thg issueshttps://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues2020-06-21T11:18:23Zhttps://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5355Filter Graph Bar Not Returning Correct Changesets2020-06-21T11:18:23ZBitbucket ImporterFilter Graph Bar Not Returning Correct Changesets*Created originally on Bitbucket by [gmb1226 (Greg Besack)](https://bitbucket.org/%7Ba2ea14bd-efe2-4789-8d6a-68324875819f%7D/)*
It seems the filter bar has some unexpected behavior when trying to find changesets using truncated hashes. ...*Created originally on Bitbucket by [gmb1226 (Greg Besack)](https://bitbucket.org/%7Ba2ea14bd-efe2-4789-8d6a-68324875819f%7D/)*
It seems the filter bar has some unexpected behavior when trying to find changesets using truncated hashes.
In 14_45_28, I highlight the commit I am trying to find: Revision 4600 with the truncated hash of d8db18eff05b.
In 14_45_39, I search for "d8", which returns many changesets, none of which contain rev 4600.
In 14_45_50, I search for "d8d", which returns one wrong commit.
In 14_46_00, I search for "d8db" which returns an empty set.
In 14_46_40, I search for "d8db1" which returns the correct commit.
I was expecting d8d to return the correct commit, based on it matching the first 3 characters of the hash d8db18eff05b.
In 14_58_35, I search for a different commit, "c09", which correctly returns rev 4530, with the hash c09a627e3ae4. This behavior does not seem consistent with above.
Separately, clicking on the filtered commit in 14_46_00 does not bring up the changelog for the commit. I think this may be broken. This may need a separate bug report.
*Attachments:* [2019-03-05_14_58_35-HiPower_Op_Code_hip_-_TortoiseHg_Workbench](/uploads/270ec0bb7ca2d65afb57750f4e16ca3f/2019-03-05_14_58_35-HiPower_Op_Code_hip_-_TortoiseHg_Workbench.png)[2019-03-05_14_45_50-HiPower_Op_Code_hip_-_TortoiseHg_Workbench](/uploads/1d661e1bfe57b18f097540559d9c9e78/2019-03-05_14_45_50-HiPower_Op_Code_hip_-_TortoiseHg_Workbench.png)[2019-03-05_14_46_00-HiPower_Op_Code_hip_-_TortoiseHg_Workbench](/uploads/fdd3ca9f88fc18679c31117d6e8ceda4/2019-03-05_14_46_00-HiPower_Op_Code_hip_-_TortoiseHg_Workbench.png)[2019-03-05_14_45_39-HiPower_Op_Code_hip_-_TortoiseHg_Workbench](/uploads/e4d72a6afb50b9542fe84f9d4ce50f49/2019-03-05_14_45_39-HiPower_Op_Code_hip_-_TortoiseHg_Workbench.png)[2019-03-05_14_45_28-HiPower_Op_Code_hip_-_TortoiseHg_Workbench](/uploads/2dd38a6721d917540678d6357495fa04/2019-03-05_14_45_28-HiPower_Op_Code_hip_-_TortoiseHg_Workbench.png)[2019-03-05_14_46_40-HiPower_Op_Code_hip_-_TortoiseHg_Workbench](/uploads/ad8c014060e7ed24d8e50207640990f2/2019-03-05_14_46_40-HiPower_Op_Code_hip_-_TortoiseHg_Workbench.png)https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5354Cannot add file after reverting largefile addition2020-06-21T11:18:15ZBitbucket ImporterCannot add file after reverting largefile addition*Created originally on Bitbucket by [Mojca Miklavec](https://bitbucket.org/%7Bcb53c37c-cf82-4243-83ba-46956c63ca0a%7D/)*
When user tries to add something as largefile, then changes his mind, reverts the addition and tries to add the fil...*Created originally on Bitbucket by [Mojca Miklavec](https://bitbucket.org/%7Bcb53c37c-cf82-4243-83ba-46956c63ca0a%7D/)*
When user tries to add something as largefile, then changes his mind, reverts the addition and tries to add the file again, the process fails.
Steps to reproduce:
* Create a file `foo.txt`
* Right-click on the file and choose `Add Largefiles...`
* Right-click on the same file, choose `Revert...`, answer `Revert`
* The file now seems to be back as if it was never added. The `hg status` commands also says `? foo.txt`, but the file `.hglf/foo.txt` is still present on the disk; note that this is also the case with command-line mercurial; manually adding and reverting a file also causes the file to remain inside `.hglf`.
* Right-click on the file, choose `Add`.
