1118: Changes between revision 1 and 2 fails when the change list contains no files in common.

Jay****@gmai***** (Google Code) (Is this you? Claim this profile.)
July 26, 2009 
What version are you running?
1.0 RC2 on Linux Redhat 4.6 using python 2.6 with perforce.

What's the URL of the page containing the problem?


What steps will reproduce the problem?
1. Use post-review to upload 2 files(a,b) using changelist 1, publish
2. Use post-review to upload 3 different files(c,d,e) using change list 2 
to the same review, publish
3. open the review, go to view diff.
4. Click on 'changes between r2 and 1.
You will see 
1. c 2 changes
2. d 3 changes
3. e 2 changes All the differences were from the 2nd diff for all 3 files.
4. e rotating cursor
5. f rotating cursor

What is the expected output? What do you see instead?

I would expect it to show me no differences.
The browser does not hang, but this is not correct.



What operating system are you using? What browser?
Running on Linux Redhat 4.6 using either firefox 3 or IE 6.
The SCM is perforce.
post-review was used to upload the files.


Please provide any additional information below.
chipx86
#1 chipx86 
It should show you differences. The second diff doesn't contain the first two files,
so those files should basically be reverts, and the new files should be as they are
in the diffs. Assuming this is happening, then the behavior is correct.
#2 Jay****@gmai***** (Google Code) (Is this you? Claim this profile.) 
It shows the 3 new files as reverts.  It then shows 2 files again. I made a mistake 
in typing.  It shows as follows;
1. c 2 changes
2. d 3 changes
3. e 2 changes All the differences were from the 2nd diff for all 3 files.
4. e rotating cursor
5. d rotating cursor
That is it shows you 2 of the 3 from the second set 2x and the first set not at all.
That is files a and b are not shown. d and e are show 2 X and the second time the 
cursor is continuously spinning,but they are not show.
chipx86
#3 chipx86 
Would you mind making diffs against the Review Board SVN tree that do this, and post
them up at http://demo.review-board.org/? It would help us to have a good, sane repro
case that we can include in unit tests.
  • +NeedInfo
david
#4 david
  • -NeedInfo
    +Incomplete