RE: svn commit: rev 4882 - trunk/subversion/libsvn_fs

> From: Karl Fogel [mailto:kfogel@newton.ch.collab.net]
> Sent: Thursday, February 13, 2003 9:37 PM

> "Sander Striker" <striker@apache.org> writes:
> > Bah. Just cache the md5 digest in the baton in close_stream.
> > txn_body_write_close_rep can use it directly from there. No hacky memcpy
> > or apr_md5_ctx_dup needed.
>
> There's some disagreement about what's more hacky here :-).
>
> Ben's solution has the nice property that the baton's state remains
> the same (and useable) no matter how many times the txn is retried.

Bah humbug. Same goes for my solution.

> In other words, the digest is really a tmp variable, and so is the
> *finalized* context, so why cache them anywhere if we don't have to?

Why copy the unfinalized context, finalize, on every retry when you don't
have to? In the end, it is about the final digest, not the unfinalized
apr md5 context.

Sander

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Feb 13 23:21:25 2003