Is it possible to detect if a commit creates a new bookmark or branch via hooks in .hgrc?
I've tried to see if I can find out using hg log, but it just shows on what branch/bookmark the commit has been created: http://hgbook.red-bean.com/read/customizing-the-output-of-mercurial.html
There don't seem to be hooks for it: http://hgbook.red-bean.com/read/handling-repository-events-with-hooks.html
It would make sense I suppose that there isn't a hook for it, because it is also not possible to make a commit which is 'just' the creation of the branch indicating branches/bookmarks only exists when added to a specific commit.
I figured I could check hg branches and hg bookmarks before and after each commit and determine which are removed and added, but is there a cleaner way for detecting branch/bookmark adds/removes?
The pushkey and prepushkey hooks can detect the addition, deletion, and moves of bookmarks.
In hgrc:
[hooks]
prepushkey=echo %HG_NAMESPACE% %HG_KEY% %HG_OLD% %HG_NEW%\n >> out.txt
HG_NAMESPACE will contain "bookmark" and HG_KEY will contain the name of the bookmark.
HG_OLD will contain the hash of the commit the bookmark was before the operation. It won't be set if the bookmark is being created.
HG_NEW will contain the hash of the commit the bookmark will be after the operation. It won't be set if the bookmark is being deleted.
Bookmarks
Bookmarks-handling does not require commit
Bookmark can be created|modified for any changeset in history, not only for current
Bookmark(s) can appear as result of data-exchange (pull|push), not local user's actions
Only part of all possible cases can be covered by hook
Branches
Changing branch always reflected in commit
Branch in changeset may differ from parent's branch not only as result of hg branch (see "merge branches" and "mergesets") - and I haven't nice and easy and ready to use solution for this case
Common notes
You can use standard precommit hook (executed before commit and can enable|disable commit) for testing different conditions in commit or or pretxncommit
Mercurial template-engine has keywords for branch and bookmark for using in hg log -T
In pretxncommit hook commit already exist, but not finalized - it means you can manipulate with commit's data using -r tip and tip's parent in any Mercurial command
Dirty skeleton for hook's body
hg log -T "{KEYWORD}\n" -r "tip + tip^" | ....
where KEYWORD may be branch or bookmarks. Result of log is two-strings output, which can be identical of different (not sure for bookmark, TBT!!), which you have to compare (as you want and can)
PS: Idea inspired by EnsureCommitPolicy Wiki and Mercurial pre commit hook topic here
Related
I'm trying to write a small script for my team that automatically updates the feature branche to a case and then creates a branch for review. I've got the commands down manually but I'm having trouble making it a bit more hands off.
For now, I want to use this templated command:
hg log --rev <changeset> --template "{branch}\n"
Which returns the branch name of a changset. Then I would like to remove a portion of the name and prepend a string. For example, a branch would be named case-1234-FeatureDescription and I would want to be creating a branch named review-1234-FeatureDescription
Ideally, I would like to pipe the output of this command to the branch command
hg branch <result-of-previous-command>
Is it possible to do this?
You suggest to create a branch by a name which you just extraced from the logs of the very same repo. That doesn't exactly look like it can succeed as it already exists.
Additionally, it likely is probably not a good idea to create a named branch for each review process as branch names are persistent. You might consider to use bookmarks for that purpose as they can be deleted without trace from the repo after review is completed.
I'd suggest to use - without piping - something like
hg bookmark -r CHANGESET $(hg log --rev CHANGESET -T"{branch}")-review
I cloned a Mercurial repo and did a bunch of local work, and forgot to make a feature branch for said work.
The normal flow is:
Clone
Create a branch
Switch to that branch
Do your work in that branch
Push that branch
Code review
If code review passes, merge branch w/ default (locally)
Push merged changes to default
Close the feature branch
So I need to create a new branch, port all my unstaged/uncommitted code changes (made to default) over to this branch (so that default is now clean and the new branch contains my changes), and then push my feature branch.
I created the new branch via hg branch new_feature. But after pouring over the Merucrial docs, I can't figure out the next step.
So I ask: How do I move (not just copy) all my unstaged/uncommitted changes from default to my new_feature branch)?
You shouldn't have to do anything in particular.
Your uncommitted working folder changes are fluid, and you can set the branch name before you commit, without losing your changes.
If you're on the command line, simply do this:
hg branch feature-X
hg commit -m "Added feature X"
If you're using TortoiseHg simply click the "Branch: default" button just above the commit message input field and select "Open a new named branch" and give it a name, then click OK, then commit as normal.
Setting the branch name to use during commit does not in fact change your working folder, it doesn't do an update, it doesn't do anything, except record in metadata what the branch name is supposed to be.
Also note that this will only allow you to create a new branch to commit to. If you want to continue on an existing branch you first need to update to the head of that branch and this may cause changes to your working folder. You should not need to do this, however, if you want to create a new branch.
I want to use hg graft to copy a changeset from one branch to another.
However, a file that is modified in the changeset has been renamed (using hg rename) in the source branch -- before the changeset that I want to graft.
When I try I get:
scanning for duplicate grafts
grafting revision 3
searching for copies back to rev 2
unmatched files in local:
f1.txt
resolving manifests
overwrite: False, partial: False
ancestor: c06936303423, local: cfeaa3003c57+, remote: e3f2a201d1e2
remote changed f1a.txt which local deleted
use (c)hanged version or leave (d)eleted? c
Neither (c) nor (d) seem like the right option.