* Select the file, enter commit message and try to commit it. This step fails with the following error: `abort: .hglf/foo.txt: file not tracked! [Code: 255]`.
* On the other hand `hg commit -m "Add foo.txt"` from the command line seems to succeed.
The only workaround seems to be to manually delete the file `.hglf/foo.txt`, but there's absolutely no indication about this, and we had an user who kept struggling with the issue of being unable to commit the file until someone else helped him figure out what the actual problem was.
The behaviour seems to be specific to TortoiseHg, but I also suspect that mercurial was supposed to delete the file from `.hglf` in the first place.https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5348Running hg pull -u in a batch file prevents subsequent commands from running2020-06-21T11:18:09ZBitbucket ImporterRunning hg pull -u in a batch file prevents subsequent commands from running*Created originally on Bitbucket by [ekolis (Ed Kolis)](https://bitbucket.org/%7B952af717-5772-4ff3-9551-eb66749b1c88%7D/)*
I created the following batch file:
```
hg pull -u
autobuild.bat
pause
```
But when I run the batch fi...*Created originally on Bitbucket by [ekolis (Ed Kolis)](https://bitbucket.org/%7B952af717-5772-4ff3-9551-eb66749b1c88%7D/)*
I created the following batch file:
```
hg pull -u
autobuild.bat
pause
```
But when I run the batch file, only the first command executes; the autobuild and pause commands are never executed.
v4.9.0 on Windows 10https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5347Column widths in log reset if unapplied patch present2020-06-21T11:18:08ZBitbucket ImporterColumn widths in log reset if unapplied patch present*Created originally on Bitbucket by Anonymous*
Any apply-patch, unapply-patch, finish-patch MQ operation will reset the column widths in log view as long as any unapplied patches remain afterwards. That is, applying the last open patch ...*Created originally on Bitbucket by Anonymous*
Any apply-patch, unapply-patch, finish-patch MQ operation will reset the column widths in log view as long as any unapplied patches remain afterwards. That is, applying the last open patch will keep the widths intact.https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5343Windows Installer takes a long time to begin "Copying files" step2020-06-21T11:18:04ZBitbucket ImporterWindows Installer takes a long time to begin "Copying files" step*Created originally on Bitbucket by [psuter (Peter Suter)](https://bitbucket.org/%7B48b289e8-68dd-4aa4-8d3a-6458c5363d69%7D/)*
The Windows Installer seems to take a very long time at the beginning of the "Copying files" step (before the...*Created originally on Bitbucket by [psuter (Peter Suter)](https://bitbucket.org/%7B48b289e8-68dd-4aa4-8d3a-6458c5363d69%7D/)*
The Windows Installer seems to take a very long time at the beginning of the "Copying files" step (before the progress bar starts filling up). Is it actually doing something else there? (If yes, maybe displaying another status text would help the user understand that e.g. the installer has not "hanged".https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5340Error after replacing "localtags" file2020-06-21T11:18:02ZBitbucket ImporterError after replacing "localtags" file*Created originally on Bitbucket by Anonymous*
Replacing localtags file by copying it from another computer leads to error message in running thg. You mast copy file from one computer by network to another (those where now working thg)....*Created originally on Bitbucket by Anonymous*
Replacing localtags file by copying it from another computer leads to error message in running thg. You mast copy file from one computer by network to another (those where now working thg).
After error message normal work continues.https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5339turned to be not an hg repository after reboot server2020-06-21T11:18:00ZBitbucket Importerturned to be not an hg repository after reboot server*Created originally on Bitbucket by Anonymous*
Everything is ok utill I rebooted the server. Remote linkage and colon reported: "does not appear to be an hg repository", while it still locally works well on the server.
I'm using tortoi...*Created originally on Bitbucket by Anonymous*
Everything is ok utill I rebooted the server. Remote linkage and colon reported: "does not appear to be an hg repository", while it still locally works well on the server.
I'm using tortoiseHg 4.8.1 with official windows UI on server, and 3.8.1 / 4.8.2 on two different client.https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5335Rebase with collapse tries to open editor and hangs2020-06-21T11:17:54ZBitbucket ImporterRebase with collapse tries to open editor and hangs*Created originally on Bitbucket by [mjacob (Manuel Jacob)](https://bitbucket.org/%7B24e0601f-3292-4277-be76-bf4178891eda%7D/)*
## How to reproduce ##
Select two revisions to rebase (the source branch should contain more than one chang...*Created originally on Bitbucket by [mjacob (Manuel Jacob)](https://bitbucket.org/%7B24e0601f-3292-4277-be76-bf4178891eda%7D/)*
## How to reproduce ##
Select two revisions to rebase (the source branch should contain more than one changeset). Select "Collapse the rebased changeset" option. Click "Rebase".
