censor: accept censored revision during upgrade They can simply be passed by as censored.

censor: show that censored revision prevent repository upgrade This is not great.

smartset: don't ignore hidden revs when intersecting This fixes the bug I demonstrated in the previous commit, but I'm not sure at all if it's the right way of doing it.

tests: demonstrate crash in `unstable()` with internal-phase orphans

rust-matchers: fix quadratic complexity in `FileMatcher` Concretely, this command: ``` $ echo hg up -r <nodeid>; time hg revert dir1 dir2 -r <othernode> --debug hg up -r <nodeid> real 0m14.690s user 0m14.766s sys 0m5.430s ``` was much slower despite using 16 cores before this change. The approach taken here is the same one used in match.py, in exactmatcher. This changeset was originally written by Valentin Gatien-Baron in a private repository. I have redacted the commit message and did a minor clean up of the code.

revlog: add a small cache of unfiltered chunk This can provides a massive boost to the reading of multiple revision and the computation of a valid delta chain. This greatly help operation like `hg log --patch`, delta computation (helping pull/unbundle), linkrev adjustment (helping copy tracing). A first round of benchmark for `hg log --patch --limit 1000` shows improvement in the 10-20% range on "small" repository like pypy or mercurial and large improvements (about 33%) for more complex ones like netbeans and mozilla's. These speeds up are consistent with the improvement to `hg pull` (from a server sending poor deltas) I saw benchmarking this last year. Further benchmark will be run during the freeze. I added some configuration in the experimental space to be able to further test the effect of various tuning for now. This feature should fit well in the "usage/resource profile" configuration that we should land next cycle. When it does not provides a benefit the overhead of the cache seem to be around 2%, a small price for the big improvement. In addition I believe we could shave most of this overhead with a more efficent lru implementation.

revlog: minor refactor in the chunk gather process We will introduce some caching in this method in the next changeset, we make some of the most "disruptive" change first as touching this could break (and maybe did during the development process).

changelog-delay: move the delay/divert logic inside the (inner) revlog Instead of hacking throught the vfs/opener, we implement the delay/divert logic inside the `_InnerRevlog` and `randomaccessfile` object. This will allow to an alternative implementation of the `_InnerRevlog` that does not need to use Python details. As a result, the new implementation can use the transaction less agressively and avoid some extra output since no data had been written yet. That seems like a good side effect.

revlog: add a `canonical_index_file` attribute on inner revlog This is currently the same and the index_file but it will become more complex when we start doing delayed write.

changelog-delay: move "delayed" check to a more official API To avoid reaching inside the inner object in the future, we needs some official API. We put one of such API early to reduce the size of the final diff.

changelog-delay: move the appender class next to randomaccessfile We want to move the delay/divert logic at the revlog level (to have all IO related logic in the _InnerRevlog) we start with small piece that are easy to move on their own.

changelog-delay: adds some check around delaying and diverting write Theses assert shows we never call delay or divert if the revlog hold file handle on the revlog's file.

revlog: consolidate cache invalidation within the inner objet The invalidation of data that belong to the inner object can and should be delegated to the inner object.

shelve: drop some weird manually stripping before transaction abort I cannot find a case where this is actually useful, so lets triggering strip in the middle of uncommitted transaction.

revlog: move entry writing in the inner object This is the second big piece of API we want to live on the inner object. With this the inner object have freedom it needs to implement things internally. (except for the delayed write feature of the changelog, that will be migrated in coming changesets)

revlog: move `sidedata` in the inner object Same object as rawtext, but for sidedata.

revlog: move the `rawtext` method on the inner object This is one of the current core goal of the inner object: having a "simple" API for some of the core revlog action, like getting a raw text, with the inner object able to do that part internally, on its own.

revlog: move the_revisioncache on the inner object The goal for this inner object is to compute that things we cache, it make more sense to have the inner object handle it directly.

revlog: move the `deltachain` method on the inner object This is a necessary step before being able to move more logic around restoring a revision content there. For now, we do a simple patch for the perf extension logic, when the implementation of the inner object changes, we will likely need some evolution of the API. However this is true of many things in the perf extension. So we will see this later.

revlog: move the `_chunks` method on the inner object This is a necessary step before being able to move more logic around restoring a revision content there. For now, we do a simple patch for the perf extension logic, when the implementation of the inner object changes, we will likely need some evolution of the API. However this is true of many things in the perf extension. So we will see this later.

