hgweb: convert comparison {lines} to a mappinggenerator with named template Bare generator can't be restarted.

hgweb: move getblock() closure out of compare() (Please use 'hg diff -w' to get readable diff from this patch.)

hgweb: move compline() closure out of compare()

hgweb: convert {diff} to a mappinggenerator with named template No more bare generator. Fortunately, this one is associated with a single template, so it can be a mappinggenerator.

hgweb: use template context to render {lines} of {diff} This is a preferred way to process nested templates.

hgweb: wrap {lines} of {diff} with mappedgenerator This can't be a mappinggenerator as it switches the templates conditionally.

hgweb: move prettyprintlines() closure out of diffs() This will be wrapped with mappedgenerator.

remotenames: enable the storage config option if extension is enabled Before this patch, the config option to store remotenames was set to False by default and remotenames extension does not set it True. So if you enable remotenames extension without setting 'experimental.remotenames=True', you won't get the remotenames. This patch makes remotenames enable the config option if then extension is enabled which makes sense. Differential Revision: https://phab.mercurial-scm.org/D3520

tests: clone the repo over ssh in tests/test-logexchange.t Cloning a local repo in tests, and normally too does not perform the actual clone, rather it copies the file contents. This makes remotenames information not being transferred. Let's clone over ssh so that we can test remotenames correctly. Differential Revision: https://phab.mercurial-scm.org/D3519

remotenames: check the remotepath with url containing user information too Current logic to match a remotepatch to a user defined path first removes the authentication information from the url and then tries to match it. However this is not true every time. If we clone a repo using ssh, the default path contains the user information too. Differential Revision: https://phab.mercurial-scm.org/D3518

bdiff: fix yet more fallout from xdiff long/int64 conversion (issue5885) "l" in Py_BuildValue's format string means long, so passing int64_t instead results in fireworks on 32bit architectures. Differential Revision: https://phab.mercurial-scm.org/D3538

shortest: avoid magic number "41" As suggested by Yuya. Feel free to fold into D3502. Differential Revision: https://phab.mercurial-scm.org/D3531

revlog: handle error from node lookup I missed this in D3499. Thanks to Yuya for spotting it. Differential Revision: https://phab.mercurial-scm.org/D3530

debugextensions: process extensions loaded from the local repository too

phabricator: drop support for the deprecated `phabricator.token` config

help: load module doc of disabled extension in extensions.disabledcmd() This helps rewriting _finddisablecmd() to not load the module.

extensions: extract closure that looks for commands from disabled module I'll rewrite this function to not load disabled extensions.

extensions: drop dead code trying to exclude deprecated disabled commands Here 'ext' is a tuple, which means ext.__doc__ describes about the Python tuple type. Fixing this would break 'hg help glog', which currently shows some nice explanation, so this patch just removes the dead code. $ hg help glog 'glog' is provided by the following extension: graphlog command to view revision graphs from a shell (DEPRECATED) (use 'hg help extensions' for information on enabling extensions)

util: drop remainder of dateutil/procutil aliases (API)

shortest: move revnum-disambiguation out of revlog I want to be able to change how we disambiguate and I think having revlog.shortest() worry only about finding a prefix that's unambiguous among nodeids makes more sense. This slows down `hg log -T '{shortest(node,1)}\n'` from 4.0s to 4.1s. Differential Revision: https://phab.mercurial-scm.org/D3502

shortest: remove unnecessary check for revnum in isvalid() We now always disambiguating against revnums after we've found an otherwise valid prefix, so we can safely remove the early isrev() check in isvalid(). Differential Revision: https://phab.mercurial-scm.org/D3501

shortest: make pure code also disambigute against revnums at end This makes the pure code more similar to the native code in that it first finds a prefix that's unambiguous among nodeids and then adds hex digits until it no longer looks like a revnum. It will allow us to even better separate the disambiguation with revnums in a later patch. With this patch `hg log -r 0::50 -T '{shortest(node,1)}'` with no native code goes from 25s to 43s. It wasn't exactly usable to begin with, so I don't feel too bad about it. Differential Revision: https://phab.mercurial-scm.org/D3500

revlog: use node tree (native code) for shortest() calculation I want to rewrite revlog.shortest() to disambiguate only among hex nodeids and then disambiguate the result with revnums at a higher level (in scmutil). However, that would slow down `hg log -T '{shortest(node,1)}\n'` from 5.0s to 6.8s, which I wasn't sure would be acceptable. So this patch makes revlog.shortest() use the node tree for finding the length of the shortest prefix that's unambiguous among nodeids. Once that has been found, it makes it longer until it is also not ambiguous with a revnum. This speeds up `hg log -T '{shortest(node,1)}\n'` from 5.0s to 4.0s. Differential Revision: https://phab.mercurial-scm.org/D3499

