2011/2/10 Branko Čibej <brane_at_e-reka.si>:
> On 10.02.2011 20:25, Paul Burba wrote:
>>> Author: danielsh
>>> Date: Wed Nov 10 21:02:46 2010
>>> New Revision: 1033709
>>>
>>> URL: http://svn.apache.org/viewvc?rev=1033709&view=rev
>>> Log:
>>> Add a regression test for 'svn patch' setting properties without
>>> trailing newlines. This could be useful, for example, to transmit
>>> 'svn:eol-style' properties via 'svn patch' without appending a newline
>>> to the property value en route.
>> Hi Daniel,
>>
>> This test seems to be checking if 'svn patch' can apply a property
>> value with no trailing newline, but is there even a way to represent
>> such a property in the unidiff format?
>>
>> Your test creates this patch:
>>
>> [[[
>> Index: .
>> ===================================================================
>> diff --git a/subversion/branches/1.6.x b/subversion/branches/1.6.x
>> --- a/subversion/branches/1.6.x (revision 1033278)
>> +++ b/subversion/branches/1.6.x (working copy)
>>
>> Property changes on: subversion/branches/1.6.x
>> ___________________________________________________________________
>> Modified: svn:mergeinfo
>> Merged /subversion/trunk:r964349
>> Added: k
>> ## -0,0 +1 ##
>> +v
>> \ No newline at end of property
>> ]]]
>>
>> But what is '\ No newline at end of property'? Is that standard? I'm
>> only familiar with the special line '\ No newline at end of file'.
>
> The magical token is the \ at the beginning of the line. The rest of the
> line can be random text as far as patch is concerned, and indeed,
> translations will localize that text.
Hi Brane,
Thanks for the info. I was under the impression that '\ No newline at
end of file' was special:
http://en.wikipedia.org/wiki/Diff#Unified_format ... but I hear you
can't believe everything you read on the Internet ;-)
Regardless, my original question for Daniel still holds:
This test seems to be checking if 'svn patch' can apply a property
value with no trailing newline, but is there even a way to represent
such a property in the unidiff format?
If there is not, then I don't see how this test is valid (in the sense
that it is never going to pass).
Paul
Received on 2011-02-14 23:37:21 CET