Ignore use case: explicit ignore necessary?

[Jo Vermeulen]

Op 18-apr-05 om 16:12 heeft Martin Pool het volgende geschreven:

> On Mon, 2005-04-18 at 15:40 +0200, Jo Vermeulen wrote:
>
>> Mmm indeed it does :-) Sorry! Ignoring is only necessary when you 
>> don't
>> want bzr status to get cluttered up with unknown files.
>>
>> One thing I did notice was that .bzrignore is not "ignored" by 
>> default.
>> The tutorial says:
>>
>> The .bzrignore file should normally be versioned, so that new copies 
>> of
>> the branch see the same patterns:
>>
>> % bzr add .bzrignore
>> % bzr commit -m "Add ignore patterns"
>
> It's pretty common to want to commit it so every contributor sees the
> same rules.

Indeed. I fully agree on the fact that each checkout should use the 
same ignore rules (and that it should be clear to the contributor which 
rules are used).

>> I understand why it couldn't be part of .bzr/ because it would still
>> have to be user-editable ofcourse (and .bzr/ should not be edited). 
>> But
>> it seems like .bzrignore is not really part of the project itself. 
>> When
>> converting a bazaar-ng repository to CVS for example, .bzrignore still
>> turns up in the CVS version. But maybe that's "desired behavior" since
>> converting this repository to bazaar-ng again would restore the
>> patterns :-)
>
> Doing it this way is consistent with cvs and some other systems, and
> seems to make sense to me.  If nothing else, it means that updates and
> mergers of the ignore patterns can be done through the usual channels,
> rather than by a special mechanism as in svn.

Haven't looked at it that way, that's right.

>> Sorry if I'm bringing up old discussions on this matter.
>
> There was some previous discussion but you're welcome to ask.

OK. Thanks for the explanation!

Kind regards,

-- 
Jo Vermeulen <jo at lumumba.luc.ac.be>
http://lumumba.luc.ac.be/jo/