P Purkayastha <
ppu...@gmail.com> writes:
> On 10/22/2012 12:41 AM, Andrey Novoseltsev wrote:
>> Thank you, that worked!
>>
>> Is this a usual workflow with git repositories? It seems that I had to
>> download 7M+ to test just a few lines, not like that was a big deal on
>> a 20Mbit connection, but if it was the same with 300M+ Sage, reviewing
>> patches would get annoying on any connection...
>>
>> Andrey
>
> I think you downloaded 7M probably because you didn't have all the
> files. AFAIK, this shouldn't happen every time. Maybe kini knows
> better.
The 7M is because you did `git fetch jhpalmieri`, which not only fetches
the particular branch you wanted, but all other things that jhpalmieri
had on his branch which your local repository didn't already "know
about". Possibly there were some large things in that category. Note
that this doesn't have to be stuff that only jhpalmieri had - it could
also be new stuff in the main sagemath master branch which you hadn't
fetched but jhpalmieri had, and which jhpalmieri had based one or more
of his branches on.
If you want to avoid this and only get the commits from a particular
branch, you should do `git fetch jhpalmieri A:B` where A is the name of
jhpalmieri's branch on his repository, and B is the name of your branch
which you want to be updated to match jhpalmieri's branch (B doesn't
need to exist yet).
Personally I'd just do `git fetch jhpalmieri` and not worry too much
about network traffic - it should only be proportional to the amount of
development that jhpalmieri is doing, not to the size of Sage. Plus, it
gets you the handy jhpalmieri/* "remote tracking branches", which `git
fetch jhpalmieri A:B` doesn't, I think. And really, 7MB is normally
pretty huge IME. I guess either your master was way behind the upstream
master, or you missed some large commit like a javascript library update
or something...
-Keshav
----
Join us in #sagemath on
irc.freenode.net !