graphmod: restore generator nature of dagwalker 9966c95b8c4f introduced the ability to walk the DAG given arbitrary revisions, but changed the behaviour of it to return a list of all nodes (and create a changectx for each one) rather than doing it lazily. This has a pretty significant impact on performance for large repositories (tested on CPython repo, with output disabled): $ time hg glog real 0m2.642s user 0m2.560s sys 0m0.080s Before 9966c95b8c4f: $ time hg glog real 0m0.143s user 0m0.112s sys 0m0.032s And after this fix: $ time hg glog real 0m0.213s user 0m0.184s sys 0m0.028s

graphlog: log -G --follow file does not work, forbid it We do not have revsets to follow file history.

graphlog: unify log -G revset translation

graphlog: fix log -G --prune follow() revset really means '::.' while we want something based on the passed argument. Also, ancestors() revset does not include the parent revisions.

graphlog: escape log -G single value option arguments

graphlog: fix log -G --removed Current version was mispelled and processed like a valued argument.

graphlog: fix log -G option types when translating to revset For instance, --keyword is a multiple value option while it was processed as a single value option.

revset: add tests for multiple and mixed ^ and ~ operators

extensions: obsolete and remove parentrevspec extension As of 305c97670d7a, revsets implements the ^ and ~ operators that this extension provides, so it's no longer necessary.

wix: a new mercurial.js has replaced graph.js

util.url: copy urllib.unquote() into util to improve startup times The ui class uses util.hasscheme() in a couple of places, causing hg to import urllib even when it doesn't need to. This copies urllib.unquote() to avoid that import. perfstartup time before the URL refactoring (8796fb6af67e): ! wall 0.050692 comb 0.000000 user 0.000000 sys 0.000000 (best of 100) before this change: ! wall 0.064742 comb 0.000000 user 0.000000 sys 0.000000 (best of 100) after this change: ! wall 0.052126 comb 0.000000 user 0.000000 sys 0.000000 (best of 100

url: move URL parsing functions into util to improve startup time The introduction of the new URL parsing code has created a startup time regression. This is mainly due to the use of url.hasscheme() in the ui class. It ends up importing many libraries that the url module requires. This fix helps marginally, but if we can get rid of the urllib import in the URL parser all together, startup time will go back to normal. perfstartup time before the URL refactoring (8796fb6af67e): ! wall 0.050692 comb 0.000000 user 0.000000 sys 0.000000 (best of 100) current startup time (139fb11210bb): ! wall 0.070685 comb 0.000000 user 0.000000 sys 0.000000 (best of 100) after this change: ! wall 0.064667 comb 0.000000 user 0.000000 sys 0.000000 (best of 100)

revlog: introduce _chunkbase to allow filelog to override Used by revlog.revision to retrieve the base-chunk in a delta chain.

filelog: extract metadata parsing and packing _parsemeta returns the dictionary and a list of keys in the order they appear in metadata. This can be used to repack the dictionary in the same order. _packmeta creates metadata from a dictionary and an optional key-order list. In _parsemeta, we use slices and re.search indead of str.index so we can accept both buffers and strings.

discovery: drop findoutgoing and simplify findcommonincoming's api This is a long desired cleanup and paves the way for new discovery. To specify subsets for bundling changes, all code should use the heads of the desired subset ("heads") and the heads of the common subset ("common") to be excluded from the bundled set. These can be used revlog.findmissing instead of revlog.nodesbetween. This fixes an actual bug exposed by the change in test-bundle-r.t where we try to bundle a changeset while specifying that said changeset is to be assumed already present in the target. This used to still bundle the changeset. It no longer does. This is similar to the bugs fixed by the recent switch to heads/common for incoming/pull.

revset: add missing whitespace

url: remove unused/obsolete functions

revset: add ^ and ~ operators from parentrevspec extension ^ (Nth parent) and ~ (Nth first ancestor) are infix operators that match certain ancestors of the set: set^0 the set set^1 (also available as set^) the first parent of every changeset in set set^2 the second parent of every changeset in set set~0 the set set~1 the first ancestor (i.e. the first parent) of every changeset in set set~2 the second ancestor (i.e. first parent of first parent) of every changeset in set set~N the Nth ancestor (following first parents only) of every changeset in set; set~N is equivalent to set^1^1..., with ^1 repeated N times.

