Issue running git-svn-mirror script on code

Geert Janssens janssens-geert at telenet.be
Tue Jan 22 12:25:36 EST 2013


On 22-01-13 16:16, Yawar Amin wrote:
> Hi folks,
>
> On 2013-01-22, at 9:51, John Ralls <jralls at ceridwen.us> wrote:
>
>> On Jan 22, 2013, at 5:30 AM, Derek Atkins <warlord at MIT.EDU> wrote:
>>
>>>> [...]
>>> Maybe..  Here's the bigger issue, if I found issues/bugs in John's
>>> svn->git conversion, what do we do?  (and yes, I found a problem in the
>>> conversion)
>>>
>>> [...]
>>>
>>> Personally, I feel that it's important to have the history "correct",
>>> even if it means resetting and invalidating the existing repos.
>> [...]
>>
>> Yes, this is a good opportunity to fix the history. We can simply delete and regenerate the Github repos. No problem, I had to do it a couple of times when I did the original conversion. Everyone will then have to re-clone and reattach any private branches. A PITA, but a one-time one.
>> [...]
> Let me play the devil's advocate here. The GitHub repos are Good Enough™ now because they work--and they have the benefit of being already published and in use (e.g. by me :-). We can always leave the SVN repo up as a read-only archive if people want to see the absolutely correct version of history. In fact as I understand it we're doing that anyway?
Several of the active developers have already switched to using git, so 
all of them will have to do the reclone dance. I personally don't 
perceive it as a big hassle. I've got about 10 private branches, all 
rebased to head. It will only take a couple of minutes to switch.

Just out of curiosity, what's the size of your private history to convert ?

Regarding svn, it is not *my* intention to keep it up read-only in the 
long term. The plan for now is to keep it alive as long as we are 
supporting 2.4. The motivation is mostly that the Windows build of 2.4 
will have some issues when starting from git. And even there, I plan to 
run additional tests in that area when everything is converted. Perhaps 
I can tweak the Windows build enough to be able to build 2.4 (and tags) 
from git as well. In that case I'd even drop svn before the end of life 
of the 2.4 branch.

But regardless, I'm pretty sure svn will disappear at some point. Just 
look at history: originally GnuCash (xacc?) used cvs for revision 
control. I don't think you can still find the original cvs repository 
online. Because svn had the full history of cvs, there was no point in 
keeping cvs alive once it was verified that svn worked ok. It would 
probably even be confusing to keep the two online. Almost certainly an 
occasional newcomer would send in patches against the wrong history.

All this applies to the current svn to git migration as well.

So at some point we would have to fix git's history anyway because the 
original svn history would disappear.

I'd rather do that now since the number of clones is still limited. When 
git becomes the main repo, I expect many more clones to exist (as 
decentralization is a very fundamental concept in git) and hence we'll 
have to ask many more people to reclone.

Only my opinion of course.

Geert


More information about the gnucash-devel mailing list