worker: raise exception instead of calling sys.exit() with child's code When a worker process returns an error code, we would call `sys.exit()` with that exit code on the main process. The `SystemExit` exception would then get caught in `scmutil.callcatch()`, which would return that error code. The comment there says "Commands shouldn't sys.exit directly", which I agree with. This patch changes it so we raise a specific exception when a worker fails so we can catch instead. I think that means that `SystemExit` is now always an internal error. (I had earlier thought that this call to `sys.exit()` was from within the child process until Matt Harbison made me look again, so thanks for that!) Differential Revision: https://phab.mercurial-scm.org/D9287

config: read system hgrc in lexicographical order This is similar to edbcf5b239f9 (config: read configs from directories in lexicographical order, 2019-04-03). Apparently I forgot to sort the system hgrc files there. That's fixed by this patch. Differential Revision: https://phab.mercurial-scm.org/D9269

relnotes: drop 5.6 release entries from next Differential Revision: https://phab.mercurial-scm.org/D9282

merge with stable

makefile: use Python 3 by default (BC) This change is long overdue IMO. .. bc:: Makefile now uses `python3` instead of `python` by default on non-Windows platforms. This means Mercurial will be built and run with Python 3 instead of Python 2.7 by default. To continue using Python 2, set the PYTHON variable. e.g. `make install PYTHON=python2.7`. Differential Revision: https://phab.mercurial-scm.org/D7258

hgweb: don't call sys.exit() in httpservice.run() If I'm reading the code correctly, `mercurial.server.createservice()` can return an hgweb service or one of three types of command server services. The caller then calls `mercurial.server.runservice()`, passing it the returned service's run method. Only the hgweb service was calling `sys.exit()`. It has been that way since 8d44649df03b (refactor ssh server., 2006-06-04). That commit message doesn't provide any explanation. Let's clean up and have the code follow the usual return path into the `dispatch` module. After this patch, there should be no remaining places left where we call `sys.exit()` except for valid uses in the `dispatch` and `worker` modules. Differential Revision: https://phab.mercurial-scm.org/D9272

serve: simply return instead of calling sys.exit() in `hg serve --stdio` The shouldn't be a reason to call `sys.exit()` instead of letting the code return normally. I've remove the call in both `hg serve` and `hg debugserve`. Differential Revision: https://phab.mercurial-scm.org/D9271

httpservice: move sys.exit() out of serve_forever() This is a simple refactoring to show the callers of the method, so it's easier to reason about the impact of removing the `sys.exit()` calls in subsequent patches. Differential Revision: https://phab.mercurial-scm.org/D9270

copies: handle more cases where a file got replaced by a copy This patch fixes the changeset-centric version in a pretty straight-forward way. It fixes it to automatically resolve the conflict, which is better than resulting in a modify/delete conflict as it was before b4057d001760 (merge: when rename was made on both sides, use ancestor as merge base, 2020-01-22). I'll leave it for later to test and explicitly handle cases where files have been renamed to the same target on different sides of the merge. Differential Revision: https://phab.mercurial-scm.org/D8653

tests: test more cases where a file got replaced by a copy This adds a test where a file is modified on one branch and is renamed onto another file in another branch. That should ideally be automatically resolved (by propagating the modification to the rename destination). Alternatively, it could be considered a modify/delete conflict. It should at least not be automatically resolved by ignoring the modification. However, that is what actually happens with the changeset-centric algorithm since I broke it in b4057d001760 (merge: when rename was made on both sides, use ancestor as merge base, 2020-01-22). Before that commit, it resulted in a modify/delete conflict. The filelog-centric algorithm was broken already before that commit. Differential Revision: https://phab.mercurial-scm.org/D8652

unionrepo: don't insert index tuples with None as int field None is not a valid size. Use -1 as placeholder instead. This will be necessary when the index starts enforcing type correctness. Differential Revision: https://phab.mercurial-scm.org/D9161

bundlerepo: don't insert index tuples with full nodes as linkrev The index format has a documented format and latter changes will start to enforce the field types. The bundlerepo uses full nodes for the linkrev field when it should be using revision numbers. Use the link mapping to resolve them, except in the special case of self-references. Those are actually indications of a missing linkrev. Differential Revision: https://phab.mercurial-scm.org/D9160

