- Oct 30, 2020
-
-
Anton Shestakov authored
-
- Oct 28, 2020
-
-
Pierre-Yves David authored
-
- Sep 19, 2020
-
-
Anton Shestakov authored
This patch backs out b723805d7fe1, which only hid the problem. Usually mergestate is cleared (or should be cleared) when creating a new changeset, but sometimes there are situations when there's nothing to commit. When solving content divergence resulted in no changes and no new changesets were created, mergestate was left untouched. That's a problem because sometimes mergestate contains files with conflicts that were already resolved by user, and such mergestate only applies to the current wdir parent. Since evolve has already dealt with the current wdir parent successfully when it reached this code, this is a good place to clean the mergestate.
-
- Oct 25, 2020
-
-
Anton Shestakov authored
-
- Oct 24, 2020
-
-
Anton Shestakov authored
-
Anton Shestakov authored
-
Anton Shestakov authored
-
- May 08, 2020
-
-
Anton Shestakov authored
-
Anton Shestakov authored
-
- Oct 09, 2020
-
-
Martin von Zweigbergk authored
CORE-TEST-OUTPUT-UPDATE: f95b23283760
-
- Oct 02, 2020
-
-
Martin von Zweigbergk authored
This adds `compat.update()`, `compat.clean_update()`, which will delegate to the same functions from `mercurial.merge` when they are available, or fall back to calling `hg.updaterepo()` when they're not.
-
Martin von Zweigbergk authored
The next patch will add a `compat.update()` that delegates to the new `merge.update()` from upstream (different from the old one, which was renamed to `merge._update()`).
-
- Sep 25, 2020
-
-
Martin von Zweigbergk authored
The function was renamed in Mercurial commit 2c86b9587740.
-
- Oct 14, 2020
-
-
Sushil Khanchi authored
-
Sushil Khanchi authored
-
Sushil Khanchi authored
-
- Oct 20, 2020
-
-
Pierre-Yves David authored
-
- Oct 12, 2020
-
-
Sietse Brouwer authored
-
- Oct 07, 2020
-
-
Sietse Brouwer authored
This commit does two things: - Firstly, it tweaks the ngtip revset test to exercise the -G/--graph flag. This successfully triggers the bug when it is present. - Secondly, it changes the `ngtip` revset to return a revset made from integer revs instead of node hash bytes. The test now passes. Details: The TypeError was triggered by running hg log -r 'ngtip("default")' --graph in a repository with more than one changeset. The --graph tag caused the flow of control to call `reachableroots2` with the changeset ID found by the `ngtip` revset. Because the changeset ID was a node hash (bytes) instead of a rev (int), reachableroots2 raised the following error: TypeError: an integer is required (got type bytes)
-
- Oct 14, 2020
-
-
Pierre-Yves David authored
If this option is set to `from-branch`, a user can call `hg merge some-topic` from a bare branch even if `some-topic` is a direct descendant of the current working copy parents. This was previously denied if the changesets was on the same branch, since the result would be an "oedipus merge". Some user have been requesting this, and this type of merge is one of Gitlab standard way of merging a "Merge Request". That new option will unlock issue `heptapod#200` and make this mode available for those who wants it.
-
- Oct 15, 2020
-
-
Pierre-Yves David authored
-
- Oct 11, 2020
-
-
Anton Shestakov authored
Switching "(%d changesets obsoleted)" to using successorsmap because it's the source of items in relationships (and the latter variable now can have less items than before this patch).
-
- Oct 12, 2020
-
-
Sietse Brouwer authored
-
- Sep 26, 2020
-
-
Martin von Zweigbergk authored
The bookmark-deactivation code will be done whether or not we're using in-memory merge in `_relocatecommit()`. It could be put in a new `_preparerelocate()` or something, but it's short enough that it doesn't seem worth it.
-
- Sep 25, 2020
-
-
Martin von Zweigbergk authored
It just feels a little weird to update the working copy while a bookmark is active.
-
Martin von Zweigbergk authored
I'm about to add a version of `_relocatecommit()` that uses in-memory merge. It will be simpler to have `_evolvemerge()` inlined then than to need a separate version of that function too. Also, the name `_evolvemerge()` has often made me think that it's about evolving merge commits, so it's nice to get rid of it for that reason too.
-
Martin von Zweigbergk authored
-
- Sep 24, 2020
-
-
Martin von Zweigbergk authored
-
Martin von Zweigbergk authored
We had very similar code for updating bookmarks to handle the case when the relocation resulted in no changes and for the normal case. This restructures the code a bit so the common part is better reused.
-
Martin von Zweigbergk authored
We haven't relied on `rebase.rebasenode()` since 58370c718183 (evolve: use `merge.graft` instead of `rebasenode`, 2014-12-13), so we don't need to compensate for it.
-
Martin von Zweigbergk authored
"unstable" was renamed to "orphan" quite a while ago, so it's about time we update the function name.
-
- Oct 14, 2020
-
-
Sushil Khanchi authored
-
- Oct 10, 2020
-
-
Sushil Khanchi authored
This patch adds the current status of divergence resolution when one of the divergent changeset is a part of split or fold.
-
- Sep 22, 2020
-
-
Sushil Khanchi authored
-
Sushil Khanchi authored
-
Sushil Khanchi authored
-
- Oct 10, 2020
-
-
Sushil Khanchi authored
-
Sushil Khanchi authored
-
- Oct 09, 2020
-
-
Martin von Zweigbergk authored
CORE-TEST-OUTPUT-UPDATE: f95b23283760
-
- Sep 17, 2020
-
-
Matt Harbison authored
The message was, and still is slightly different from core (which mentions 'remote' heads). But it's probably useful to have slightly different messages so that we can easily determine the source.
-