Re: svn commit: rev 2803 - trunk/subversion/include trunk/subversion/libsvn_wc trunk/subversion/libsvn_client trunk/subversion/tests/clients/cmdline

Philip Martin <philip@codematters.co.uk> writes:
> In the case above it's a pointer-to-pointer, so I could make it
>
> "Return a pointer to a new access baton in *ADM_ACCESS"

Right, that's the only case I'm talking about.

> Lots of functions take a parameter "svn_wc_adm_access_t *adm_access"
> and I've generally written things like
>
> "Ensure ADM_ACCESS has a write lock".
>
> Do you think that should be changed as well? svn_wc_adm_access_t is
> never passed by value.

No, not those. They're fine the way they are.

> Do people prefer "working copy context" to "access baton"?

I like access baton myself.

> > Btw, we haven't generally done advance prototyping like this, except
> > where absolutely necessary. Any objection to just moving things
> > around so you don't need the extra prototype?
>
> As I remember, the use of pool cleanup handlers means that at least
> one function needs a forward declaration. So it's absolutely
> necessary.

Oh, I see what you mean, thanks. Okay, then; imho documenting it at
the definition itself is best, but your call.

-Karl

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Jul 30 23:25:37 2002