## Expected behavior ##
There should be a way to edit the changeset description of the resulting changeset.
## Actual behavior ##
Mercurial tries to open an editor and hangs.https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5328Removing bookmark scrolls revision graph2020-06-21T11:17:46ZBitbucket ImporterRemoving bookmark scrolls revision graph*Created originally on Bitbucket by [psuter (Peter Suter)](https://bitbucket.org/%7B48b289e8-68dd-4aa4-8d3a-6458c5363d69%7D/)*
Hi, a minor irritation I often encounter is the following:
In TortoiseHG Workbench I right-click a changes...*Created originally on Bitbucket by [psuter (Peter Suter)](https://bitbucket.org/%7B48b289e8-68dd-4aa4-8d3a-6458c5363d69%7D/)*
Hi, a minor irritation I often encounter is the following:
In TortoiseHG Workbench I right-click a changeset with a bookmark, and select "Bookmark...", then click remove and close the dialog. This works fine, but now the changeset I had selected was scrolled away in a weird way. (The list of revisions sort of jumps around up and down.)
This is not the end of the world, but makes removing many bookmarks more tedious than necessary.
(Maybe it's more pronounced because I have many unapplied MQ patches?)
Thanks! ☃️https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5317Repository password visible in temporary directory2020-06-21T11:17:33ZBitbucket ImporterRepository password visible in temporary directory*Created originally on Bitbucket by Anonymous*
When click "Check for incoming changes" you can see your password in temporary directory name (although it's hidden in url part of the command):
% hg incoming --quiet --bundle "c:\temp\thg...*Created originally on Bitbucket by Anonymous*
When click "Check for incoming changes" you can see your password in temporary directory name (although it's hidden in url part of the command):
% hg incoming --quiet --bundle "c:\temp\thg.1p5jlc\http__MyLoginMyPassword@RepoAddress_jehtpn.hg" http://MyLogin:***@RepoAddresshttps://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5316Workbench: repository password visible in hint2020-06-21T11:17:33ZBitbucket ImporterWorkbench: repository password visible in hint*Created originally on Bitbucket by Anonymous*
When mouse cursor is over repositories combobox (Switched on in Settings/Sync/Target Combo) you can see repository password in hint.*Created originally on Bitbucket by Anonymous*
When mouse cursor is over repositories combobox (Switched on in Settings/Sync/Target Combo) you can see repository password in hint.https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5309Command line "thg revdetails" not working2020-06-21T11:17:26ZBitbucket ImporterCommand line "thg revdetails" not working*Created originally on Bitbucket by [DerColonel (Nikolaos Gkinos)](https://bitbucket.org/%7B709568af-882c-4f0e-84c9-f6e64c111aee%7D/)*
Unfortunately, the command line feature "thg revdetails" is not working. It crashes immediately.
S...*Created originally on Bitbucket by [DerColonel (Nikolaos Gkinos)](https://bitbucket.org/%7B709568af-882c-4f0e-84c9-f6e64c111aee%7D/)*
Unfortunately, the command line feature "thg revdetails" is not working. It crashes immediately.