revlog: add a couple more of useful method on the inner object This will be needed for the next changeset.

revlog: move the `_chunk` method on the inner object This is a necessary step before being able to move more logic around restoring a revision content there. For now, we do a simple patch for the perf extension logic, when the implementation of the inner object changes, we will likely need some evolution of the API. However this is true of many things in the perf extension. So we will see this later.

revlog: move the compression/decompression logic on the inner object This is a necessary step before being able to move more logic around restoring a revision content there. For now, we do a simple patch for the perf extension logic, when the implementation of the inner object changes, we will likely need some evolution of the API. However this is true of many things in the perf extension. So we will see this later.

revlog: move the splitting-inline-revlog logic inside the inner object This is another large IO block that we need to move within the inner object if we want's it to be self sufficient.

revlog: synchronise the various attribute holding the index filename The segmentfile and the attribute need to be synchronized, let's enforce that.

revlog: drop reference to docket in the inline-splitting code revlog with a docket do not use inline revlog and do not need to split them. So we can remove some code handling docket there.

revlog: move _getsegmentforrevs on the internal object See inline documentation for details.

revlog: create a iteration of a _InnerRevlog object within the revlog The goal of this object is to isolate a sub-API that can be implemented by a compiled object (e.g. Rust). So the boundary of this object will be arbitrary depending of what can we easily implemented in the Compiled code. For now, we start simple, and move the code that manage the IO objects in the inner object. More will come in the coming changesets. Note: the object definition could live in the different module to thin the `revlog.py` file, however there are other better candidate for extraction first and I have enought patch stacked on top of the this one for the split in this patch not to be worth it. So I leave this to future me.

revlog: drop the unused `_chunkcache` attribute Apparently, some time ago, the chunk cache moved in the randomaccessfile object.

revlog: drop the unused `_sidedatareadfp` method It has no caller anywhere and is probably the remains of some older code.

revlog: also migrates `revlog.upperboundcomp` to ConfigClass This was planned but overlooked when doing the rest of the migration.

revlog: small doc to the `files` method

revlog: remove the `_indexfp` method The method is only used two time and we want to move that code to a lower level object. So we simply inline the method to simplify further refactoring.

revlog: avoid opening and closing the file for each cloned revision The previous code was flushing files after each new revision, slowing things down. For exemple, with this change, the evolve repository can run `hg debugupgraderepo --run --optimize re-delta-parent` in about 3.4s instead of 4.5 seconds.

censor: accept censored revision during upgrade They can simply be passed by as censored.

censor: show that censored revision prevent repository upgrade This is not great.

stream-clone: fix a crash when a repo with an empty revlog is cloned

debug-delta-chain: print less data by default This is faster and simpler to read.

debugdeltachain: add a parameter to display all info This will be useful with the next changeset that change the defaul output to display the minimum amount of information.

debug-delta-chain: actually skip unrequested computation Doing quick test on mozilla-unified show we can run up to 2× faster by skipping some of these computation.

debug-delta-chain: add options to control what we compute Currently this mostly controls what we display, but actual computation saving will come soon.

debug-delta-chaing: add a parameter to select revision to look at This allows for much faster runtime when we are interrested in some revisions only.

delta-chain: extract some debugdeltachain logic is object Moving from a closure to an object's method will help us to extend the command logic.

delta-chain: move the debugdeltachain command in revlogutils There is a dedicated `mercurial.revlogutils.debug` module were this code fits well.

dirstate: document the `changing_*` context manager The methods that requires them have documentation, but the context themselves had none. This is now fixed.

revlog: deprecate the compatibility config property Now that core is no longer use them, we can deprecated them. Let us aim at removing them in the next version.

revlog: remove legacy usage of `_lazydeltabase` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_lazydelta` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `canonical_parent_order` All core code is now getting the setting from the FeatureConfig object.

revlog: remove legacy usage of `_compute_rank` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_srmingapsize` All core code is now getting the setting from the DataConfig object.

revlog: remove legacy usage of `_srdensitythreshold` All core code is now getting the setting from the DataConfig object.

revlog: remove legacy usage of `hassidedata` All core code is now getting the setting from the FeatureConfig object.

revlog: remove legacy usage of `_sparserevlog` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_withsparseread` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_maxdeltachainspan` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_compengineopts` All core code is now getting the setting from the FeatureConfig object.

revlog: remove legacy usage of `_compengine` All core code is now getting the setting from the FeatureConfig object.

