thank you very much for your explanations. Actually, I am not complaining
about what default behavior should be - either the recursive or not recursive
one would be to me perfectly valid. Unfortunately, svn diff is not able to
recursively follow externals, as it does not accept the option -R
svn diff -R
Subcommand 'diff' doesn't accept option '-R [--recursive]'
so I have not any option to check all dirs at once (sigh)
So, as you have suggested, I am going to file an enhancement request, it may
be someone will take up the challenge ;) I will be also willing to test the
modification once it will be implemented!
Have a nice day,
On Saturday 13 March 2010, Stefan Sperling wrote:
> On Sat, Mar 13, 2010 at 12:19:03PM +0100, iprmaster wrote:
> > Hi Neels,
> > thank you for your answer, but I can confirm that other commands
> > (i.e. svn update) do step into external folders, even in the latest
> > version ;)
> > Does anyone else have any suggestion?
> Some commands step into externals, and some don't.
> E.g. commit certainly does *not* recurse into externals by default.
> I don't know what the plan was for 'svn diff' wrt externals.
> It happens to not recurse, but I'm not sure if this was a concious
> decision or not. Skimming the code (in libsvn_wc/diff.c) for 5 minutes
> does not really turn up an answer. It probably just happened to behave
> the way it does now, and nobody ever questioned this behaviour (until
> you showed up :)
> If you really care a lot about this, please file an issue asking
> for an enhancement such that svn diff shows modifications to externals
> by default, but does not if --ignore-externals is specified.
> But there really is no reason for defaulting to one way or the other,
> either use case is perfectly valid. Though having to list all externals
> explicitly to have them included in the diff is probably less easy than
> specifying --ignore-externals if you don't want to see modifications
> made to externals in the diff. Not sure.
Received on 2010-03-15 15:56:57 CET