Steps to reproduce:
* Open command line in an hg repository
* execute "thg revdetails [revid]" where "[revid]" is any revision id
* e. g. "thg revdetails 08fc4b2"
```
#!python
#!python
** Mercurial version (4.8.1). TortoiseHg version (4.8.1)
** Command: revdetails d1a734f5caf6
** CWD: D:\Projekte-local\Solutions
** Encoding: cp1252
** Extensions loaded: histedit, rebase, strip, fsmonitor, beautifygraph, tortoisehg.util.configitems
** Python version: 2.7.13 (v2.7.13:a06454b1afa1, Dec 17 2016, 20:53:40) [MSC v.1500 64 bit (AMD64)]
** Windows version: sys.getwindowsversion(major=6, minor=2, build=9200, platform=2, service_pack='')
** Processor architecture: x64
** Qt-5.9.2 PyQt-5.9.1 QScintilla-2.10.2.dev1711012104
Traceback (most recent call last):
File "tortoisehg\hgqt\run.pyo", line 70, in dispatch
File "tortoisehg\hgqt\run.pyo", line 273, in _runcatch
File "tortoisehg\hgqt\run.pyo", line 364, in runcommand
File "tortoisehg\hgqt\run.pyo", line 415, in _runcommand
File "tortoisehg\hgqt\run.pyo", line 369, in checkargs
File "tortoisehg\hgqt\run.pyo", line 363, in <lambda>
File "tortoisehg\hgqt\qtapp.pyo", line 459, in __call__
File "tortoisehg\hgqt\qtapp.pyo", line 523, in _createdialog
File "mercurial\util.pyo", line 1644, in check
File "tortoisehg\hgqt\run.pyo", line 1114, in revdetails
File "tortoisehg\hgqt\revdetails.pyo", line 543, in __init__
File "tortoisehg\hgqt\revdetails.pyo", line 83, in __init__
File "tortoisehg\hgqt\thgrepo.pyo", line 772, in __getitem__
File "mercurial\localrepo.pyo", line 1277, in __getitem__
ProgrammingError: unsupported changeid '' of type <type 'str'>
```https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5305Error adding files to new repository2020-06-21T11:17:22ZBitbucket ImporterError adding files to new repository*Created originally on Bitbucket by Anonymous*
Mercurial threw an error while adding files to a newly created repository. Note: the repository had a subrepo added to it, i.e. `.hgsub` file.
Here's the bug trace message...
```
...*Created originally on Bitbucket by Anonymous*
Mercurial threw an error while adding files to a newly created repository. Note: the repository had a subrepo added to it, i.e. `.hgsub` file.
Here's the bug trace message...
```
#!python
** Mercurial version (4.4.1). TortoiseHg version (4.4.1)
** Command: --nofork workbench
** CWD: C:\Users\abir\Projects\lads\tools\lads_tools
** Encoding: cp1252
** Extensions loaded:
** Python version: 2.7.13 (v2.7.13:a06454b1afa1, Dec 17 2016, 20:53:40) [MSC v.1500 64 bit (AMD64)]
** Windows version: sys.getwindowsversion(major=6, minor=2, build=9200, platform=2, service_pack='')
** Processor architecture: x64
** Qt-5.9.2 PyQt-5.9.1 QScintilla-2.10.2.dev1711012104
Traceback (most recent call last):
File "tortoisehg\hgqt\revdetails.pyo", line 392, in _onFileSelected
File "tortoisehg\hgqt\fileview.pyo", line 492, in display
File "tortoisehg\hgqt\filedata.pyo", line 700, in load
TypeError: _isdir() argument 1 must be encoded string without null bytes, not str
Traceback (most recent call last):
File "tortoisehg\hgqt\status.pyo", line 724, in run
File "mercurial\context.pyo", line 332, in sub
File "mercurial\subrepo.pyo", line 405, in subrepo
File "mercurial\subrepo.pyo", line 680, in __init__
File "mercurial\vfs.pyo", line 114, in exists
File "genericpath.pyo", line 26, in exists
TypeError: stat() argument 1 must be encoded string without null bytes, not str
```https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5295Explore command not working for UNC paths2020-06-21T11:17:12ZBitbucket ImporterExplore command not working for UNC paths*Created originally on Bitbucket by [Kuul Laager (Rando)](https://bitbucket.org/%7B83f22eb0-d09f-40b1-9ad9-a0a2f3f053db%7D/)*
Explore (open the repository in file browser) command does not work for UNC paths.
Windows explorer does not ...*Created originally on Bitbucket by [Kuul Laager (Rando)](https://bitbucket.org/%7B83f22eb0-d09f-40b1-9ad9-a0a2f3f053db%7D/)*
Explore (open the repository in file browser) command does not work for UNC paths.
Windows explorer does not open and no error messages are shown.
Running tortoisehg from console with SET THGDEBUG=1 shows this when clicking 'explore':
```
#!bash
ShellExecute '%5C%5Csome-pc$%5CUsers%5CMe%5CDesktop%5Cmyrepo' failed (error 2).
```
This worked fine in 4.3.1.
TortoiseHg 4.8.1 with Mercurial-4.8.1, Python-2.7.13, PyQt-5.9.1, Qt-5.9.2.
Windows 10 (1607).https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5294Home and End keys don't work in the revision list2020-06-21T11:17:11ZBitbucket ImporterHome and End keys don't work in the revision list*Created originally on Bitbucket by [DaveInCaz (David Hollman)](https://bitbucket.org/%7B2dd8bec0-e2b4-43d0-b274-6ee79db66b75%7D/)*
In the main list of revisions, the Home and End keys don't do anything.
