![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
dmazeLet's say I'm doing $INVOLVED_THING at work, which involves changes in a half-dozen widely used files. I don't want to check them in until I'm actually done for various reasons. So I'm halfway done, and my boss comes over and says "there's this bug you should fix". The fix involves changes to the same files. What do you do?
Right now, my "solution" is to have two separate check-outs, one of which is "the head" and one of which is "my working copy". This scales poorly, though, especially if the fix to "the head" also turns out to be involved. I think the thing I want is a "cheap branch": I type something like vc push -r dzm-temp-branch and my version control system creates a "branch" for things that I've modified (maybe just on the local machine) and gets the head. I make my changes, then run vc pop -r dzm-temp-branch, which attempts to merge my changes in. I could live without version control on the "branch" (especially after I actually commit it to the mainline); I definitely don't want things like automated mail.
...I could do this, almost trivially, with a shell script and CVS, huh. Thought.
Right now, my "solution" is to have two separate check-outs, one of which is "the head" and one of which is "my working copy". This scales poorly, though, especially if the fix to "the head" also turns out to be involved. I think the thing I want is a "cheap branch": I type something like vc push -r dzm-temp-branch and my version control system creates a "branch" for things that I've modified (maybe just on the local machine) and gets the head. I make my changes, then run vc pop -r dzm-temp-branch, which attempts to merge my changes in. I could live without version control on the "branch" (especially after I actually commit it to the mainline); I definitely don't want things like automated mail.
...I could do this, almost trivially, with a shell script and CVS, huh. Thought.
