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

Re: [RFC] 1.4.x platform dependent blame output

From: Erik Huelsmann <ehuels_at_gmail.com>
Date: 2006-06-11 23:15:48 CEST

On 6/11/06, D.J. Heap <djheap@gmail.com> wrote:
> On 6/11/06, Erik Huelsmann <ehuels@gmail.com> wrote:
> > Hi!
> >
> > As part of solving issue #2431 (blame ignores svn:eol-style), blame
> > now gives different output for (some eol) -> native svn:eol-style
> > changes. Brane, Lundblad and I all agree this is not correct.
> >
> > There is also another issue: I actually think blame output is client
> > output (like svn status) rather than raw output (like svn cat). This
> > means the issue is invalid and needs no fixing: we just need to make
> > sure we generate the right platform specific eols.
> >
> > The data passed to the blame callback is insufficiently defined to
> > guarantee the right platform specific eol style, because it doesn't
> > say whether the file-eols are in- or excluded from the line data. In
> > practice, we pass CRs when the eol style is CRLF.
> >
> > If the eol style is CR we give completely broken results: the full
> > file is passed as line 1.
> >
> > I'd like your comments to this:
> >
> > - blame output is 'client output', not 'raw output'
> That seems like the natural thing to do to me -- I most often see
> blame output sent to a file and loaded into a text editor to be looked
> at more closely.

I have a patch which does this ...

> > - blame output should be the same for all platforms

... and this ...

> Does this mean in contrast to the above or something else?

What I mean by that item is: if a files svn:eol-style is switched from
LF -> native, that should show the same blame output on CRLF platforms
as on LF-only platforms. Currently, they give different outputs.

> > - we should fix the callback to use the right eol style
> > - we should fix the line data to be better defined
> > - all the above needs to be done before 1.4.0 final (I have the patches)
> Fixing the callback and the line data in the callback to be well
> defined seems like a very reasonable thing for a 3rd party client of
> the libraries to expect, so I would say yes.

... and this ...

So, I'll go and post the patch if no other reactions come in tomorrow.

Thanks for your comments!



To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Jun 11 23:16:14 2006

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