Re: svn commit: r35444 - trunk/tools/client-side

2009-01-24 19:20:14 Blair Zajac napisał(a):
> On Jan 24, 2009, at 10:09 AM, Fabien Coelho wrote:
>
> > Author: fabien
> > Date: Sat Jan 24 10:09:36 2009
> > New Revision: 35444
> >
> > Log:
> > Fix completion bug for revert when only a prop was modified.
> >
> > * tools/client-side/bash_completion:
> > fix IFS not to ignore leading spaces from "svn status" output.
> >
> > Modified:
> > trunk/tools/client-side/bash_completion
> >
> > Modified: trunk/tools/client-side/bash_completion
> > URL: http://svn.collab.net/viewvc/svn/trunk/tools/client-side/bash_completion?pathrev=35444&r1=35443&r2=35444
> > =
> > =
> > =
> > =
> > =
> > =
> > =
> > =
> > ======================================================================
> > --- trunk/tools/client-side/bash_completion Sat Jan 24 04:52:57 2009
> > (r35443)
> > +++ trunk/tools/client-side/bash_completion Sat Jan 24 10:09:36 2009
> > (r35444)
> > @@ -44,8 +44,13 @@ function _svn_read_hashfile()
> > function _svn_grcut()
> > {
> > local re=$1 line=
> > + # fix IFS, so that leading spaces are not ignored by next read.
> > + # (there is a leading space in svn status output if only a prop
> > is changed)
> > + # outside IFS is not broken because the function is called in a
> > subshell
> > + IFS='
> > +'
> > while read -r line ; do
> > - [[ ! $re || $line == $re ]] && echo ${line/???????/}
> > + [[ ! $re || $line == $re ]] && echo "${line/???????/}"
> > done
>
> Don't you have to reset IFS when you're done with it, or is the change
> local to this function?

I have already fixed it before reading your e-mail in r35445.

-- 
Arfrever Frehtes Taifersar Arahesis