shelve: reduce scope of merge tool config override The config override seems to have a much greater scope than it needed to. I *think* it's only relevant in the while merging files. The rebase step also cares about the merge tool, but we seem to be explicitly passing it to rebase (around line 755). Differential Revision: https://phab.mercurial-scm.org/D3517

util: drop _deprecatedfunc() It was only needed for the previously removed forwarding.

util: drop deprecated forwarding to procutil (API)

tests: migrate demandimport.py away from deprecated `util` module symbols

templatekw: drop deprecated showlist() and showdict() (API)

util: drop deprecated forwarding to dateutil (API)

util: drop deprecated forwarding to stringutil (API)

scmutil: remove deprecated revpairnodes method (API)

merge: drop support for using updateresults as tuples (API)

cmdutil: drop deprecated log helper methods (API)

cmdutil: drop deprecated precursor of registrar.command (API)

cmdutil: remove deprecated _revertprefetch code (API)

tests: fix deprecation warning in test-url.py

hgweb: remove unused argument 'tmpl' from listfilediffs()

hgweb: use template context to render {files} of changelist entries This is a preferred way to process nested templates.

hgweb: wrap {files} of changelist entries with mappedgenerator This also switches the associated templates conditionally, which can't be a mappinggenerator.

hgweb: use template context to render {files} of changesetentry() This is a preferred way to process nested templates.

hgweb: wrap {files} of changesetentry() with mappedgenerator This can't be easily adapted to mappinggenerator because the associated templates are switched per item.

hgweb: extract generator of {files} from changesetentry() This will be wrapped with mappedgenerator.

templatefilters: document the json filter I don't know the exact reason why it was undocumented, but that would be because unicode is difficult for humans. Since the json filter can now get localstr back to UTF-8 as long as it is unprocessed (i.e. {desc|json} works, but {desc|firstline|json} doesn't), it seems good enough to advertise the json filter.

encoding: introduce tagging type for non-lossy non-ASCII string This fixes the weird behavior of toutf8b(), which would convert a local string back to UTF-8 *only if* it was lossy in the system encoding. Before b7b26e54e37a "encoding: avoid localstr when a string can be encoded losslessly (issue2763)", all local strings were wrapped by the localstr class. I think this would justify the round-trip behavior of toutf8b(). ASCII strings are special-cased, so the cost of wrapping with safelocalstr is negligible. (with mercurial repo) $ export HGRCPATH=/dev/null HGPLAIN= HGENCODING=latin-1 $ hg log --time --config experimental.evolution=all > /dev/null (original) time: real 11.340 secs (user 11.290+0.000 sys 0.060+0.000) time: real 11.390 secs (user 11.300+0.000 sys 0.080+0.000) time: real 11.430 secs (user 11.360+0.000 sys 0.070+0.000) (this patch) time: real 11.200 secs (user 11.100+0.000 sys 0.100+0.000) time: real 11.370 secs (user 11.300+0.000 sys 0.070+0.000) time: real 11.190 secs (user 11.130+0.000 sys 0.060+0.000)

encoding: fix toutf8b() to resurrect lossy characters even if "\xed" in it If 's' is a localstr, 's._utf8' must be returned to get the original UTF-8 sequence back. Because of this, it was totally wrong to test if '"\xed" not in s', which should be either '"\xed" not in s._utf8' or just omitted. This patch moves the localstr handling to top as the validity of 's._utf8' should be pre-checked by encoding.tolocal().

sshserver: redirect stdin/stdout early and use duplicated streams This is what we achieved with hook.redirect(True) plus ui.fout = ui.ferr. The hook.redirect() function can't be completely removed yet since hgweb still depends on it. I'm not sure if it is necessary for WSGI servers. CGI needs it, but does WSGI communicate over stdin/stdout channels?

sshserver: do setbinary() by caller (API) In most cases, stdio should be set to binary mode by the dispatcher, so the sshserver does not have to take care of that. The only exception was hg-ssh, which is fixed by this patch. .. api:: ``sshserver()`` no longer sets stdin and stdout to binary mode.

test-ssh: add some flush() to make output deterministic We shouldn't rely on buffering mode/state of file handles.