Steve Cohen wrote on Wed, Dec 01, 2010 at 17:40:12 -0600:
> On 12/01/2010 03:33 PM, Ryan Schmidt wrote:
>> On Dec 1, 2010, at 15:19, Steve Cohen wrote:
>>
>>> It seems to me that
>>> svn --recursive propset svn:ignore xyz
>>>
>>> is basically just syntactic sugar for manually going through issuing
>>> svn propset svn:ignore xyz
>>>
>>> on every node of the directory structure, It is syntactic sugar in the
>>> sense that the end product in either case would be the same. There is no
>>> "recursive" data attribute in a project's svn:ignore tree.
>>
>> Yes, that's right.
>>
>>
>>> Which leads me to think that a one-time shell script or python script or
>>> whatever might be written to walk the directory tree, look for all the
>>> executable files and manually add them to that directory's svn:ignore list.
>>
>> That should work.
>>
>> The other approach that initially occurred to me was to clean the directory
>> so there are no unversioned files ("make clean" perhaps), then build the
>> software ("make"), then copy the output of "svn status" into an editor and
>> massage it a bit to turn it into something you can hand to "svn propset
>> svn:ignore --file". Though this would have to be done on a per-directory
>> basis, so if there are many directories involved this may be impractical and
>> a script as you suggest may do better.
>>
>>
>>
>>
> Thanks for confirming, this should work, but one more question.
> I've already got some other files in svn:ignore in each directory. There
> is no command to append a name to the svn:ignore property. So my script
> would have to call
> svn propget svn:ignore
> and then
> svn propset svn:ignore
> appending the list, right?
>
>
svn propedit --editor-cmd foo.sh
where foo.sh gets some filename as argv[1] and appends to it.
