Re: [Monotone-devel] automate show_conflict

[Thomas Keller]

Stephen Leake schrieb:
> As I understand it, the only downside of leaving use_workspace_options
> true is that mtn might accidently read some random file with
> unpredictable results if run outside a workspace. But the name of the
> file is quite specific, and there are format checks on the structure
> of the file, so I think that risk is acceptably small.

Hey, I was recently blamed to do a lookup in a map twice - what will 
people say if we read a file and its contents without using it... :)

> > And thats why I also like to have automate show_conflicts behave like
> > any other workspace automate command (f.e. inventory), which bails out
> > with "workspace required, but not found" before anything else. 
>
> That also seems like a good convention, but since
> use_workspace_options is currently true, I don't think we can be sure
> that all automate commands actually follow this convention.
>
> In addition, automate show_conflicts is _not_ "like inventory"; it
> only needs a workspace if it must search for the heads to compare, and
> --branch is not given. So if the revids are given as arguments, or
> --branch is given, it doesn't care whether a workspace is present.

Ok, not like inventory, correct, more like content_diff (but even this 
is not a 100% match, because it shows the workspace changes in this 
mode). Anyways, for show_conflicts its still just a convenience that it 
has "more magic" when run inside a workspace. The more I think about all 
this magic (which also needs to be tested separately) the more I think 
we should just drop it and make --branch mandatory.

> > And, lastly, I surely want that commands behave equally on stdio and
> > non-stdio usage :)
>
> That is true now (with use_workspace_options true), so I don't see
> what the problem is.

I updated to 25c98a0758c205f4e5e3f3c278588aefec7c342f and the test fails 
in line 66 with

mtn: misuse: branch 'net.venge.monotone.automate_show_conflict' has 0 
heads; must be at least 2 for show_conflicts

Somehow the new error message is not used...

And again, I get two different error messages for stdio and non-stdio use:

$ ../../../mtn au show_conflicts
mtn: misuse: branch 'invalid-add' has 0 heads; must be at least 2 for 
show_conflicts

vs.

$ echo l14:show_conflictse | ../../../mtn au stdio
0:2:l:53:misuse: please specify a branch, with --branch=BRANCH

Thomas.

--
GPG-Key 0x160D1092 | address@hidden | http://thomaskeller.biz
Please note that according to the EU law on data retention, information
on every electronic information exchange might be retained for a period
of six months or longer: http://www.vorratsdatenspeicherung.de/?lang=en

signature.asc
Description: OpenPGP digital signature