rhg: add full node id support for `debugdata` command Unlike other later implemented commands `debugdata` only supported revision number. This changeset add full node id support for consistency with other commands. Differential Revision: https://phab.mercurial-scm.org/D9230

commit: warn the user when a commit already exists Sometimes, a commit will result in an exact match of a preexisting commit, and if that commit isn't a branch head, hg will incorrectly note that it created a new head. Instead, we should warn the user that commit already existed in the repository. In practice, this bug is rather uncommon, and will only occur when the usr explicitly sets the date. Please note that this commit contains an API change to cmdutil.commitstatus() Differential Revision: https://phab.mercurial-scm.org/D9257

revlog: don't cache parsed tuples in the C module A cached entry creates ~8 Python objects per cached changeset, which comes to around 200 Bytes per cached changeset on AMD64. Especially for operations that touch a lot of changesets, that can easily sum up to more than a 100MB of memory. Simple tests on large repositories show <2% runtime penalty for ripping out the cache, even for cache heavy operations like "hg log" for all revisions. Differential Revision: https://phab.mercurial-scm.org/D9155

fix: only check for obsolete commits in the --rev case With both `--all` and `--source`, we already exclude obsolete revisions in the revset, so there's no need to call `checkfixablectx()` in those cases. Differential Revision: https://phab.mercurial-scm.org/D9227

fix: don't include obsolete descendants with -s The `-s/--source` option is for regular users (`-r` is there for power users). If there are obsolete commits that are descendants of the given revision(s), then they almost definitely should just be left alone. That's what `hg rebase` does as well. So this patch makes it so we skip obsolete commits (including those in the input set itself). Differential Revision: https://phab.mercurial-scm.org/D9226

tests: add test showing how `hg fix -s` deals with obsolete and orphan nodes We didn't have any tests for how `hg fix -s` behaves with obsolete commits among the descendants. The next patch will change the behavior in this area. Differential Revision: https://phab.mercurial-scm.org/D9225

fix: suggest --source instead of --rev on empty revset `--source` is the recommended flag for regular users (`--rev` is available for advanced users). Differential Revision: https://phab.mercurial-scm.org/D9224

hg-core: fix path encoding usage 1. Hash encoded path are in `.hg/store/dh` instead of `.hg/store/data`. 2. Path encoded index and data files may not have the same parent path. It is not just about replacing `.i` by `.d` Differential Revision: https://phab.mercurial-scm.org/D9121

push: add --all-bookmarks option Differential Revision: https://phab.mercurial-scm.org/D9261

rhg: strip copied files metadata from `cat` output Differential Revision: https://phab.mercurial-scm.org/D9264

merge with stable

chg: close file descriptors when starting the daemon It's good practice to close file descriptors when forking to start a daemon. This did not appear to happen yet, which results in flock hanging in one location in our system (because the chg daemon keeps the locked file open). Differential Revision: https://phab.mercurial-scm.org/D9268

crecord: render chunkpad on Windows (issue6427) When using Windows wrappers of PDCurses (e.g., windows-curses), the chunkpad does not render when executing `hg commit -i`. This is due to attempting to refresh one too many columns of the pad. Differential Revision: https://phab.mercurial-scm.org/D9267

Added signature for changeset 18c17d63fdab

Added tag 5.6 for changeset 18c17d63fdab

copies: make sure deleted copy info do not overwriting unrelated ones See test documentation for details. This is yet another corner case for copy tracing. Differential Revision: https://phab.mercurial-scm.org/D9200

utils: helper function to print top memory allocation site The memorytop function uses Python's tracemalloc module to show the source lines / backtraces with the largest remaining allocations. This allows identifying the origins of active memory by placing calls in strategic locations. Allocations from C extensions will show up as long as they are using the Python allocators. Differential Revision: https://phab.mercurial-scm.org/D9236

