remove a branch from a repo

John Arbash Meinel john at arbash-meinel.com
Wed Sep 27 21:15:36 BST 2006


William Dode wrote:
>

...

>> We've talked about garbage collection a lot. And it is pretty easy to
>> implement. The biggest issue is that destroying history is not something
>> to do lightly. So it gets low priority from us, because it isn't
>> something that you want to mess up, and isn't super critical for proper
>> functioning.
>>
>> I have what needs to be done mapped out in my head, and if someone is
>> interested in implementing it, I can go over it with them. I could write
>> up a spec one day, just hasn't been worth the time yet.
>>
>> So William, how important is this for you?
> 
> Thank you but it's not actualy important for me, i just pointed this 
> because when we read the tutorials about repository it really miss 
> something, specialy whith example of branch to correct one bug. The 
> shared repository are specialy interesting for big tree, so people will 
> think about this.

Well if you are branching to fix a bug, and then merge that into the
source. You still have that revision in the ancestry, so you *don't*
want to get rid of it anyway.

> 
> Maybe for now it'll be enough to just add somes comments on the 
> tutorials to say that 1/ it will be a garbage collection one time, and 
> 2/ even for big tree an unused branch doesn't use a lot of place in the 
> repository.
> 

Any given revision generally only occupies space relative to the size of
the change. A small fraction is a full text so that we don't have to go
too far back into history. But yes, a revision is ~O(changes) not
~O(tree size). (I can give examples where the constant factors dominate,
but this is the general case)

John
=:->


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 254 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060927/eea5972d/attachment.pgp 


More information about the bazaar mailing list