revlog: remove legacy usage of `_candidate_group_chunk_size` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_debug_delta` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_deltabothparents` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_maxchainlen` All core code is now getting the setting from the DeltaConfig object.

revlog: remove legacy usage of `_chunkcachesize` All core code is now getting the setting from the DataConfig object.

revlog: remove legacy usage of `_censorable` All core code is now getting the setting from the FeatureConfig object.

revlog: remove legacy usage of `_mmaplargeindex` All core code is now getting the setting from the DataConfig object.

revlog: remove legacy usage of `_checkambig` All core code is now getting the setting from the DataConfig object.

revlog: drop the _generaldelta attribute from the imanifeststorage interface It never belonged here.

revlog: remove legacy usage of `_generaldelta` All core code is now getting the setting from the DeltaConfig object.

revlog: use the new Config classes in _testrevlog the mock object need to follow the new interface. We allow ourself a small hacky import since this is testing code. The legacy attribute are still here because some code still use them. We will drop them when this code is updated.

revlog: skip opener options to pass sparse reading values We can directly set the option in the config object now.

revlog: skip opener options to pass sparse_revlog value We can directly set the option in the config object now.

revlog: skip opener options to pass mmap_index_threshold value We can directly set the option in the config object now.

revlog: skip opener options to pass enable_ellipsis We can directly set the option in the config object now.

revlog: skip opener options to pass compression option values We can directly set the option in the config object now.

revlog: skip opener options to pass compression_engine value We can directly set the option in the config object now.

revlog: skip opener options to pass max_chain_len We can directly set the option in the config object now.

revlog: skip opener options to pass max_deltachain_span We can directly set the option in the config object now.

revlog: skip opener options to pass lazy_delta values We can directly set the option in the config object now.

revlog: skip opener options to pass debug_delta value We can directly set the option in the config object now.

revlog: skip opener options to pass candidate_group_chunk_size value We can directly set the option in the config object now.

revlog: skip opener option to pass delta_both_parents value We can directly set the option in the config object now.

revlog: skip opener option to pass chunk_cache_size value We can directly set the option in the config object now.

revlog: create the revlog object at the repository level There is currently no value set in it, but we will be able to start centralise config parsing at the repository level.

revlog: overwrite revlog config through copy of the config object The new objects allow for this kind of blanket approach that make things cleaner. If we have more cases, it would probably deserve a context manager, but since we only have two usage, I don't think it is worth it.

revlog: move configuration attribute into dedicated object First this make things clearer as the number of configuration attributes has grown out of control, so gathering them make things clearer. Second, this will make future sharing them at different level simpler. Third it will allow us to make initialisation and temporary override simpler soon.

manifestrevlog: flag some inconsistency in bundle/union repo inheritance This "work" but still, this is weird, let us point it out.

randomaccessfile: drop explicit passing of file description The goal of this object is to manage IO, we still have to open the file if necessary, but this is all internal now.

revlog: drop more file description passing between private function They are no longer used as we are covered by the `reading` context

revlog: use a `reading` context in `_enforceinlinesize` We are about to enforce reading context on various operation, so we make sure top level method are in the right context. In the future we might move the responsibility of opening the revlog for reading higher in the call chain but lets limit the disruption for now.

revlog: drop the df argument to `sidedata` The intend of this argument is better filled by the `revlog.reading` or `revlog._writing` context. So we drop it to leave rooms for further cleanup and improvements.

revlog: drop the df argument to `rawdata` The intend of this argument is better filled by the `revlog.reading` or `revlog._writing` context. So we drop it to leave rooms for further cleanup and improvements.

revlog: drop the df argument to `revision` The intend of this argument is better filled by the `revlog.reading` or `revlog._writing` context. So we drop it to leave rooms for further cleanup and improvements.

interfaces: drop the `raw` parameters on ifiledata interface In practice it was not implemented since forever. The `rawdata` method fit that purpose. The interface checking does not detect that because it seems to confuse the `_df` argument for the `raw` one. Dropping that `_df` argument in a descendant changesets revealed the issue.

delta-computer: stop explicitly taking file handle The revlog has all the logic for opening and caching such handles, so no need to duplicate it here. In addition, this let the revlog handle that logic by itself which is better.

censors: simply use `revlog.reading` to keep things open dring rewrite We have a dedicated context now, we can simply use it. In practice, we cannot "simply" use it, as `finddeltainfo` still requires the file pointer to be passed explicitly. We will keep the old context manager open for a small bit until we get rid of it in a future changeset.

