pytype: add assertions to explain revlogv2 invariants to Pytype Differential Revision: https://phab.mercurial-scm.org/D11241

typing: add several assertions to dirstatemap to appease pytype (grafted from default to stable) I think it's been mentioned in IRC that these can't be None in this case. This fixes: File "/mnt/c/Users/Matt/hg/mercurial/dirstatemap.py", line 213, in addfile: unsupported operand type(s) for &: 'None' and 'int' [unsupported-operands] No attribute '__and__' on None or '__rand__' on int Called from (traceback): line 290, in reset_state File "/mnt/c/Users/Matt/hg/mercurial/dirstatemap.py", line 214, in addfile: unsupported operand type(s) for &: 'None' and 'int' [unsupported-operands] No attribute '__and__' on None or '__rand__' on int Called from (traceback): line 290, in reset_state Differential Revision: https://phab.mercurial-scm.org/D11235

windows: avoid a bytes vs unicode crash reading passwords on py2 This broke in 5b3513177f2b. Specifically, after typing in the password on py2, it would crash with: TypeError: putwch() argument 1 must be cannot convert raw buffers, not str

streamclone: ensure the server sends the right amount of data Otherwise, the client would fail with some confusing error. I have seen an error which I think is this, perhaps due to a concurrent revlog split, which streamclones do not handle correctly and would result in a short read of the index of the revlog being split. Differential Revision: https://phab.mercurial-scm.org/D11236

rewriteutil: fix crash when a rewritten message references f{6,64} Without this, the rewriteutil logic thinks it's found a reference to the wdir pseudo-revision, and then tries to look it up and rewrite it. Stop it from doing that. Amusingly, I had trouble working with this changeset when I didn't describe the defect above using a regular expression, because it would trigger the bug in my installed version of hg. Differential Revision: https://phab.mercurial-scm.org/D11232

tests: add explicit coverage for update_hash_refs from rewriteutil I couldn't find any evidence this is covered by tests in core (but there's a good chance I missed it). We've seen a cute bug in that code, but first let's just cover the cases that work correctly. Differential Revision: https://phab.mercurial-scm.org/D11231

store: document the decoding discrepancy in store.py This will help future people that might be looking into this. Differential Revision: https://phab.mercurial-scm.org/D11220

clone: add a file with special character while testing uncompressed This will make sure we currently do not have any actual impact from issue6548. Differential Revision: https://phab.mercurial-scm.org/D11219

clone: test local clone in `test-clone-uncompressed.t` too This is not an uncompressed test but needs to be tested in the same kind of constraints regarding special file name. Differential Revision: https://phab.mercurial-scm.org/D11218

test-transaction-safety: relax some of the synchronisation schedule We can have `internal` and `external` runs at the same time, so we unlock `external` a bit sooner. Differential Revision: https://phab.mercurial-scm.org/D11216

test-transaction-safety: document the test schedule This helps reader to understand how the test work and actually helped me to detect a missing synchronisation step. Differential Revision: https://phab.mercurial-scm.org/D11215

rhg: Propagate permission errors when finding a repository The Rust standard library has a `Path::is_dir` method that returns false for any I/O error (such as a permission error), not just "No such file or directory". Instead add an `is_dir` function that returns false for non-directories and for "No such file or directory" errors, but propagates other I/O errors. Differential Revision: https://phab.mercurial-scm.org/D11230

rhg: Add build and config instructions to the README file This adds documentation explaining how to compile, configure, and use rhg as well as how the fallback mechanism works. Differential Revision: https://phab.mercurial-scm.org/D11229

dirstate: use `add` in the deprecation message about `add` The previous message was talking about `remove`. Differential Revision: https://phab.mercurial-scm.org/D11224

help: correct config.profiling.freq name (frequency->freq) Differential Revision: https://phab.mercurial-scm.org/D11217

