| Submitter | Isaac Jurado |
|---|---|
| Date | April 18, 2013, 7:56 p.m. |
| Message ID | <380de75c43f4a186a1b8.1366315008@findus> |
| Download | mbox | patch |
| Permalink | /patch/1436/ |
| State | Accepted, archived |
| Headers | show |
Comments
On Thu, 2013-04-18 at 21:56 +0200, diptongo@gmail.com wrote: > # HG changeset patch > # User Isaac Jurado <diptongo@gmail.com> > # Date 1366314160 -7200 > # Thu Apr 18 21:42:40 2013 +0200 > # Node ID 380de75c43f4a186a1b8255393cf52c3af43820a > # Parent 7d31f2e42a8afb54c8fae87e8e3e29a63578aea4 > dirstate: use repository root as base for relative paths read from ui.ignore > > Convert relative paths from the ui.ignore* configurations to absolute by > prepending the path to the repository root. This breaks the old, and probably > unused, behaviour where relative paths were interpreted from the working > directory. > > Absolute paths are kept untouched. Seems safe enough. Queued for default with a (BC) marker.
On Thu, 2013-04-18 at 17:47 -0500, Matt Mackall wrote: > On Thu, 2013-04-18 at 21:56 +0200, diptongo@gmail.com wrote: > > # HG changeset patch > > # User Isaac Jurado <diptongo@gmail.com> > > # Date 1366314160 -7200 > > # Thu Apr 18 21:42:40 2013 +0200 > > # Node ID 380de75c43f4a186a1b8255393cf52c3af43820a > > # Parent 7d31f2e42a8afb54c8fae87e8e3e29a63578aea4 > > dirstate: use repository root as base for relative paths read from ui.ignore > > > > Convert relative paths from the ui.ignore* configurations to absolute by > > prepending the path to the repository root. This breaks the old, and probably > > unused, behaviour where relative paths were interpreted from the working > > directory. > > > > Absolute paths are kept untouched. > > Seems safe enough. Queued for default with a (BC) marker. Or maybe not. Something just destroyed my rsync-to-box-in-cloud-and-test script, and this is the most likely culprit.
On Fri, Apr 19, 2013 at 1:16 AM, Matt Mackall <mpm@selenic.com> wrote: > On Thu, 2013-04-18 at 17:47 -0500, Matt Mackall wrote: >> On Thu, 2013-04-18 at 21:56 +0200, diptongo@gmail.com wrote: >> > # HG changeset patch >> > # User Isaac Jurado <diptongo@gmail.com> >> > # Date 1366314160 -7200 >> > # Thu Apr 18 21:42:40 2013 +0200 >> > # Node ID 380de75c43f4a186a1b8255393cf52c3af43820a >> > # Parent 7d31f2e42a8afb54c8fae87e8e3e29a63578aea4 >> > dirstate: use repository root as base for relative paths read from ui.ignore >> > >> > Convert relative paths from the ui.ignore* configurations to absolute by >> > prepending the path to the repository root. This breaks the old, and probably >> > unused, behaviour where relative paths were interpreted from the working >> > directory. >> > >> > Absolute paths are kept untouched. >> >> Seems safe enough. Queued for default with a (BC) marker. > > Or maybe not. Something just destroyed my > rsync-to-box-in-cloud-and-test script, and this is the most likely > culprit. Ok, I'll resend on May. Meanwhile, it could be useful to have more details abut such failure, if possible. Cheers.
On 04/18/2013 09:56 PM, diptongo@gmail.com wrote: > # HG changeset patch > # User Isaac Jurado <diptongo@gmail.com> > # Date 1366314160 -7200 > # Thu Apr 18 21:42:40 2013 +0200 > # Node ID 380de75c43f4a186a1b8255393cf52c3af43820a > # Parent 7d31f2e42a8afb54c8fae87e8e3e29a63578aea4 > dirstate: use repository root as base for relative paths read from ui.ignore > > Convert relative paths from the ui.ignore* configurations to absolute by > prepending the path to the repository root. This breaks the old, and probably > unused, behaviour where relative paths were interpreted from the working > directory. I agree that the current behaviour is useless, but think that would you describe would be surprising. I would expect a relative filename in a config file to be relative to the file it is specified in (or cwd if specified on the command line). See also http://markmail.org/message/fcgvz5zij62triwk for a similar discussion and some unanswered questions you might want to pick up. /Mads
On Fri, 2013-04-19 at 10:53 +0200, Isaac Jurado wrote: > On Fri, Apr 19, 2013 at 1:16 AM, Matt Mackall <mpm@selenic.com> wrote: > > On Thu, 2013-04-18 at 17:47 -0500, Matt Mackall wrote: > >> On Thu, 2013-04-18 at 21:56 +0200, diptongo@gmail.com wrote: > >> > # HG changeset patch > >> > # User Isaac Jurado <diptongo@gmail.com> > >> > # Date 1366314160 -7200 > >> > # Thu Apr 18 21:42:40 2013 +0200 > >> > # Node ID 380de75c43f4a186a1b8255393cf52c3af43820a > >> > # Parent 7d31f2e42a8afb54c8fae87e8e3e29a63578aea4 > >> > dirstate: use repository root as base for relative paths read from ui.ignore > >> > > >> > Convert relative paths from the ui.ignore* configurations to absolute by > >> > prepending the path to the repository root. This breaks the old, and probably > >> > unused, behaviour where relative paths were interpreted from the working > >> > directory. > >> > > >> > Absolute paths are kept untouched. > >> > >> Seems safe enough. Queued for default with a (BC) marker. > > > > Or maybe not. Something just destroyed my > > rsync-to-box-in-cloud-and-test script, and this is the most likely > > culprit. > > Ok, I'll resend on May. Meanwhile, it could be useful to have more > details abut such failure, if possible. Turns out this was probably the fault of the recent color extension bug and nothing wrong with your patch.
Patch
diff --git a/mercurial/dirstate.py b/mercurial/dirstate.py --- a/mercurial/dirstate.py +++ b/mercurial/dirstate.py @@ -103,7 +103,11 @@ files = [self._join('.hgignore')] for name, path in self._ui.configitems("ui"): if name == 'ignore' or name.startswith('ignore.'): - files.append(util.expandpath(path)) + p = util.expandpath(path) + if os.path.isabs(p): + files.append(p) + else: + files.append(self._join(p)) return ignore.ignore(self._root, files, self._ui.warn) @propertycache