templates: include all non-branch namespaces in default one-line summary I left out branches and custom namespaces on purpose from D9252 because I figured that people like us (Google) who have custom namespaces can also have custom configs. However, I just realized that this makes everyone with the topic extension lose the topic they've had in rebase output for a long time (ever since someone was nice enough to add it in D741). Sorry about the churn. The more generic template couldn't easily keep the "log.bookmark" label in the template because the namespace is called "bookmarks" (plural). That means that we can't be compatible with users' existing configs for "log.bookmark", so I decided to change the labels to be in a brand-new "oneline-summary" namespace. Differential Revision: https://phab.mercurial-scm.org/D9262

relnotes: mention improved memory use and underlaying API changes Differential Revision: https://phab.mercurial-scm.org/D9258

branching: merge with stable

help: update command synopsis to clarify "cp --forget" only takes destinations I'm a bit confused while reading 03690079d7dd, which says "a destination file", but the code loops over matched files.

rebase: update commit hash references in the new commits This excludes the --collapse case because degenerating to p1 is almost certainly as wrong as leaving the old hashes in place. I expect most people to amend the message explicitly when using that. Differential Revision: https://phab.mercurial-scm.org/D9229

histedit: drop fallback to empty string from rendertemplate() AFAICT, `cmdutil.rendertemplate()` always returns bytes (never e.g. `None`), so we don't need to fall back to empty (byte-)string. The fallback has been there since the code was added in 11c076786d56 (histedit: add templating support to histedit's rule file generation, 2019-01-29). Differential Revision: https://phab.mercurial-scm.org/D9244

phases: convert registernew users to use revision sets Differential Revision: https://phab.mercurial-scm.org/D9233

phases: allow registration and boundary advancement with revision sets The core internals either use revision sets already or can trivially use them. Use the new interface in cg1unpacker.apply to avoid materializing the list of all new nodes as it is normally just a revision range. This avoids about 67 Bytes / changeset on AMD64 in peak RSS. Differential Revision: https://phab.mercurial-scm.org/D9232

revlog: extend addgroup() with callback for duplicates The addgroup() interface currently doesn't allow the caller to keep track of duplicated nodes except by looking at the returned node list. Add an optional second callback for this purpose and change the return type to a boolean. This allows follow-up changes to use more efficient storage for the node list in places that are memory-sensitive. Differential Revision: https://phab.mercurial-scm.org/D9231

tags: add safety check for len(record) while reading hgtagsfnodescache I am trying to fix a breakage where somehow we end up getting a node of 12 length from `getfnode()`. Understanding the hgtagsfnodescache code, it seems highly unlikely that it can happen unless one of `mctx.readfast().get()` or `ctx.filenode()` is returning a node of 12 length. For safety, I think it's better to add a check to make sure that record which we are parsing is of same length we are expecting otherwise we consider that as invalid record. Differential Revision: https://phab.mercurial-scm.org/D9169

procutil: allow to specify arbitrary stdin bytes to runbgcommand For automatic clonebundles generation I need to pass arbitrary large amount of data to the process (eg: common nodes, target nodes). I am updating the `runbgcommand` to allow for this. Previously not stdin input was possible, now, one can provide raw bytes and they will be feed to the command through an unnamed temporary files. Differential Revision: https://phab.mercurial-scm.org/D9212

clonebundle: move the manifest filename to a constant I am about to add more reference to it, so I would rather have it an explicit constant. This allow to unify various call too. Differential Revision: https://phab.mercurial-scm.org/D9209

clonebundles: move a bundle of clone bundle related code to a new module In the process on general clone bundle automatically, we need to make some function available more widely. This is a good opportunity to extract a significant amount of code from `mercurial.exchange` into a new `mercurial.bundlecaches`. This make `mercurial.exchange` move under the 3K line range (hooray…). The module is called `bundlecaches` because I expect it to be eventually useful for more than just clone bundle (like pull bunbles). Differential Revision: https://phab.mercurial-scm.org/D9208

config: add a .hg/hgrc-not-shared which won't be shared in share-safe mode Previous patches add a safe mode for sharing repositories which involve sharing of source requirements and config files. In certain situations we might need to add a config to source repository which we does not want to share. For this, we add a `.hg/hgrc-not-shared` which won't be shared. This also adds a `--non-shared` flag to `hg config` command to see the non-shared config. Differential Revision: https://phab.mercurial-scm.org/D8673

