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

Re: trees and test suites -- a proposal

From: Sam TH <sam_at_uchicago.edu>
Date: 2001-04-18 19:59:34 CEST

On Wed, Apr 18, 2001 at 10:17:27AM -0500, Ben Collins-Sussman wrote:
> Wow, quite a lot of work!

It was just this morning (but starting pretty early).

> Have you looked at svn_output.py? I'm worried that you've reinvented
> a lot of things I've already done.

I did look at the stuff you had done, and I think most of what I did
is different.

> 1. The compare_sets() routine takes two lists and a comparison func,
> and decides if the lists are the same, regardless of ordering.
> For comparing actual vs. expected output, I'm handing it a list of
> output lines and a list of regexps. The 'comparison' func tries to
> match a line with a regexp.

I think the problem here is that not everything maps nicely to a list
the way everything maps nicely to a tree (mostly because it's a
representation of a tree). For example, to compare SVN/entries to
CVS/Entries would require a really hairy comparison function, which
would fundamentally have to convert both of them to some intermediate
format and compare them. What I'm proposing is to make that explicit,
and use the fact that they're all trees to our advantage.

> 2. svn_entry.py loads up an entries file from disk into a hash of
> hashes, just like the C function svn_wc_get_entries. The TODO comment
> you saw was referring to the fact that I was planning on writing
> another comparison func that compared entry objects, and again using
> this func with compare_sets().

As I mentioned above, comparing entry objects to some other kind of
object would be really painful, I think. And your svn_entry only
parses one entries file, so it doesn't let you look at the same
picture as the whole list of files you just checked out. But the code
you've already written is most of the work done to transform a whole
working-copy worth of entries files into a tree.

> I'll be working on python stuff later this afternoon... and I'll take
> a closer look at your stuff. Maybe your system is better. :)

Well, I like my system, but I'm definitely flexible on the design. :-)
sam th --- sam_at_uchicago.edu --- http://www.abisource.com/~sam/
OpenPGP Key: CABD33FC --- http://samth.dyndns.org/key
DeCSS: http://samth.dyndns.org/decss

  • application/pgp-signature attachment: stored
Received on Sat Oct 21 14:36:28 2006

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