[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

RE: [Subclipse-users] Hiding Unchanged Files Within a Branch

From: Chris Spencer <cspencer_at_finestationery.com>
Date: Thu, 28 Feb 2008 09:51:32 -0600

>> The problem is, when I
>> try to use Team->Merge to merge these changes back into the main trunk, SVN
>> says the two codebases are unrelated, presumably because my branch is
>> missing most of the trunk's files.
>
> You can try the --ignore-ancestry setting on merge.  It sounds like
> you would have to do your merges on a file by file basis though.

Yeah, I tried both ignore-ancestry and force options. If I select the entire revision history of the branch, I get the error "Unable to find repository location for 'svn://<branch path>' in revision <blah>". If I only select the first couple revisions, it merges, by all modified files are shown as completely conflicted, even if the change is just a single line.

>> I know SVN branching/tagging is relatively cheap, but with a gig of code
>> even a cheap branch takes a long time to create and work with.
>
> No.  creating a branch of 1 file and 1 million files takes the same
> amount of time and disk space in the repository.
>
>> files may not be duplicated in SVN, but checking out the branch essentially
>> duplicates all the files on my local file-system. Is there any way to
>> selectively checkout only the files from a branch that differ from the
>> trunk, or merge an incomplete branch?
>
> Normally you would use the Switch command to change your working copy
> to point to your branch.  When you are done you switch back to trunk
> and then merge the changes from the branch to trunk and commit.
> Switch is very intelligent and efficient.
>
> You can switch specific files and folders, but I would typically
> recommend switching the entire project so you do not accidentally
> change something on trunk that you meant to change on the branch.

I wasn't referring to SVN's internal performance. You're absolutely right, creating a branch/tag in the repo is very cheap. I was referring to Eclipse's "switch to branch/tag..." feature. It may be smart, but it still takes forever to process my codebase, especially if I'm working remotely and there's considerable network latency. The difference between the branch including every file, and only the files I care about can mean the difference between a 1-hour and 1-minute wait.

Regards,
Chris

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe_at_subclipse.tigris.org
For additional commands, e-mail: users-help_at_subclipse.tigris.org
Received on 2008-02-28 16:54:58 CET

This is an archived mail posted to the Subclipse Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.