I thought I remembered that t...*Created originally on Bitbucket by [DaveInCaz (David Hollman)](https://bitbucket.org/%7B2dd8bec0-e2b4-43d0-b274-6ee79db66b75%7D/)*
In the main list of revisions, the Home and End keys don't do anything.
I thought I remembered that they used to... so this seems like a regression.
In any case, using Home to get back to the working directory line item would be particularly useful.https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5279abandoned transaction issues2020-06-21T11:16:53ZMatt Harbisonabandoned transaction issuesI tried rebasing through the GUI in 4.8, not realizing there was an abandoned transaction. Hit the rebase button on the dialog, it greys out for a few seconds, then changes to an enabled "Continue". And then you can hit continue foreve...I tried rebasing through the GUI in 4.8, not realizing there was an abandoned transaction. Hit the rebase button on the dialog, it greys out for a few seconds, then changes to an enabled "Continue". And then you can hit continue forever, with it simply disabling briefly, and then re-enabling. Eventually I aborted it, opened the log window, and say this:
```
% hg rebase --dest=18761 --source=18757 --config=ui.merge=internal:merge
rebasing 18757:288d19619dff "settings: add a panel to support the Phabricator extension"
abandoned transaction found
[command returned code 255 Fri Nov 16 22:30:47 2018]
% hg rebase --continue --config=ui.merge=internal:merge
rebasing 18757:288d19619dff "settings: add a panel to support the Phabricator extension"
abandoned transaction found
[command returned code 255 Fri Nov 16 22:30:49 2018]
```
I assume there's some mechanism to catch that error and display a messagebox. Also notice that the hint isn't printed there, which is if run from the command line:
```
C:\Users\Matt\Projects\thg>hg rebase -s 18757 -d .
rebasing 18757:288d19619dff "settings: add a panel to support the Phabricator extension"
abort: abandoned transaction found!
(run 'hg recover' to clean up transaction)
```https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5260"Diff to parent" does nothing2020-06-21T11:16:30ZBitbucket Importer"Diff to parent" does nothing*Created originally on Bitbucket by [teo8976 (Teo Tei)](https://bitbucket.org/%7B4d381d15-4c61-4a29-9de3-d1b5a89dbb2a%7D/)*
Steps to reproduce:
- open any repository
- select the working directory or any revision where there are any f...*Created originally on Bitbucket by [teo8976 (Teo Tei)](https://bitbucket.org/%7B4d381d15-4c61-4a29-9de3-d1b5a89dbb2a%7D/)*
Steps to reproduce:
- open any repository
- select the working directory or any revision where there are any files listed
- in the list of modified files, right click on any of them that has been modified
- select "diff to parent"
Expected:
should open some tool to show the diff. If the tool isn't available or there is any issue opening it, an informative error message should be displayed
Observed:
Nothing at all happens.
I'm pretty sure this used to work and I think it used to open a program called "Meld Diff Viewer", which I still have and I can open it normally on its own.
*Attachments:* [Untitled](/uploads/8726109fd0f6eaf1f051db66738b0857/Untitled.png)https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5217Refresh error dialog when pruning '.' and amend is selected in commit widget ...2020-06-21T11:15:29ZMatt HarbisonRefresh error dialog when pruning '.' and amend is selected in commit widget in 4.6.1I amended the tip of a series last night, and pulled the hg-committed result this morning. The first three commits obsoleted, and I pruned the orphan in thg. I didn't realize I had left the commit widget with 'amend' still selected, an...I amended the tip of a series last night, and pulled the hg-committed result this morning. The first three commits obsoleted, and I pruned the orphan in thg. I didn't realize I had left the commit widget with 'amend' still selected, and after a few seconds, got a message saying "filtered revision '<some binary>'". There's no stacktrace or other error in the log window, and `--nofork --debug` doesn't seem to have anything useful.