run-tests: use a small timeout for chg instance There is case where the test runner fails to clean up the temporary files in that case, spawned chg instance can stay around for 1 hours. Getting them to shut down sooner cannot hurt. (We should also use a more robust approach for this cleanup) Differential Revision: https://phab.mercurial-scm.org/D11227

run-tests: do not inherit file descriptor when running a command This is one of the difference between python2 and python3 and could have been a reason why test hang with python2 + chg. This does not seems to help the hanging issue at all… However, now that this is written lets reduce the difference between python2 and python3. Differential Revision: https://phab.mercurial-scm.org/D11226

dummyssh: make sure we don't inherit files descriptor to the children This is one of the difference between python2 and python3 and could have been a reason why test hang with python2 + chg. This does not seems to help the hanging issue at all… However, now that this is written lets reduce the difference between python2 and python3. Differential Revision: https://phab.mercurial-scm.org/D11225

tests: make test-serve works on system that allow user to bind low port It is possible to encounter linux system configured in a way that allow port < 1024 to be bound by any users. So we update a test to focus more on the actual test: "resolving `daytime` to an actual port number. The new auto-scaling CI runner provided by clever cloud is doing so. This is the initial motivation for this change, but it seems useful anyway. Differential Revision: https://phab.mercurial-scm.org/D11228

git: restore basic functionality (issue6545) It looks like a big refactor happened on dirstate, and the git extension was just ignored. Differential Revision: https://phab.mercurial-scm.org/D11234

dirstate: fix typo in docstring Spotted while repairing git extension tests. Differential Revision: https://phab.mercurial-scm.org/D11233

typing: add several assertions to dirstatemap to appease pytype I think it's been mentioned in IRC that these can't be None in this case. This fixes: File "/mnt/c/Users/Matt/hg/mercurial/dirstatemap.py", line 213, in addfile: unsupported operand type(s) for &: 'None' and 'int' [unsupported-operands] No attribute '__and__' on None or '__rand__' on int Called from (traceback): line 290, in reset_state File "/mnt/c/Users/Matt/hg/mercurial/dirstatemap.py", line 214, in addfile: unsupported operand type(s) for &: 'None' and 'int' [unsupported-operands] No attribute '__and__' on None or '__rand__' on int Called from (traceback): line 290, in reset_state Differential Revision: https://phab.mercurial-scm.org/D11235

help: indicate how to run downgrade a repository using persistent nodemap We have a couple of report of people needing this. I am not fan of pointing people to a debug command, but we don't have much better to offer right now Differential Revision: https://phab.mercurial-scm.org/D11214

help: use the correct spelling for `slow-path` in persistent nodemap help Differential Revision: https://phab.mercurial-scm.org/D11213

fix: use scmutil.movedirstate() instead of reimplementing I wrote this patch 2 years ago as a little cleanup. I wanted to generally used `scmutil.movedirstate()` instead of manually updating the dirstate because that is easy to get wrong. I didn't know until today that the current code had a bug. So I added the test case two patches before this one and dusted off this one patch. This is a little slower than the previous code, as it diffs two manifests. However, it fixes the bug and I don't think it's going to be noticeably slower anyway. Differential Revision: https://phab.mercurial-scm.org/D11210

fix: rewrite writeworkingdir() to explicitly not work with merges `hg fix` errors out early if there is an unfinished merge, so we should have only one parent here. Making that explicit makes my next patches simpler. Differential Revision: https://phab.mercurial-scm.org/D11209

tests: demonstrate bug in `hg fix` with incorrectly dirty working copy If the parent commit needs formatting but the working copy already has the correct formatting, then the working copy will be reported as modified even though it's clean (because the size in the dirstate is incorrect). Because the bug only occurs when the size changes, I modified the formatter used in the test case to remove repeated spaces. Differential Revision: https://phab.mercurial-scm.org/D11208

upgrade: byteify an i18n message Differential Revision: https://phab.mercurial-scm.org/D11207

typing: convert an annotation to an assertion in commands.py Pytype was happy with the annotation at one point, but 2021.04.15 now complains. Differential Revision: https://phab.mercurial-scm.org/D11206

test: remove `sleep` usage in `test-nointerrupt.t` (issue6271) We move from sleep based synchronisation to file creation based synchronisation. Sleeps is the path to the dark side. Sleeps leads to flakiness. Flakiness leads to anger. Anger leads to hate. Hate leads to suffering. Differential Revision: https://phab.mercurial-scm.org/D11205