dispatch: load shared source repository config in share-safe mode It seems to me now that there are two steps when config is loaded: 1) on dispatch 2) repository object creation Recent patches added functionality that there can be shares in share-safe mode where config of the source repository is shared with the the shares. However we missed adding logic to read the source config on dispatch. This leads to extensions not being loaded on dispatch and hence extensions command not being recognized. This patch fixes it by reading the shared source config on dispatch. Differential Revision: https://phab.mercurial-scm.org/D9047

tests: add test showing broken extension loading in case of share-safe When we are using share-safe and loading source repository config, we are still not loading the extensions which are enabled in the source repository. This patch demonstrates the bug. Meanwhile it also shows how unreliable `hg debugextensions` is. Differential Revision: https://phab.mercurial-scm.org/D9046

dirstate-tree: simplify the control flow in the Node.insert method But explicitly with the special case early, laying out the various case become simpler. (The initial motivation was to make some future lifetime error simpler). Differential Revision: https://phab.mercurial-scm.org/D9203

revlog: use LRU for the chain cache For a large repository, this reduces the number of filelog instances and associated data a lot. For a 1% speed penalty, it reduces peak RSS by 20% for the full NetBSD test repository. Differential Revision: https://phab.mercurial-scm.org/D9235

tests: add notes about broken `hg log --follow <file>` with copies in extras I also removed some unnecessary `#if no-changeset` where the `#else` was the same :P Differential Revision: https://phab.mercurial-scm.org/D9204

config: move message about leading spaces in config to config.py When the config parser raises a ParseError, it uses the line that failed to parse as the error message. It doesn't currently tell the user anything about why it failed to parse. b13b99d39a46 (config: highlight parse error caused by leading spaces (issue3214), 2014-03-16) added a checked based on the error *message* having leading spaces. That has worked fine because only the config parser uses the line itself as error message (I think the revset and fileset parsers use more user-friendly proper messages). It still feels like a hack. Let's make the config parser give a useful message about leading whitespace instead. We should ideally follow up with more useful messages for other parse errors in config files. Differential Revision: https://phab.mercurial-scm.org/D9241

errors: name arguments to ParseError constructor As with similar previous patches, this is to improve readability. Differential Revision: https://phab.mercurial-scm.org/D9240

errors: remove unnecessary override of __bytes__ in RevlogError StorageError already provides the same definition. Differential Revision: https://phab.mercurial-scm.org/D9234

commit: leverage cmdutil.check_incompatible_arguments() Differential Revision: https://phab.mercurial-scm.org/D9218

split: use default one-line summary Differential Revision: https://phab.mercurial-scm.org/D9255

phabsend: use default one-line summary This means that the `phabricator.node` and `phabricator.desc` color configs will no longer be respected, but `log.changeset` and `log.desc` will be respected instead. Differential Revision: https://phab.mercurial-scm.org/D9254

rebase: change and standarize template for rebase's one-line summary This removes the default template in rebase and switches to a centrally defined template. I've simplified it a bit to avoid the conditional parenthesis. I've also added labels so the different parts can be easily colored. The template is somewhat similar to what we've used internally at Google for a few years. I'm happy to change the template if others have opinions. Should we reuse the `color.log.` names as I have? Differential Revision: https://phab.mercurial-scm.org/D9252

rebase: make summary template configurable, with default to shared template Differential Revision: https://phab.mercurial-scm.org/D9251

rebase: use hard-coded template for one-line commit description This is to prepare for making making the one-line summary customizable. The template ended up pretty complicated because of the conditional output of "(<bookmarks etc>)". Maybe we can simplify the template later. Differential Revision: https://phab.mercurial-scm.org/D9250

config: move ui.pre-merge-tool-output-template into [command-templates] Differential Revision: https://phab.mercurial-scm.org/D9249

config: rename ui.mergemarkertemplate to command-templates.mergemarker Differential Revision: https://phab.mercurial-scm.org/D9247

config: rename ui.graphnodetemplate to command-templates.graphnode Differential Revision: https://phab.mercurial-scm.org/D9246