[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: validate_eol_prop_against_file() inefficiency?

From: C. Michael Pilato <cmpilato_at_collab.net>
Date: Wed, 02 Jan 2008 13:55:51 -0500

C. Michael Pilato wrote:
> Mark Phippard wrote:
>> On Dec 28, 2007 1:13 PM, C. Michael Pilato <cmpilato_at_collab.net> wrote:
>>> While updating the svnbook, I noticed something odd. In Subversion 1.4,
>>> when you try to set svn:eol-style on a binary file, you get the obvious error:
>>> % svn pset svn:eol-style CRLF suse.png
>>> svn: File 'suse.png' has binary mime type property
>>> In trunk, however, you get something far more bizarre:
>>> % svn pset svn:eol-style CRLF suse.png
>>> subversion/libsvn_wc/props.c:2425: (apr_err=200009)
>>> svn: File 'suse.png' has inconsistent newlines
>>> subversion/libsvn_subr/subst.c:736: (apr_err=135000)
>>> svn: Inconsistent line ending style
>>> I see that libsvn_wc/props.c:validate_eol_prop_against_file() has been
>>> ordered differently in trunk. What I can't figure out is why the
>>> far-more-expensive newline consistency check was moved ahead of the
>>> super-cheap is-binary check? Besides being really inefficient, the new
>>> error is just not as clear.
>>> Was this fallout from the URL propsetting support or something?
>> Not sure, but wasn't there also a patch recently to add validation of
>> the eol-style values? Perhaps that is when this happened.
> Well, I'm going to make the obvious optimization (which will restore the
> more-obvious error message in the case presented).

Oops, disregard this. I hadn't yet read David's mail, which points out why
this little switcheroo won't accomplish much.

C. Michael Pilato <cmpilato_at_collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Received on 2008-01-04 06:38:26 CET

This is an archived mail posted to the Subversion Dev mailing list.