stream-clone: use `revlog.reading` in `revlog.get_streams` We can get the same result while using more standard method, so we do.

debug-revlog: keep the revlog open for the analysis duration This is more explicit. No performance differences were observed.

repo-upgrade: keep the revlog open while cloning This is more explicit. No performance differences were observed.

verify: keep the revlog open for reading while verifying it This is more explicit. No performances difference were observed.

convert: use `changelogrevision` to fetch changeset data instead of read This `changelogrevision` is a higher level function that will remain simple to use as we refactor the revlog.

revlog: adapt the `reading` check for `unionrepo` We cannot just rely on the length check for the `unionrepo` as the local revlog might be empty while the other revlog contains data. In addition, we need to also open the second revlog for reading when needed.

revlog: adapt the `reading` check for `bundlerepo` We cannot just rely on the length check for the `bundlerepo` as the local revlog might be empty with all data in the bundle.

revlog: make `reading` not crash on empty repository If the revlog is empty, the file might not exist and the open will fails. This is not great, but that details or this is now contained in the revlog itself.

changelog: make the `_delayopener` a real proxy We use the same approach as the `_divertopener`. This make it possible to use other methods of the vfs and not just the `__call__` API.

perf: ensure all readlog's reading is done within a `reading` context We are about to enforce this at the revlog level, so we update the perf code in advance.

perf: change the way we approach revlog reading If the `reading` context manager is available, we should use it over explicit file handle management. This will help us to make file handle management a matter more internal to the revlog.

revlog: document the `rev` method A small docstring does not hurt.

revlog: document the `clearcaches` method A small docstring does not hurt.

revlog: document the `update_caches` method A small docstring does not hurt.

revlog: make the `candelta` method private It is only used internally, so lets make it clear.

revlog: document the `get_streams` method A small docstring does not hurt.

revlog: document the `is_inline_index` method A small docstring does not hurt.

opemvms: add specific files and utilities

openvms: do not check for shebang in openvms script It does not apply to them.

openvms: consider worker costly on OpenVMS As for Windows, the parallel worker feature is less appealing on OpenVMS.

openvms: fix the pager spawning and cleanup Unsurprisingly, pager handling needs some adjustment.

openvms: make process spawning works on OpenVMS We need to adjust some behavior for OpenVMS.

openvms: mark symlink as unsupported on OpenVMS OpenVMS's symlink emulation is broken on some OpenVMS versions, so we better disable them altogether for now.

openvms: duck-punch a bugfix into `environb` object The official Python3 build for OpenVMS has some crippling bug that we need to patch dynamically OpenVMS patches

fastexport: simplify code

branching: merge stable into default

cleanup: typos, formatting

tests: backed out changeset 8037ddacad47 This issue with 3.11 was real, but it eventually got fixed after people started building exploit and filling CVE See https://github.com/python/cpython/issues/106242 And https://github.com/python/cpython/pull/106816

httppeer: fix static-http: scheme autodetection (issue6833) Since statichttprepo.make_peer() assumes path is a urlutil.path object (by accessing path.loc), we no longer can just provide a bytestring there. See also 5f71fff8dc74 and cfe8d88a453e.

admin-command: add verify command Start using the 'admin' namespace by adding a 'verify' command. Invocation is 'admin::verify'. The idea is to progressively add more focused checks than the existing verify command. To do so we need an advanced way to express what we want to check. The first check for admin::verify is 'working-copy.dirstate' which has no options, because it was an easy first check to implement, which verifies the integrity of the dirstate. This changeset was created with the help of Franck Bret.

transaction: abstract away the detection of an abandoned transaction We're about to add another place where this is used, let's not spill this implementation detail.

commands: add admin namespace In order to stop abusing the 'debug' namespace for non-debug command, it adds a new 'admin' namespace dedicated to 'admin' operations on a repository (i.e commands suitable for administration tasks). This namespace entry would be used to migrate in the future some existing commands from the 'debug' namespace, or other top level commands that may not be directly exposed to end users. (verify command is a perfect candidate for this case)

revlog: make the rust test for node hex prefix resolution exercise the nodemap

revlog: fix a bug where NULL_NODE failed to be resolved to NULL_REV The problem is that nodemap already takes care about NULL_NODE resolution (in `validate_candidate` in `nodemap.rs`), so the special handling in `rev_from_node` is unnecessary and incorrect.