Added signature for changeset d7515d29761d

Added tag 5.9rc0 for changeset d7515d29761d

branching: merge default into stable This mark the start of the 5.9 freeze.

relnotes: document the change in the dirstate API Differential Revision: https://phab.mercurial-scm.org/D11200

dirstate: deprecate the `add` method All users have been migrated. Differential Revision: https://phab.mercurial-scm.org/D11199

mq: replace `add` call with newer API Differential Revision: https://phab.mercurial-scm.org/D11198

largefile: use `update_file` instead of `add` in `synclfdirstate` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11197

largefile: use `update_file` instead of `add` in `mergerecordupdates` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11196

dirstate: deprecate the `drop` method All users have been migrated. Differential Revision: https://phab.mercurial-scm.org/D11195

narrow: stop using `drop` in the `updateworkingcopy` Let us use the new API instead. Differential Revision: https://phab.mercurial-scm.org/D11194

mq: replace `drop` call with newer API Differential Revision: https://phab.mercurial-scm.org/D11193

largefile: use `update_file` instead of `drop` during `rollback` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11192

largefile: use `update_file` instead of `drop` in `synclfdirstate` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11191

sparse: use `update_file` instead of `drop` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11190

sparse: use `update_file` instead of `drop` in `refreshwdir` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11189

amend: use `update_file` instead of `drop` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11188

context: use `update_file` instead of `drop` in `markcommitted` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11187

dirstate: deprecate the `merge` method in all cases All code have been migrated to the new APIs. Differential Revision: https://phab.mercurial-scm.org/D11186

mq: drop the use of `dirstate.merged` Since mq does not work on merge, we do not need this. Differential Revision: https://phab.mercurial-scm.org/D11185

dirstate: deprecate the `otherparent` method in all cases All code have been migrated to the new APIs. Differential Revision: https://phab.mercurial-scm.org/D11184

dirstate: deprecate the `normallookup` method in all cases All code have been migrated to the new APIs. Differential Revision: https://phab.mercurial-scm.org/D11183

narrow: stop using `normallookup` during a test Differential Revision: https://phab.mercurial-scm.org/D11182

narrow: use `update_file` instead of `normallookup` after update This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11181

test: use the internal `_normallookup` in `test-rebuildstate.t` This is a low level test doing low level manipulation. Differential Revision: https://phab.mercurial-scm.org/D11180

sparse: use `update_file` instead of `normallookup` in refreshwdir This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11179

keyword: use `update_file_p1` instead of `normallookup` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11178

fix: use `set_possibly_dirty` instead of `normallookup` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11177

eol: use `set_possibly_dirty` instead of `normallookup` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11176

mq: replace `normallookup` call with newer API This align MQ behavior with what what `scmutil.movedirstate` does. (it should probably use `scmutil.movedirstate` itself. Differential Revision: https://phab.mercurial-scm.org/D11175

revert: use `set_possibly_dirty` instead of `normallookup` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11174

dirstate: enforce `possibly_dirty` in `set_tracked` There is some usecase (like `hg revert`) that requires more agressive `possibly_dirty` setting. So we take a safe path and enforce it in one of the main `normallookup` replacement. Differential Revision: https://phab.mercurial-scm.org/D11173

record: use `update_file` instead of `normallookup` We are cheating a bit to enforce we are in a pending change context. This will have to be cleaned up later. The end goal is reached, we are now using the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11172

copy: use `set_tracked` instead of `normallookup` in `dirstatecopy` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11171

amend: use `update_file` instead of `normallookup` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11170

mergestate: use `update_file` to handle for `ACTION_EXEC` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11169