The file f1a.txt was renamed from f1.txt in the source branch. f1a.txt never even existed in the target branch.
Is this possible?
This is an excellent question. It looks like it is not possible. I replayed your scenario and also tried transplant and rebase -- with the same effect.
I guess the basic reason why this does not work is because a graft explicitly only picks the changes of one revision, i.e. past changes (including file renames) explicitly are not considered (update: could be a bug or missing feature as well, see #Soulman's answer).
You did not ask for, but here's a work around ..
Get a patch of the changeset you want to graft:
$ hg export --git -r <rev-to-graft> > my.patch
In this patch, replace f1a.txt with f1.txt, so you get:
...
diff --git a/f1.txt b/f1.txt
--- a/f1.txt
+++ b/f1.txt
## -1,1 +1,1 ##
...
Sed is your friend here: sed -i my.patch -e "s,\([ab]\)/f1a\.txt,\1/f1.txt,g"
Import the edited patch to the target revision:
$ hg up <target-rev>
$ hg import my.patch
I think this is simply a bug that will hopefully be fixed. There is a ticket: Graft fails when files have been moved in an ancestor revision.
A rename is a delete + add + a note saying to treat history of the old file as the history of the new file. When doing a graft (that is making a new changeset by duplicating the same changes without creating a relation as opposed to a merge), you have to incorporate the changes of the changeset some how. You can only graft an "add" operation in this case or decide not to add the file (leave it deleted). I am sorry to say, that it is fundamentally impossible to graft a rename in this case, because it is meaningless if the other branch did not have that file.
I am sorry, if I didn't understand correctly the situation you are in. If that is the case, then please provide an example. (Few commands to set up a dummy repo, which reproduces your case will do the trick).
I want to write a hook that performs some actions each time I run hg branch branch_name (e.g. set "In progress" status for a JIRA ticket), but I can't find anything that runs during branching. Is there a way I can do it?
The is a pre-<command> hook (with a hyphen) for each command. Note that is is distinct from any hook that may exist without a hyphen, sush as precommit.
Thus you can do:
[hooks]
pre-bookmark = /usr/bin/notify_jira.sh ${HG_ARGS#bookmark }
to invoke:
/usr/bin/notify_jira.sh PROJ-415
when you run:
hg bookmark PROJ-415
Full details on the generic pre-<command> (and post-<command>) hooks can be found on the hgrc man page.
It also looks like pushkey hook might do what you want, but pre-bookmark (or better, post-bookmark) is probably more straightforward.
Before, when I was using perforce, I could work on multiple bugs at once as long as the code did not affect the same files, by having multiple change sets open at once.
Changeset 1:
A.txt
B.txt
C.txt
Changeset 2:
D.txt
E.txt
F.txt
I could submit changeset 2 to the repository without submitting changeset 1 (because it's still in progress)
Is this possible with Mercurial? other than doing each file as a separate commit?
You can always just do: hg commit D.txt E.txt F.txt to commit just those files which will leave A.txt, B.txt, and C.txt uncommited. Using the -I option to commit lets you do those with patterns if they're, for example, in a common directory: hg commit -I 'dir1/**'
You can have two separate branches (working copies) and make one change in and the other in the other. That's one way.
Another is to use Mercurial Queues. You can use qpush --move to change the order of changesets if they have no dependencies on one another, so you can then use qfinish to 'commit' the first changeset that's ready.
You don't actually hold changesets "open" in Mercurial.
Either you've committed your changes, or you haven't.
You can, however, have multiple files with uncommitted changes, and only commit a few of them. It is even possible to commit parts of files, chunks, instead of the whole file.
If I was to do what you're asking I would simply make another clone locally from my first one, and work on the two different fixes in two different working folders. Then I have all the freedom I need to combine the two (push/pull locally), push to the server, etc.
In Mercurial, cloning is a cheap operation when done locally. Just clone the repository for each thing you are working on. Push back to the main repository as each thing is ready:
hg clone http://project project
hg clone project project-bugfix // uses hardlinks and is inexpensive.
hg clone project project-feature
<edit bugfix and feature as needed>
But remember that the default push location is the project cloned from, so you'll either need to edit .hg\hgrc's default push path for the "project" clones, or make sure to hg push http://project from "project-bugfix" and "project-feature" when complete.
Even if your modifications touches the same file you can select what to include (hunk by hunk). You just have to activate the record extension (it's installed by not activated by default).
To activate the record extension, put this in the [extensions] section of your .hgrc:
[extensions]
hgext.record=
Then to commit, replace hg commit by hg record. Let's assume your example change sets with a G.txt file that have changes for both change sets. Commit with:
hg record D.txt E.txt F.txt G.txt
Answer questions, for example:
2 hunks, 6 lines changed
examine changes to 'D.txt'? [Ynsfdaq?] f # f for full file (help with ?)
... skipped file E and F
6 hunks, 35 lines changed
examine changes to 'G.txt'? [Ynsfdaq?] y
## ... (patch hunk here)
record change 6/16 to 'G.txt'? [Ynsfdaq?] y # y to include the hunk, n to skip