tests: reproduce the nullrev bug in an end-to-end test

rust-config: fix incorrect coercion of null values to false As explained in the previous changeset: Probably being too trigger happy about boolean values, I incorrectly set the transform for a `None` to a `Some(false)`. It would cause for example the `ui.formatted` value to be set to `Some(false)`, which turns off the colors among other things, when `None` would trigger the automatic behavior.

rust-config: show default `null` is coerced incorrectly to `false` Probably being too trigger happy about boolean values, I incorrectly set the transform for a `None` to a `Some(false)`. It would cause for example the `ui.formatted` value to be set to `Some(false)`, which turns off the colors among other things, when `None` would trigger the automatic behavior. This is fixed in the next commit.

rust-config: fix fallback to default not parsing the default value When a config item's default is a string, it sometimes needs to be parsed into another type, like in the case of `cmdserver.max-log-size`, that returns a number of bytes from a human-readable amount like `25MB`. The logic for the fix is explained inline.

rust-config: demonstrate a bug when falling back to non-trivial default values The default value (when the user hasn't configured anything) is not run through the value parser, causing a programming error to happen because of type mismatch. This will be fixed in the next commit.

rust: make `Revision` a newtype This change is the one we've been building towards during this series. The aim is to make `Revision` mean more than a simple integer, holding the information that it is valid for a given revlog index. While this still allows for programmer error, since creating a revision directly and querying a different index with a "checked" revision are still possible, the friction created by the newtype will hopefully make us think twice about which type to use. Enough of the Rust ecosystem relies on the newtype pattern to be efficiently optimized away (even compiler in codegen tests¹), so I'm not worried about this being a fundamental problem. [1] https://github.com/rust-lang/rust/blob/7a70647f195f6b0a0f1ebd72b1542ba91a32f43a/tests/codegen/vec-in-place.rs#L47

rust: implement the `Graph` trait for all revlogs This is trivial and makes all the algorithms relying on the trait usable for more use cases.

rust: use the new `UncheckedRevision` everywhere applicable This step converts all revisions that shouldn't be considered "valid" in any context to `UncheckedRevison`, allowing `Revision` to be changed for a stronger type in a later changeset. Note that the conversion from unchecked to checked is manual and requires at least some thought from the programmer, although directly using `Revision` is still possible. A later changeset will make this mistake harder to make.

rust-revlog: split logic for `rawdata` to prepare for `UncheckedRevision` use In the next changeset, we will change the logic to use `UncheckedRevision` in the non-general delta case. The general delta case will use the "checked" path since `base_rev_or_base_of_delta_chain` will be checked.

rust: remove unused error variant All paths check that the working directory revision is not used.

rust: add `UncheckedRevision` type This is the start of a series whose aim is to separate "checked" and "unchecked" revision numbers. A "checked" revision number is valid for a given index, allowing us to have faster algorithms that don't do redundant checks all the time and have a clearer view of the kinds of revisions we're working with.

debugwireproto: migrate `opts` to native kwargs

debugwireargs: migrate `opts` to native kwargs

debugwalk: migrate `opts` to native kwargs

debug_revlog_stats: migrate `opts` to native kwargs

debugbackupbundle: migrate `opts` to native kwargs

debugsidedata: migrate `opts` to native kwargs

debugserve: migrate `opts` to native kwargs

debugrevspec: migrate `opts` to native kwargs

debugrevlogindex: migrate `opts` to native kwargs

debugrevlog: migrate `opts` to native kwargs

debugrename: migrate `opts` to native kwargs

debugrebuildfncache: migrate `opts` to native kwargs

debugpickmergetool: migrate `opts` to native kwargs

debugp2copies: migrate `opts` to native kwargs

debugp1copies: migrate `opts` to native kwargs

debugobsolete: migrate `opts` to native kwargs

debugmergestate: migrate `opts` to native kwargs

debugknown: migrate `opts` to native kwargs

debuginstall: migrate `opts` to native kwargs

debugindexdot: migrate `opts` to native kwargs

debuggetbundle: migrate `opts` to native kwargs

debugformat: migrate `opts` to native kwargs

debugfileset: migrate `opts` to native kwargs

debugextensions: migrate `opts` to native kwargs

debugdiscovery: migrate `opts` to native kwargs

debugdeltafind: migrate `opts` to native kwargs

debugdeltachain: migrate `opts` to native kwargs

debugdata: migrate `opts` to native kwargs

debugcapabilities: migrate `opts` to native kwargs

debugbundle: migrate `opts` to native kwargs This is a utility function that's only called by `debugbundle()`. The rest of the command was previously ported.