mergestate: use `update_file` to handle for `ACTION_ADD_MODIFIED` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11168

largefile: use `update_file` for `synclfdirstate` "m" case This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11167

largefile: use `update_file` for `synclfdirstate` "n" case This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11166

largefile: replace the `normallookup` call when creating a new lfdirstate This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11165

largefile: use `update_file` instead of `normallookup` in `updatelfiles` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11164

largefile: use `set_possibly_dirty` instead of `normallookup` Differential Revision: https://phab.mercurial-scm.org/D11163

dirstate: add a `set_possibly_dirty` method This method will be able to replace some usage of `normallookup` in the future. Differential Revision: https://phab.mercurial-scm.org/D11162

dirstate: deprecate the `normal` method in all cases All code have been migrated to the new APIs. Differential Revision: https://phab.mercurial-scm.org/D11161

test: use `set_tracked` in `test-context` This is apparently the last user of the old API. Differential Revision: https://phab.mercurial-scm.org/D11160

mq: replace usage of `normal` with newer API Differential Revision: https://phab.mercurial-scm.org/D11159

largefile: use `update_file` instead of `normal` in `synclfdirstate` This is the newer, more semantic API. This was the last call to `normal` in largefile. Differential Revision: https://phab.mercurial-scm.org/D11158

largefile: use `set_clean` instead of `normal` in `repo.status` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11157

largefile: use `set_clean` instead of `normal` in `lfdirstatestatus` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11156

largefile: use `update_file` instead of `normal` in `updatelfiles` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11155

largefile: respect the `normal` signature The `parentfiledata` parameter was missing, we throw it away because I am not sure what to do with it in the largefile case. Differential Revision: https://phab.mercurial-scm.org/D11154

keyword: use `set_clean` instead of `normal` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11153

sparse: use `update_file` instead of `normal` in `refreshwdir` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11152

sparse: use `update_file` instead of `normal` during `applyupdates` This is the newer, more semantic API. (more content end up "nonnormal", see previous commit) Differential Revision: https://phab.mercurial-scm.org/D11151

sparse: apply update with in a `parentchange` context We are changing the reference so we should be within such context manager. This seems make the actual dirstate content a bit flaky with content being sometime ambigous (but maybe it was flaky before?) So I increased the matching. Differential Revision: https://phab.mercurial-scm.org/D11150

revert: use `set_clean` instead of `normal` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11149

context: use `update_file` instead of `normal` in `markcommitted` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11148

context: use new API instead of `normal` during fixup post status We use `set_clean` or `update_file` depending of if we are in the middle of an update or not. Differential Revision: https://phab.mercurial-scm.org/D11147

dirstate: introduce a `set_clean` method This will provide a good alternative to `normal` call with a narrower semantic. Differential Revision: https://phab.mercurial-scm.org/D11146

largefile: use `update_file` instead of `normal` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11145

largefile: adjust the dirstate post update within a `parentchange` context This is doing an update, so this seems more appropriate. No other change have been made, but the next changeset will :-) Differential Revision: https://phab.mercurial-scm.org/D11144

dirstate: deprecate `dirstate.remove` in all cases This will make sure that code will use the newer, more semantic API sooner than later. Differential Revision: https://phab.mercurial-scm.org/D11143

largefile: use `update_file` instead of `remove` in `synclfdirstate` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11142

largefile: rearrange conditionnal in `synclfdirstate` We can liquidate the special case early, so do we. Differential Revision: https://phab.mercurial-scm.org/D11141

largefile: use `update_file` instead of `remove` in `mergerecordupdates` This is the newer, more semantic API. Differential Revision: https://phab.mercurial-scm.org/D11140

mq: use `update_file_p1` instead of `remove` when adjusting dirstate This is the new API. Differential Revision: https://phab.mercurial-scm.org/D11139

dirstate: properly update `_lastnormaltime` in `update_file_p1` This was previously overlooked. Differential Revision: https://phab.mercurial-scm.org/D11138