Martin von Zweigbergk <martinvonz@google.com> [Wed, 15 Jul 2020 22:24:02 -0700] rev 45132
extensions: make `hg help extensions` list disabled extensions with PyOxidizer
`setup.py` has a mechanism for generating a list of bundled extensions
and their documenation and putting it in a `hgext/__index__.py`
file. That file is then read from `extensions.py`, which makes `hg
help extensions` work in py2exe binaries. This patch makes it so
`setup.py` produces the `hgext/__index__.py` file also when called by
PyOxidizer.
Thanks to Matt Harbison for pointing out the existence of
`__index__.py`!
Differential Revision: https://phab.mercurial-scm.org/D8758
Joerg Sonnenberger <joerg@bec.de> [Wed, 08 Jul 2020 00:36:36 +0200] rev 45131
phases: sparsify phaseroots and phasesets
As final step of dealing with the holes in the phase numbers, make
phaseroots and phasesets both dictionaries indexed by the phase number.
Further adjust the interface of the C module by pushing the node to
revision mapping down as it is cheaper on the C side to deal with
revision numbers.
Overall, the patch series improves a no-change "hg up" for my NetBSD test
repository from 4.7s to 1.3s.
Differential Revision: https://phab.mercurial-scm.org/D8698
Pulkit Goyal <7895pulkit@gmail.com> [Tue, 07 Jul 2020 14:01:12 +0530] rev 45130
extdiff: add some comments in diffrevs()
It was not obvious to understand the code so I added some comments.
Differential Revision: https://phab.mercurial-scm.org/D8690
Pulkit Goyal <7895pulkit@gmail.com> [Tue, 07 Jul 2020 13:26:57 +0530] rev 45129
extdiff: add comments and minor variable renames diffpatch()
Some variable names were now confusing as we refactored the code in a separate
function. For example, `node1a` leads to ideas why `1a` and not `1`. The
variable storing path to patch file was named as `dirX` instead of `fileX`.
Renamed these variables and added couple of comments.
Differential Revision: https://phab.mercurial-scm.org/D8689
Pulkit Goyal <7895pulkit@gmail.com> [Tue, 07 Jul 2020 13:24:13 +0530] rev 45128
extdiff: remove unrequired do3way argument to diffpatch()
When we are diffing patches, there is no 3 way diff.
Differential Revision: https://phab.mercurial-scm.org/D8688
Pulkit Goyal <7895pulkit@gmail.com> [Tue, 07 Jul 2020 13:13:18 +0530] rev 45127
extdiff: refactor logic to diff revs of versions of files
Now that code for both cases, diffing patches or files is in separate function,
it will be better to refactor them more and understand.
Differential Revision: https://phab.mercurial-scm.org/D8687
Pulkit Goyal <7895pulkit@gmail.com> [Tue, 07 Jul 2020 12:42:39 +0530] rev 45126
extdiff: refactor logic which does diff of patches
The current extdiff code is hard to understand on first look. Took me few hours
to grasp the code. Before adding more things, decided to do some refactoring.
Differential Revision: https://phab.mercurial-scm.org/D8686
Manuel Jacob <me@manueljacob.de> [Mon, 01 Jun 2020 09:55:31 +0200] rev 45125
absorb: make it explicit if empty changeset was created
If the config rewrite.empty-successor=skip is set, a message "became empty and
was dropped" is shown if the changeset became empty. If the config
rewrite.empty-successor=keep is set, absorb may create changesets even if they
became empty. It’s probably a good idea to make that explicit. Therefore the
message is changed to be a combination of both: "became empty and became ...".
Repeating the word "became" is not very elegant. This results from the fact
that "became" was and is overloaded to indicate both the change from non-empty
to empty and the successor relation. In the combinated message, both meanings
are used in one sentence.
Manuel Jacob <me@manueljacob.de> [Mon, 01 Jun 2020 09:42:41 +0200] rev 45124
absorb: consider rewrite.empty-successor configuration
This adds support for the recently added rewrite.empty-successor configuration.
Manuel Jacob <me@manueljacob.de> [Mon, 01 Jun 2020 08:38:42 +0200] rev 45123
rebase: consider rewrite.empty-successor configuration
This adds support for the recently added rewrite.empty-successor configuration.