Notification mails for git repos

Derek Atkins warlord at MIT.EDU
Thu Jan 31 09:46:57 EST 2013


Yawar Amin <yawar.amin at gmail.com> writes:

> Hi,
>
> On 2013-01-30 22:44, John Ralls wrote:
>> [...]
>> 2.4 was branched off of 2.3, not trunk. And it was done 4 years and 4500 changes ago. If the bit of code that you're working on happens not to have changed in those 4 years (there's plenty that hasn't) it makes no difference whether you base your change on r18000 or r22737. If it *has* changed in either branch, the only thing you accomplish by working with r18000 is that your changeset might not apply to *either* branch.
>
> Agreed, this _could_ happen. But we don't know for sure. There are
> several factors here. How old is the common ancestor? Did the bug
> exist at that point or was it introduced later? Is it a large fix or a
> small one? Were the lines that we plan to modify in r18000 changed
> after that commit?

Often, yes, there is great divergence.  This happens frequently when
major subsystems change.  While you might find a good percentage of code
that's common (at least on a line-by-line basis), in general the code
does diverge over the multi-year lifetime of the stable releases.

> I guess daggy fixes are more of a personal preference than a panacea :-)

Yes.  In my experience patches often require manual porting to work on
both branches (trunk and "release").  Sometimes it applies cleanly to
both branches, but the more time that passes the less likely that would
happen.  So in general I would say that it does require manual
intervention and actual porting.

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available


More information about the gnucash-devel mailing list