On Saturday, June 2, 2012 11:53:02 AM UTC+2, Christian Hammond wrote:
>
> Hi Ryan,
>
> Branches don't matter in Git at this point. A branch is just a handy tag 
> that gives a human-readable name to a SHA1 and updates on each commit. From 
> the point of view of anything that needs to interact with files on Git, the 
> only information needed are a SHA1 and a filename. A branch will never come 
> into play here.
>

Yes, you is right. I have not dug so much before into SHA1 numbers and 
git... 

>
> What it sounds like is that your remote repo isn't updated properly. 
> Perhaps you have changes in your master that aren't pushed, and the diff is 
> therefore based on a revision that is not available to Review Board by way 
> of cgit?
>
>  
However, this is the breakdown...

For a particular file, post-review (git diff ), generates a patch for a 
file x. This contains a parent SHA id and the new sha id, right?

index 
ebd63fe54cf46839f6ca5582699b249783aca538..3ce9a396ebbda257d1625e9101ca0f9a27d9ed29

The log for reviewboard shows that there is a fetch 
for ebd63fe54cf46839f6ca5582699b249783aca538 id, but cgit returns "not 
found" for that
file.... I can see in the HTML formatted cgit 
site, ebd63fe54cf46839f6ca5582699b249783aca538  exists for the blob link.

Just for clarity

************************************** Reviewboard.log 
***************************************************
2012-06-02 14:12:10,925 - INFO - Fetching file from 
http://machx/cgit/buildroot/plain/package/libxml2/Config.in?id=ebd63fe54cf46839f6ca5582699b249783aca538
2012-06-02 14:12:10,932 - ERROR - 404

*************************************** post-review 
********************************************************
rra@pc-9999  (branchX)$ post-review -r 36 --debug --parent origin/branchX 
 --output-diff
>>> RBTools 0.4.1
>>> Home = /home/rra
>>> HTTP GETting api/
>>> HTTP GETting http://machx/reviews/api/info/
>>> Using the new web API
diff --git a/package/libxml2/Config.in b/package/libxml2/Config.in
index 
ebd63fe54cf46839f6ca5582699b249783aca538..e66740a4dd4263112d10dd0eba0d5c534437a842
 
100644
--- a/package/libxml2/Config.in
+++ b/package/libxml2/Config.in
@@ -4,3 +4,9 @@ config BR2_PACKAGE_LIBXML2
          XML C Parser

          http://xmlsoft.org/
+
+config BR2_PACKAGE_LIBXML2_MINIMAL
+       bool "libxml2 minimal"
+       depends on BR2_PACKAGE_LIBXML2
+       help
+         Build a minimally sized library.



I am digging further, but is there anything blatantly obvious I am 
overlooking? This really doesn't make sense...


Ryan





Christian
>
> -- 
> Christian Hammond - chip...@chipx86.com
> Review Board - http://www.reviewboard.org
> VMware, Inc. - http://www.vmware.com
>
>
> On Sat, Jun 2, 2012 at 2:21 AM, Ryan Raasch <ryan.raa...@gmail.com> wrote:
>
>>
>>
>> On Friday, June 1, 2012 6:54:08 PM UTC+2, Christian Hammond wrote:
>>>
>>> The branch doesn't come into play. All that's needed for checking out a 
>>> file is a filename and revision.
>>>
>>> Something like cgit is actually required for remote Git repositories, as 
>>> Git has no ability to fetch individual blobs remotely.
>>>
>>> Yea, that makes sense. But when I am using post-review to generate a 
>> diff on a local branch to a remote branch, the diff is correct. However, 
>> when the server tries to apply the patch, it fails. The server returns an 
>> error that one of the files is not found in the remote branch. After 
>> investigating this,
>> it turns out that the master on remote repos does not have that file, and 
>> the remote branch does. So it seems that the patch was applied to the 
>> default branch, master on the server (which is what cgit delivers in the 
>> RAW URL without the branch specification).
>>
>> The diff generated by post-review is correct (diff is created from the 
>> correct remote branch to local branch), but the apply on the server side 
>> seems to not apply the diff on the correct branch. What I cannot understand 
>> is that once the diff is generated on the client side, the branch 
>> information is lost. How does the server know which branch to apply the 
>> received patch (through the cgit interface)? BTW, everything worked when I 
>> used a local clone residing on the server.
>>
>> Ryan
>>   
>>
>>> Christian
>>>
>>>
>>>
>>> On Jun 1, 2012, at 2:33, Ryan Raasch <ryan.raa...@gmail.com> wrote:
>>>
>>> Hello,
>>>
>>> I have now decided it easier to use the cgit interface for the backend. 
>>> However, the
>>> raw url format does not include the parameter for the branch to be 
>>> diffed (in git)
>>>
>>> http://cgit.server/repos/**plain/<filename>?h=<branch>&**id=<revision><http://cgit.server/repos/plain/%3Cfilename%3E?h=%3Cbranch%3E&id=%3Crevision%3E>
>>>
>>> I cannot find where the branch can be specified. Is this possible?
>>>
>>> I guess also (now that i think about it), uploading a diff does not ask 
>>> which branch the
>>> diff should be based upon...
>>>
>>> BTW i am using v1.6.6
>>>
>>> Cheers,
>>> Ryan
>>>
>>> -- 
>>> Want to help the Review Board project? Donate today at 
>>> http://www.reviewboard.org/**donate/<http://www.reviewboard.org/donate/>
>>> Happy user? Let us know at 
>>> http://www.reviewboard.org/**users/<http://www.reviewboard.org/users/>
>>> -~----------~----~----~----~--**----~----~------~--~---
>>> To unsubscribe from this group, send email to reviewboard+unsubscribe@**
>>> googlegroups.com <reviewboard+unsubscr...@googlegroups.com>
>>> For more options, visit this group at http://groups.google.com/**
>>> group/reviewboard?hl=en<http://groups.google.com/group/reviewboard?hl=en>
>>>
>>>  -- 
>> Want to help the Review Board project? Donate today at 
>> http://www.reviewboard.org/donate/
>> Happy user? Let us know at http://www.reviewboard.org/users/
>> -~----------~----~----~----~------~----~------~--~---
>> To unsubscribe from this group, send email to 
>> reviewboard+unsubscr...@googlegroups.com
>> For more options, visit this group at 
>> http://groups.google.com/group/reviewboard?hl=en
>>
>
>

-- 
Want to help the Review Board project? Donate today at 
http://www.reviewboard.org/donate/
Happy user? Let us know at http://www.reviewboard.org/users/
-~----------~----~----~----~------~----~------~--~---
To unsubscribe from this group, send email to 
reviewboard+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/reviewboard?hl=en

Reply via email to