- 12 Apr, 2019 2 commits
-
-
Matt Harbison authored
Converted with `2to3 -w -n -f dict .`, and then remove a bunch of list() wrapping.
-
Matt Harbison authored
Converted with `2to3 -w -n -f dict .`, and then remove a bunch of list() wrapping. The only place it seemed necessary was in repoview.py, as the collection is traversed more than once.
-
- 05 Apr, 2019 1 commit
-
-
muxator authored
-
- 06 Apr, 2019 3 commits
-
-
muxator authored
Let's try to better witespace-separate the various subfunctions, since in the next commits we are going to mess with them.
-
muxator authored
Without this patch, clicking on an unapplied mq patch causes the following stack trace: [...] ** Mercurial version (4.9+7-a62690659d6b+20190223). TortoiseHg version (4.9+25-a4f12d0cab7c) [...] Traceback (most recent call last): File "<base>/thg/tortoisehg/hgqt/revdetails.py", line 397, in _onFileSelected self.fileview.display(model.fileData(index)) File "<base>/thg/tortoisehg/hgqt/fileview.py", line 498, in display fd.load(self.isChangeSelectionEnabled()) File "<base>/thg/tortoisehg/hgqt/filedata.py", line 494, in load self.diff = ctx.thgmqpatchdata(wfile) File "<base>/thg/tortoisehg/util/patchctx.py", line 190, in thgmqpatchdata NameError: global name 'pycompat' is not defined This was a small overlook in 679593c2dbe6 (py3: replace cStringIO with pycompat.bytesio)
-
muxator authored
No functional changes
-
- 27 Mar, 2019 4 commits
-
-
Matt Harbison authored
This is kind of messy, and I've got no way to test bugtraq, so I don't want to change anything. There's no thg equivalent of pycompat that I know of, and the windows module handles this already, so it seems OK to use.
-
Matt Harbison authored
I'm not sure if these hoops are needed. I've got `configparser.py` under lib/site-packages, and `ConfigParser.py` under lib on pythpn 2.7.15. Listing the attributes on each looks similar, but there are more with the lowercase name.
-
Matt Harbison authored
-
Matt Harbison authored
-
- 24 Mar, 2019 1 commit
-
-
Matt Harbison authored
Fixed up after running `2to3 -w -n -f imports -f imports2 .`. There are other imports that need fixing, but not all have a pycompat alias. So I'm splitting these up. TortoiseHgOverlayServer.py and thg replace stdio with byte buffers, which won't work on py3. They need to be revisited to see if this is still necessary with py2exe[1]. In the meantime, I conditionalized a couple things there to prevent extra stacktraces on py3, since the local variable holding the buffer is used further down in each case. [1] https://groups.google.com/d/msg/thg-dev/dZbFjZugORs/sGOnxBpnBgAJ
-
- 23 Mar, 2019 1 commit
-
-
Matt Harbison authored
Fixed up after running `2to3 -w -n -f xrange .`.
-
- 27 Mar, 2019 1 commit
-
-
Matt Harbison authored
Spotted by Yuya.
-
- 24 Mar, 2019 2 commits
-
-
Matt Harbison authored
Fixed up after running `2to3 -w -n -f map .` It also replaced map+lambda with a list comprehension in qt_repotreemodel_test.py.
-
Matt Harbison authored
Converted with `2to3 -w -n -f idioms .`. The changes it wanted to make to tests/qt_repotreemodel_test.py are skipped because Yuya thought they are meant to test for specific class types. Additionally, a check in reporegistry.py of the selected indices when a drag is started is skipped, as the comment there implies it is testing for an exact type. Same goes for setModel() in repoview.py.
-
- 23 Mar, 2019 1 commit
-
-
Matt Harbison authored
Fixed up after running `2to3 -w -n -f zip .` Instances of `dict(zip(..))` in filedialogs.py, shelve.py, and status.py are unchanged, as dict can take the output of `zip()` on py2 and py3.
-
- 24 Mar, 2019 4 commits
-
-
Matt Harbison authored
Converted with `2to3 -w -n -f standarderror .`. Not sure if we should be catching and re-raising StopIteration et al prior to this handler to maintain the same behavior on py2 and py3.
-
Matt Harbison authored
Fixed up after running `2to3 -w -n -f nonzero .`.
-
Matt Harbison authored
Converted with `2to3 -w -n -f funcattrs -f methodattrs .`.
-
Matt Harbison authored
Converted with `2to3 -w -n -f filter .`
-
- 23 Mar, 2019 6 commits
-
-
Matt Harbison authored
Updated with `2to3 -w -n -f next .` The changes it wanted to make in backout.py and merge.py are bogus rewrites of QWizard::next().
-
Matt Harbison authored
-
Matt Harbison authored
This used `2to3 -w -n -f except .`, fixed an extra parenthesis in bugreport.py, and a commented out clause in pbranch.py.
-
Matt Harbison authored
Manually converted, but compared with `2to3 -w -n -f print .`
-
Matt Harbison authored
Caught by 2to3 wanting to convert it to `except ClassNotFound as TypeError`.
-
Matt Harbison authored
Converted with `2to3 -w -n -f raise .`, but excluded the changes in tests/helpers.py because it wants to use `.with_traceback(..)`.
-
- 03 Feb, 2019 2 commits
-
-
Matt Harbison authored
-
Matt Harbison authored
The tuple attributes are stored in unicode already. The only other place I question is in _phabsendopts(), where `**opts` will get a unicode value for the username. But that's how the tag command does it, so I assume this is proper.
-
- 20 Feb, 2019 1 commit
-
-
Manuel Jacob authored
It’s already implied by calling self.cslist.update().
-
- 24 Feb, 2019 2 commits
-
-
Yuya Nishihara authored
-
Steve Borho authored
--HG-- branch : stable
-
- 21 Feb, 2019 1 commit
-
-
Matt Harbison authored
Yet another ProgrammingError in __getitem()__ complaining '' isn't supported. I can't effectively test this because fiddling with the manifest view crashes thg.exe with the PyQt libraries that I have available locally. --HG-- branch : stable
-
- 26 Jan, 2019 1 commit
-
-
chrisjbillington authored
QTableView has word wrapping enabled by default, but Qt <= 5.11 had a bug such that it had no effect. Qt 5.12 fixes this bug, resulting in the repo view wrapping lines unless word wrapping is explicitly disabled. --HG-- branch : stable
-
- 22 Jan, 2019 4 commits
-
-
Steve Borho authored
--HG-- branch : stable
-
Yuya Nishihara authored
--HG-- branch : stable
-
Matt Harbison authored
I intend to find icons that distinguish between different roles, e.g. a regular user and a mailing list. With the reviewer info being saved locally to avoid the delay of hitting the server, I think it makes sense to save this out of the gate too. Otherwise, a user that builds up a reviewer list might never refresh it from the server.
-
Matt Harbison authored
For the repositories I work with, the same reviewers are selected the vast majority of the time. I still intend to experiment with the auto completion idea, but I think this is a useful convenience with or without that.
-
- 21 Jan, 2019 1 commit
-
-
Matt Harbison authored
This avoids needing to hit the server to get a list of the previously used reviewers. The list is tied to the configured server, instead of repo specific, to build up faster.
-
- 22 Jan, 2019 1 commit
-
-
Matt Harbison authored
This makes the following changes more readable, and models how scmutil.status does it.
-
- 19 Jan, 2019 1 commit
-
-
Yuya Nishihara authored
-