Bisect Inconsistent State

[Tony Mechelynck]

On 16/06/10 23:30, venizio krups wrote:
>> Shouldn't I get an error message on the update in that case?  Something like
>> this:
>>
>> $ hg up -r THUNDERBIRD_3_1_rc2_RELEASE
>> abort: crosses branches (use 'hg merge' to merge or use 'hg update -C' to
>> discard changes)
>
> if you had uncommitted changes you would have seen the message.
>
>> Or maybe I have branches and heads confused.
>
> probl. not, but these are helpful:
> hg help heads
> http://mercurial.selenic.com/wiki/Branch
>
>>   Basically I just want to find
>> an older revision without my bug,
>
> what about hg log --follow? or try a graphical tool like 'hgtk log',
> or 'hg log --graph'
>
>> and since updating to 34299:671931955b3b
>> "seemed" to work, I was unexpectedly surprised when bisect told me it had
>> marked that revision both good and bad.
>
> yeah, it should say something like 'tip is unreachable from here' or
> some such thing.
>
> -vk
> _______________________________________________
> Mercurial mailing list
> Mercurial at selenic.com
> http://selenic.com/mailman/listinfo/mercurial
>

@vk:

Yeah, in this case I would suggest "hg log -f"or "hg log --follow" 
(which mean the same) with an appropriate -l <number> (or --limit 
<number>) argument if you know an upper bound to the number of 
changesets "related in direct line to the tip" that you'll need.

Beware that -f / --follow is incompatible with log -G / --graph (or 
glog). You can use graphical logging, or follow the chain from children 
to ancestors, but not both.


Best regards,
Tony.
-- 
How can you be in two places at once when you're not anywhere at all?