```
...
locked, aborting
locked, aborting
stat: D:\mercurial\.hg\branch ((8L, 1536934998.8368795, 1536934998.8378775) -> (8L, 1536935112.8868291, 1536935112.8868291))
stat: D:\mercurial\.hg\dirstate ((79245L, 1536934999.5359898, 1536934999.539951) -> (79245L, 1536935112.9068289, 1536935112.9068289))
stat: D:\mercurial\.hg/store\obsstore ((17962530L, 1526304044.0672638, 1536934981.0) -> (17962656L, 1526304044.0672638, 1536935112.8868291))
data: _readparents ("!\xb7\x87=b2'7\xc5\x85\r\x07\x16/\xe8\xfd\xaf\x98\xfb\xf5\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00" -> '\xc5\xe6\xc1\xba\x1cy\xffw\x1b\xac\xaf\xa0\'\x06\xff\x06"X\xc30\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00')
change found (flags = 0xb)
invalid branchheads cache (visible): tip differs
invalid branchheads cache (served): tip differs
lfs: assuming remote store: https://www.mercurial-scm.org/repo/hg//.git/info/lfs
stat: D:\mercurial\.hg\dirstate ((79245L, 1536935112.9068289, 1536935112.9068289) -> (79245L, 1536935113.576829, 1536935113.576829))
change found (flags = 0x8)
GarbageCollector.check: 10063 1 9
collected gen 0, found 2848 unreachable
stopping service: D:\mercurial
closing repo: D:\mercurial
restoring excepthook
```
Unlike some other strange behaviors with amend and evolve that I still need to nail down, this one doesn't seem stateful. I can rollback and re-prune from the command line, and open and close thg, and reproduce it on demand. It doesn't seem to be an issue if 'commit' is active instead of 'amend'.
![thg small prune error.png](https://bitbucket.org/repo/48R4X/images/1824680402-thg%20small%20prune%20error.png)https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5199hgeol policy changes don't take effect without a restart2020-06-21T11:15:12ZMatt Harbisonhgeol policy changes don't take effect without a restartI have a repo where the policy started as:
```
[patterns]
.hgtags = LF
; Shell scripts are Unix LF
**.sh = LF
; Source files
**.c = CRLF
```
I recently changed the policy to LF, and did the update to null and back aga...I have a repo where the policy started as:
```
[patterns]
.hgtags = LF
; Shell scripts are Unix LF
**.sh = LF
; Source files
**.c = CRLF
```
I recently changed the policy to LF, and did the update to null and back again. But when I edited the file, I noticed that in the diff window, the entire file's content was marked as removed and re-added. The same is true when running `hg diff` in the log window. But `hg diff` in a terminal window correctly shows only the few added lines. Additionally, committing from tortoise will commit the wrong line endings. Using tortoise vs separate command terminal to do the updates didn't matter. Closing and reopening fixes the diff.
I thought maybe this was a command server thing, but I wasn't able to write a test that reproduces it.
```
$ hg init eol-change
$ cd eol-change
$ cat >> .hgeol << EOF
> [patterns]
> **.c = CRLF
> EOF
$ hg ci -Aqm 'add CRLF eol policy'
>>> with open('file.c', 'wb') as fp:
... fp.write('line1\r\nline2\r\nline3\r\n')
$ hg ci -Aqm 'add CRLF file'
$ cat > .hgeol << EOF
> [patterns]
> **.c = LF
> EOF
>>> from __future__ import absolute_import
>>> from hgclient import check, readchannel, runcommand
>>> @check
... def diff(server):
... readchannel(server)
... runcommand(server, ['ci', '-m', 'change eol policy'])
... runcommand(server, ['up', '-C', 'null'])
... runcommand(server, ['up', '-C', 'tip'])
... with open('file.c', 'wb') as fp:
... fp.write('line1\r\nline2\r\nline2.5\r\n\r\nline3\r\n')
... runcommand(server, ['diff'])
*** runcommand ci -m change eol policy
*** runcommand up -C null
0 files updated, 0 files merged, 2 files removed, 0 files unresolved
*** runcommand up -C tip
2 files updated, 0 files merged, 0 files removed, 0 files unresolved
*** runcommand diff
diff -r a5a9447127e7 file.c
--- a/file.c Thu Jan 01 00:00:00 1970 +0000
+++ b/file.c Thu Jan 01 00:00:00 1970 +0000
@@ -1,3 +1,5 @@
line1
line2
+line2.5
+
line3
```https://foss.heptapod.net/mercurial/tortoisehg/thg/-/issues/5166I sometimes get an "abort: missing support for check phases" error when I pus...2020-06-21T11:14:32ZBitbucket ImporterI sometimes get an "abort: missing support for check phases" error when I push from the workbench.*Created originally on Bitbucket by [ekolis (Ed Kolis)](https://bitbucket.org/%7B952af717-5772-4ff3-9551-eb66749b1c88%7D/)*
If I do, I can push again and it usually works.*Created originally on Bitbucket by [ekolis (Ed Kolis)](https://bitbucket.org/%7B952af717-5772-4ff3-9551-eb66749b1c88%7D/)*
If I do, I can push again and it usually works.