branching: merge with stable

tags: avoid expensive access to repo.changelog in a loop repo.changelog needs some cache invalidation when run on filtered repository. Accessing it in that loop can be expensive when there is many heads (e.g. mozilla try and it 25 000 heads). Note that the loop itself seems useless, but after this patch it no longer take about ⅛ of the time we spend computing cache for mozilla try. before : ! wall 0.350994 comb 0.350000 user 0.330000 sys 0.020000 (median of 28) after : ! wall 0.319520 comb 0.310000 user 0.290000 sys 0.020000 (median of 30)

blackbox: add a option to duplicate output to stderr too This is useful when debugging some performance operation.

perf: add a --update-last flag to perf::tags See inline documentation for details.

perf: add a `--clear-fnode-cache-rev` argument to perf::tags This will help us test this in situation closer to an update after a commit or a push.

perf: introduce more cache invalidation option in perf::tags This will help us to assert the performance of different versions in more diverse situations. We introduce helper function in the tags module in case invalidating such cache becomes different in the future.

hgweb: pass strings in WSGI environment correctly from wsgicgi According to PEP 3333, the strings in the WSGI environment must be translatable to bytes using the ISO-8859-1 codec.

branching: merge stable into default

Added signature for changeset 5a8b54201039

Added tag 6.5.2 for changeset 5a8b54201039

relnotes: add 6.5.2

run-tests: detect HGWITHRUSTEXT value Without this, some manual check in tests/test-dirstate.t could get confused by the lack of `rust` in module policy and break the test.

tests: avoid test environment affecting setup.py In particular [RHG_FALLBACK_EXECUTABLE] is being set prematurely, before rhg is built, but probably the rest of the env changes don't help, either.

setup: make the error "Unable to find a working hg binary" more informative The error message now shows the attempted hg commands and their stderr, to make it easier to investigate why things are not working. Here's an example output /!\ /!\ Unable to find a working hg binary /!\ Version cannot be extracted from the repository /!\ Re-run the setup once a first version is built /!\ Attempts: /!\ attempt #0: /!\ cmd: ['hg-missing', 'log', '-r.', '-Ttest'] /!\ exception: [Errno 2] No such file or directory: 'hg-missing': 'hg-missing' /!\ attempt #1: /!\ cmd: ['/usr/bin/python3', 'hg', 'log', '-r.', '-Ttest'] /!\ return code: 255 /!\ std output: /!\ std error: *** failed to import extension "topic": No module named 'topic' *** failed to import extension "evolve": No module named 'evolve' abort: accessing `dirstate-v2` repository without associated fast implementation. (check `hg help config.format.use-dirstate-v2` for details) /!\ /!\ Could not determine the Mercurial version /!\ You need to build a local version first /!\ Run `make local` and try again /!\

pycompat: drop usage of hasattr/getattr/setattr/delatt proxy The function remains to ease extensions transition, but we no longer use them in core.

safehasattr: drop usage in favor of hasattr The two functions should now be equivalent at least in their usage in core.

check-code: drop the `safehasattr` rule Since we no longer supports Python 2 we no longer needs to use a custom function to compensate the short comings of the global `hasattr` function in it. We cannot use the Python 3 `hasattr` function instead.

url: ignore some future pytype error As soon as we start using the builtins `hasattr` function, pytype will start getting confused about which types are available or not. So we ignore this error beforehand.

url: move the _generic_proxytunnel in the `has_https` block It is only used when https exists. If you stay out of the block, pytype complains that the type of its argument are not declared.

pycompat: deprecate using bytes Python2 has been dropped for a while, so lets comply to the signature of the global function. This open the way to drop the use of `pycompat.getattr` and company, and, especially, the associated `util.safehasattr`.

dynamic-import: use sysstr for importing extension and others This logic is used by extensions, and python hooks and merge-tools. All this logic eventually deals with native string (unicode in Python 3). This patch makes it handle `str` directly instead of relying on some pycompat low lever layer to do the conversion at the last minutes. We adjust the Python version filtering of a test as the output seems to be present with Python 3.7 too.

extension: check the command attributes using `sysstr` Since we are checking attributes, lets use the native representation instead of bytes.

