Opened 14 years ago
Closed 14 years ago
#611 closed defect (fixed)
Description for "hg:pull" is too vague
Reported by: | krigstask | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | 0.6 |
Component: | Documentation | Version: | 0.6b2 |
Keywords: | Cc: | ||
Operating System: |
Description
hg:pull tells us:
"dir" parameter: local subdirectory with repository (optional, defaults to '.')
It is not a "subdirectory with repo". It is a "where-to-pull-from", be it a local directory, remote path (http:// or something) or an entry in hgrc/[paths].
Therefore I could hardly agree with dir="." as a sensible default: local repo won't be updated from remote one and build would fail with "unknown changeset" from Mercurial, because new changeset (reported to slave by master) will be absent in local repo, which is at least weird. I'd suggested dir="default", which results in usual "pull then update to ${revision}".
Attachments (0)
Change History (7)
comment:1 Changed 14 years ago by osimons
- Milestone set to 0.6.1
comment:2 Changed 14 years ago by krigstask
- Version set to 0.6b2
Ah, sorry. What I wrote is true for 0.6b2 release, now it is fixed in source:trunk/biiten/build/hgtools.py (_dir is now handled differently). So maybe a little addition to documentation? Most people use at least beta, not the live svn version.
comment:3 Changed 14 years ago by osimons
- Milestone changed from 0.6.1 to 0.6
Like this perhaps:
-
doc/commands.txt
a b 1129 1129 1130 1130 As the command depends on a pre-existing repository, bitten-slave must be 1131 1131 started with ``--build-dir=`` option for locating and working with the 1132 repository. 1132 repository. Also, seeing no remote location can be specified the repository 1133 ``.hg/hgrc`` configuration file needs a ``[paths] default = ....`` location 1134 defined. 1133 1135 1134 1136 Parameters 1135 1137 ----------
I'll commit it unless someone wants to phrase it better.
comment:4 follow-up: ↓ 5 Changed 14 years ago by krigstask
I mean, add something like
In latest release, 0.6b2, hg:pull behaves a little bit differently: it passes dir parameter directly to hg command, which then interprets it as a path (or alias) to where to pull from.
If you're planning to release beta3 or something in a short time, there's no need to bother, of course, but it really took me more than a day to investigate. Right now latest release behaves not like docs tell it should.
comment:5 in reply to: ↑ 4 Changed 14 years ago by osimons
Replying to krigstask:
I mean, add something like...
Sorry, can't really do that as the docs are served straight for repository and follow latest source. It won't make sense to use current docs (for upcoming 0.6b3) to document flaws in earlier betas.
comment:6 Changed 14 years ago by krigstask
I see... Sorry for report then, close it somehow, and let's hope nobody would fall into that and beta3 would become ready soon.
Thank you for responding.
comment:7 Changed 14 years ago by osimons
- Resolution set to fixed
- Status changed from new to closed
Committed in [888:889] for active development lines. A continuation of the "Improvements to the hg:pull command" already mentioned in milestone:0.6 docs for upcoming release.
No, I don't think so. The way the command is written to use the regular working-dir input for shell execute, all it does is change current working directory to the local directory before running hg pull.
Seeing that the command does not take any arguments to locate remote repository, it is presumed that a [paths] default = .... setting is configured in the .hg/hgrc config file for the repository so that there is a default location for pulling.
That information could no doubt be added to the docs.