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

Re: ABI breakage

From: Philip Martin <philip_at_codematters.co.uk>
Date: 2005-03-23 23:10:23 CET

"Peter N. Lundblad" <peter@famlundblad.se> writes:

> Yeah, and then I think we can safely extend the struct.

Depends on your definition of safe, it is an ABI change and it is
trivial to write code that will be broken.

However the impenetrability of the libsvn_wc code is probably an
advantage for once! I suspect it's opaqueness means that there is
little, if any, outside code that manipulates the libsvn_wc data
structures. Does anyone know of any applications doing anything more
complex than calling the main svn_client_xxx/svn_wc_xxx functions? I
wonder if code generators might be affected? I suspect Swig only
really manipulates pointers and is thus OK.

As I noted before, the svn_ra_plugin_t already had an incompatible ABI
change in 1.1, so there is precedence for the svn_wc_entry_t change.
I think we might need to be more careful about our "binary
compatibility is guaranteed" claims.

Philip Martin
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Mar 23 23:11:46 2005

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

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.