encoding: add an encoding-aware lower function

add: notify when adding a file that would cause a case-folding collision On a case-sensitive file system, files can be added with names that differ only in case (a "case collision"). This would cause an error on case-insensitive filesystems. A warning or error is now given for such collisions, depending on the value of ui.portablefilenames ('warn', 'abort', or 'ignore'): $ touch file File $ hg add --config ui.portablefilenames=abort File abort: possible case-folding collision for File $ hg add File warning: possible case-folding collision for File

scmutil: refactor ui.portablefilenames processing The ui.portablefilenames config handling is generally useful for notifying the user of various portability problems.

pure bdiff: don't use a generator Generators are slow, and currently defeat the PyPy JIT.

pure mpatch: avoid using list.insert(0, ...) In Python lists are implemented as arrays with overallocation. As a result, list.insert(0, ...) is O(n), whereas list.append() has an amortised running time of O(1). Reversing the internal representation of the list should cause a slight speedup for pure Python builds.

remove unused imports and variables

Zeroconf: catch both ValueError and KeyError in get() This prevents a traceback during discovery of available hosts.

run-tests: print a new line before writing the diff

revsets: add a last function last(set, n) = reverser(limit(reverse(set), n))

bundle: more comments about the different header types, remove useless if

merge with mpm

hgweb: initialize permhooks at definition time This is simpler than creating it empty and then appending the default checkauthz hook 50 lines below.

revset: fix undefined name ParseError

branchcache: improve speed relative to the amount of heads Updating the branch cache is quadratic to the amount of heads in the repository. One consequence of this was that cloning a pathological repository with 10,000 heads (and nothing else) took hours of CPU time. This patch makes one of the inner loop much faster, by removing a changectx instantiation, and removes another entirely in cases where there are no candidate branch heads which descend from other branch heads.

drop {short,hex}(ctx.node()) calls in favor of ctx methods

mq: be more explicit on invalid patch name message

fix broken tests test-debugcomplete.t broken by 58e58406ed19 test-highlight.t broken by b24e5a708fad

merge with stable

mq: add '.' and '..' to list of forbidden patch names When an empty string is being passed to normname it would return '.' causing checkfile() to always return that a patch with that name exists.

subrepo: handle svn tracked/unknown directory collisions This happens more often than expected. Say you have an svn subrepository with python code. Python would have generated unknown .pyc files. Now, you rebase this setup on a revision where a directory containing python code does not exist. Subversion is first asked to remove this directory when updating, but will not because it contains untracked items. Then it will have to bring back the directory after the merge but will fail because it now collides with an untracked directory. Using --force is not very elegant and only works with svn >= 1.5 but the only alternative I can think of is to write our own purge command for subversion.

subrepo: backout bcc6ed0f6c3b

wireproto: add test for new optional arg missing on server New argument is silently ignored by both HTTP and SSH servers. This means we can, for instance, add new flags to getbundle() to request advanced features (like lightweight-copy-aware bundles), and older servers will silently ignore this request and send back a plain bundle.

merge with stable

web: Made elapsed time calculation dynamic (javascript). This allow safe caching of the pages by the browser and still display the right amount of elapsed time upon page refresh. If javascript is disabled, absolute time is displayed, leaving it readable. All the templates have been updated.

extdiff: add repository root as a variable Some external diff tools (notably Plan 9 diff(1)) require the absolute path to the file being diffed for proper function. A root variable was added to inform an external tool of the repository root (the tool is invoked with the cwd set to tmproot).

help: move hgignore man page into built-in help (issue2769)

graphlog: make use of graphmod's revset support

graphmod: use revsets internally Thanks for the idea and most of the implementation to Klaus Koch Backs revisions() and filerevs() with DAG walker which can iterate through arbitrary list of revisions instead of strict one by one iteration from start to stop. When a gap occurs in a revisions (i.e. in file log), the next topological parent within the revset is searched and the connection to it is printed in the ascii graph. File graph can draw sometimes more connections than previous version, because graph is produced according to the revset, not according to a file's filelog. In case the graph contains several branches where the left parent is null, the graphs for each are printed sequentially, not in parallel as it was a case earlier (see for example the graph for README in hg-dev).

svn subrepo: attempt work around obstructed checkouts (issue2752) It should be possible to do better than this with 'svn switch', but the logic required woud be significantly more complex. Until someone needs the performance improvements of using switch, we'll just use the same strategy for everything.

churn: strip key earlier to avoid false negative seach in aliases