linkrev commit is not an >> ancestor. >> I¹ll resend with the new test. >> >> > Actually, I lied. It does affect it :( My test was just wrong. > Sigh. K I'll figure out a fix and resend. Good news everyone! hg log -f <file> was broken for linkrevs even without my change! So I'll send my patch out with a better test and will log a bug about log -f <file> being broken, since it will require a more indepth fix around linkrevs.
On 12/5/14 4:05 PM, Durham Goode wrote: > > On 12/5/14 3:22 PM, Durham Goode wrote: >> >> On 12/5/14 3:19 PM, Durham Goode wrote: >>> >>> On 12/5/14, 2:56 PM, "Pierre-Yves David" >>> <pierre-yves.david@ens-lyon.org> >>> wrote: >>> >>>> >>>> On 12/05/2014 02:54 PM, Durham Goode wrote: >>>>> # HG changeset patch >>>>> # User Durham Goode <durham@fb.com> >>>>> # Date 1417818452 28800 >>>>> # Fri Dec 05 14:27:32 2014 -0800 >>>>> # Node ID 96d916622019f448bd012884ee4fd78fe8dc63ba >>>>> # Parent 406dfc63a1ad71213dccc1a45de99a3c5d5ac460 >>>>> log: fix log -f slow path to actually follow history >>>>> >>>>> log -f was using ':.' instead of '::.' so it wasn't actually >>>>> following >>>>> history. >>>>> This fixes it and adds a test. It's been broken for years, so I guess >>>>> it wasn't >>>>> that big of a deal... >>>> good catch. >>>> >>>> Does this open us to linkrev related bug when using it on a file? >>> Good question. I updated the test to check for that, and it doesn¹t >>> seem >>> to affect file log traversals when the linkrev commit is not an >>> ancestor. >>> I¹ll resend with the new test. >>> >>> >> Actually, I lied. It does affect it :( My test was just wrong. >> Sigh. K I'll figure out a fix and resend. > Good news everyone! hg log -f <file> was broken for linkrevs even > without my change! So I'll send my patch out with a better test and > will log a bug about log -f <file> being broken, since it will require > a more indepth fix around linkrevs. And once again I'm proven wrong because my test was wrong. Ok, I'm going to stop spamming the list until I have it done.
On Fri, 2014-12-05 at 14:54 -0800, Durham Goode wrote: > # HG changeset patch > # User Durham Goode <durham@fb.com> > # Date 1417818452 28800 > # Fri Dec 05 14:27:32 2014 -0800 > # Node ID 96d916622019f448bd012884ee4fd78fe8dc63ba > # Parent 406dfc63a1ad71213dccc1a45de99a3c5d5ac460 > log: fix log -f slow path to actually follow history > > log -f was using ':.' instead of '::.' so it wasn't actually following history. > This fixes it and adds a test. It's been broken for years, so I guess it wasn't > that big of a deal... Actually, log wasn't able to use revsets until we made them lazy, so this bug is still newish.
Patch
diff --git a/mercurial/cmdutil.py b/mercurial/cmdutil.py --- a/mercurial/cmdutil.py +++ b/mercurial/cmdutil.py @@ -1836,7 +1836,7 @@ def getgraphlogrevs(repo, pats, opts): possiblyunsorted = True else: if follow and len(repo) > 0: - revs = repo.revs('reverse(:.)') + revs = repo.revs('reverse(::.)') else: revs = revset.spanset(repo) revs.reverse() @@ -1880,7 +1880,7 @@ def getlogrevs(repo, pats, opts): if opts.get('rev'): revs = scmutil.revrange(repo, opts['rev']) elif follow: - revs = repo.revs('reverse(:.)') + revs = repo.revs('reverse(::.)') else: revs = revset.spanset(repo) revs.reverse() diff --git a/tests/test-log.t b/tests/test-log.t --- a/tests/test-log.t +++ b/tests/test-log.t @@ -1543,3 +1543,19 @@ issue3772: hg log -r :null showing revis $ cd .. + +hg log -f dir across branches + + $ hg init acrossbranches + $ cd acrossbranches + $ mkdir d + $ echo a > d/a && hg ci -Aqm a + $ echo b > d/b && hg ci -Aqm b + $ hg up -q 0 + $ echo c > d/c && hg ci -Aqm c + $ hg log -f d -T '{desc}' -G + @ c + | + o a + + $ cd ..