Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Pig >> mail # dev >> Making git the repo of choice for Pig?


Copy link to this message
-
Re: Making git the repo of choice for Pig?
+1 workday

Thanks,
Prashant

On Apr 3, 2012, at 5:11 PM, Jonathan Coveney <[EMAIL PROTECTED]> wrote:

> Do people prefer a workday or a weekend?
>
> 2012/4/3 Russell Jurney <[EMAIL PROTECTED]>
>
>> Ok, so we just need a date.  Twitter on April <blank> we will be moving
>> piggybank to github (and other awesome things).
>>
>> On Tue, Apr 3, 2012 at 3:55 PM, Alan Gates <[EMAIL PROTECTED]> wrote:
>>
>>> Major +1 for moving Piggybank to github.
>>>
>>> Alan.
>>>
>>> On Apr 3, 2012, at 3:36 PM, Russell Jurney wrote:
>>>
>>>> Some more info is available here, for the Cassandra project:
>>>>
>>>> https://issues.apache.org/jira/browse/INFRA-4254
>>>> https://github.com/apache/cassandra
>>>> http://wiki.apache.org/cassandra/HowToContribute
>>>>
>>>> I have a proposal to make... since we're not yet up for this, but do
>>> agree
>>>> that it is the future: what if we move Piggybank to github?
>>>>
>>>> We can do this at the hackathon we're having at Twitter - I forgot the
>>> date?
>>>>
>>>> On Wed, Mar 28, 2012 at 6:16 PM, Scott Carey <[EMAIL PROTECTED]
>>>> wrote:
>>>>
>>>>> Git was built with use case #1 being Linux developers exchanging
>> patches
>>>>> as text via e-mail.  I am not sure what binary patch support Linus
>> plans
>>>>> or what it would help other than being more compact.  One disadvantage
>>> is
>>>>> that it is then hard to review simple changes by simply reading the
>>> patch
>>>>> itself in a text editor.  I have +1'd many near-trivial changes after
>>> only
>>>>> glancing at the text of the patch.
>>>>>
>>>>> SVN is (slowly) catching up to have better patch features.  'svn
>> patch'
>>> is
>>>>> now in 1.7, 1.8 will enhance it, and eventually there is planned
>> support
>>>>> for git patch extensions. In the long run the differences between the
>>>>> tools for day-to-day work will shrink from where it is now.
>> Subversion
>>> is
>>>>> planning 'svn shelve' to make it more git-like when working locally,
>> for
>>>>> example.
>>>>>
>>>>> On 3/27/12 10:43 PM, "Dmitriy Ryaboy" <[EMAIL PROTECTED]> wrote:
>>>>>
>>>>>> Could you enumerate the problems you have with text patches?
>>>>>>
>>>>>> I can think of 1, but would like to hear your list.
>>>>>>
>>>>>> Dmitriy
>>>>>>
>>>>>> On Tue, Mar 27, 2012 at 2:48 PM, Russell Jurney
>>>>>> <[EMAIL PROTECTED]> wrote:
>>>>>>> Using text patches is the problem as I see it.  We should be
>> using...
>>>>>>> something else.  Doesn't git's functionality enable phasing out text
>>>>>>> patches?
>>>>>>>
>>>>>>> On Tue, Mar 27, 2012 at 11:16 AM, Scott Carey
>>>>>>> <[EMAIL PROTECTED]>wrote:
>>>>>>>
>>>>>>>> On the Avro project, several folk generate a patch using git and
>>> upload
>>>>>>>> that to the JIRA, it patches fine using unix 'patch'.
>>>>>>>> Why do you need to switch to svn to generate the patch?
>>>>>>>>
>>>>>>>> A committer needs to apply the patch to a location checked out with
>>>>>>>> svn,
>>>>>>>> but that is trivial.
>>>>>>>>
>>>>>>>> On 3/27/12 2:26 AM, "Gianmarco De Francisci Morales" <
>>> [EMAIL PROTECTED]>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> My 2 cents, based on the S4 project experience so far (they have
>>> git)
>>>>>>>> is
>>>>>>>>> +1
>>>>>>>>> but later.
>>>>>>>>>
>>>>>>>>> The process with git is still a bit fuzzy and not fully integrated
>>> and
>>>>>>>>> standardized like the one with svn.
>>>>>>>>> I personally use git for all my dev work with Pig, but then switch
>>> to
>>>>>>>> svn
>>>>>>>>> to generate the final patch. A bit cumbersome but not so bad.
>>>>>>>>>
>>>>>>>>> Switching now will generate confusion and is not that urgent in my
>>>>>>>> humble
>>>>>>>>> opinion.
>>>>>>>>>
>>>>>>>>> Cheers,
>>>>>>>>> --
>>>>>>>>> Gianmarco
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Sat, Mar 24, 2012 at 01:27, Dmitriy Ryaboy <[EMAIL PROTECTED]
>>>
>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> There is a check box you check when you upload a patch. If we
>>>>>>>> committed
>>>>>>>>>> without verifying you checked it, thats an unfortunate oversight.