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

Re: svn commit: r1575428 - /subversion/trunk/subversion/libsvn_fs_fs/fs.h

From: Stefan Fuhrmann <stefan.fuhrmann_at_wandisco.com>
Date: Sat, 8 Mar 2014 00:11:52 +0100

On Fri, Mar 7, 2014 at 11:32 PM, Ben Reser <ben_at_reser.org> wrote:

> This is downright horrible for code maintainability. We have a type for
> revision numbers.
> It's bad enough we don't have a stable type for revision numbers across
> platforms. But now you're abusing it and using 32-bit integers in some
> places
> for them and 64-bit values in others.
> Now that you're doing this I'm now of the opinion that I should veto the
> code
> that limits them to 32-bits at the RA layer. Because all it's done is
> encourage us to do things like this.

Thanks for the review. r1575447 fix the inconsistency.

Two comments:

* The keys are being used one-way, i.e. any type > input type will do
* Padding is horrible because it inserts sections of random data
  into a *hash* key. Valgrind might find these instances, if run on Windows.

-- Stefan^2.
Received on 2014-03-08 00:12:23 CET

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