RFR: CODETOOLS-7901115: webrev -N -r generates broken patch files

[Bengt Rutisson]

On 05/05/15 22:21, Stuart Marks wrote:
> In case anyone is wondering, the changeset that Bengt is proposing to 
> revert was mine. He and I have discussed this, and reverting it seems 
> to be the best course of action in the short term.

Thanks for clarifying this, Stuart! I should have written more about it 
in my original post.

>
> This script has too many %@#$& global variables, otherwise I think 
> we'd fix it instead of backing it out.

Totally agree!

Bengt

>
> s'marks
>
> On 5/5/15 1:20 AM, Bengt Rutisson wrote:
>>
>> Hi everyone,
>>
>> Can I have a couple of reviews for this change to fix a P1 issue in 
>> the webrev
>> script? I also need a sponsor to push the change.
>>
>> http://cr.openjdk.java.net/~brutisso/CODETOOLS-7901115/webrev.00/
>> https://bugs.openjdk.java.net/browse/CODETOOLS-7901115
>>
>> There are some details for what is happening in the bug report. The 
>> conclusion
>> is that the problem was introduced by this change:
>>
>> http://hg.openjdk.java.net/code-tools/webrev/rev/09eb35524f9a
>>
>> Since this is a feature (a good feature in my opinion) but causes 
>> problems we
>> decided that the easiest way forward is to just back out this change 
>> now to make
>> sure we can trust the generated patch files again. Later we can try 
>> to implement
>> the feature to generate hg exports for "webrev -r" again.
>>
>> So, the changes proposed in this review request basically just undoes 
>> what was
>> done for 09eb35524f9a.
>>
>> Thanks,
>> Bengt