extension: access special module members using sysstr These extensions variables and mapping are module attributes so they should be dealt with unicode `str` in Python 3. We move the part that deal with reading theses variable and checking their validity to use unicode `str` string.

sshpeer: convert command name to sysstr before accessing method Method name are `str` so we should pass them as `str`.

localrepo: purge filecache attribute using there unicode name This could be better, but that's a good step.

fastannotate: use sysstr to check for attribute presence We do not need bytes here.

fastannotate: use sysstr to deal with some attributes We are now Python3 only and attribute are always unicode string. So no needs to managed them using bytes.

hgweb: use sysstr to set attribute on diff option Attribute identifier should be `str` not `bytes`.

hgweb: use sysstr to get attribute on diff option Attribute identifier should be `str` not `bytes`.

diff-option: move attributes handling to sysstr Attributes are `str` and should be handled as such.

compression: use sysstr to specify attribute to fetch for priority These are attributes so they should be `str`.

smartset: set attribute using sysstr instead of bytes We do not needs bytes here.

cachedlocalrepo: use sysstr for attribute name of "file of interest" The first item is an attribute name, the second one is a filename. So the first one can (and should) be a `str`.

terse-status: use `sysstr` to specify attributes to set We use unicode string to specify attribute and we add a explicit char for the other usage.

path-suboption: deprecated specifying the attributes as bytes The known suboptions have been updated, so we can deprecated passing bytes in case some extensions use them.

path-suboption: use str for "_pushloc" suboptions That second argument refer to an attribute and do not needs to be bytes.

path-suboption: use str for "pushrev" suboptions That second argument refer to an attribute and do not needs to be bytes.

path-suboption: use str for "bookmarks_mode" suboptions That second argument refer to an attribute and do not needs to be bytes.

path-suboption: use str for "delta_reuse_policy" suboptions That second argument refer to an attribute and do not needs to be bytes.

path-suboption: use str for "multi_urls" path suboptions That second argument refer to an attribute and do not needs to be bytes.

convert: use sysstr to check for attribute presence We do not need bytes here.

mq: use sysstr to check for attribute access We do not need bytes here.

lfs: use sysstr to check for attribute presence We do not need bytes here.

mq: use sysstr to check for attribute presence We do not need bytes here.

relink: use sysstr to check for attribute presence We do not need bytes here.

ui: use sysstr to check for attribute presence We do not need bytes here.

remotefilelog: use sysstr to access for attributes We do not need bytes here.

hgwebmod: use sysstr to check for attribute presence We do not need bytes here.

remotefilelog: use sysstr to check for attribute presence We do not need bytes here.

extensions: use sysstr to check for attribute presence We do not need bytes here.

bundle2: use sysstr to check for attribute presence We do not need bytes here.

largefile: use sysstr to check for attribute presence in `hgsubrepoarchive` We do not need bytes here.

largefile: use sysstr to check for attribute presence in `openstore` We do not need bytes here.

largefile: use sysstr to check for attribute presence in `rebase` wrapping We do not need bytes here.

largefile: use sysstr to check for attribute presence in `getstatuswriter` We do not need bytes here.

run-tests: detect HGWITHRUSTEXT value Without this, some manual check in tests/test-dirstate.t could get confused by the lack of `rust` in module policy and break the test.

convert: stabilize subversion date sorter In case of tie, this provides unstable result on my system. So I am adding a couple of tie breaker to fix the test.

rhg: allow setting defaults.cmd to an empty string This is used by at least one hg UI to prevent defaults from affecting the command: https://github.com/emacs-mirror/emacs/blob/b71beb7ae7c60a5c5af608420d28fdda5265a264/lisp/vc/vc-hg.el#L245 Let's let it work.

uncommit: migrate `opts` to native kwargs

split: migrate `opts` to native kwargs

releasenotes: migrate `opts` to native kwargs

narrow: hoist a variable to a higher level to avoid use-before-init warning In practice, this shouldn't generate an IOError, so there wouldn't have been a problem. But PyCharm didn't know that.

narrow: migrate `opts` to native kwargs

largefiles: migrate `opts` to native kwargs

keyword: migrate `opts` to native kwargs

gpg: use a context manager to write `.hgsigs`

gpg: migrate `opts` to native kwargs

fetch: migrate `opts` to native kwargs

fastexport: migrate `opts` to native kwargs

closehead: wrap transaction handling in a context manager