On Mon, 9 May 2005 kfogel@tigris.org wrote:
> Author: kfogel
> Date: Mon May 9 01:07:30 2005
> New Revision: 14643
>
> Modified: trunk/subversion/include/svn_subst.h
> Url: http://svn.collab.net/viewcvs/svn/trunk/subversion/include/svn_subst.h?rev=14643&p1=trunk/subversion/include/svn_subst.h&p2=trunk/subversion/include/svn_subst.h&r1=14642&r2=14643
> ==============================================================================
> --- trunk/subversion/include/svn_subst.h (original)
> +++ trunk/subversion/include/svn_subst.h Mon May 9 01:07:30 2005
> @@ -126,7 +126,10 @@
> svn_boolean_t compare_values);
>
>
> -/** Copy and translate the data in stream @a src into stream @a dst. It is
> +/**
> + * @since New in 1.2.
> + *
> + * Copy and translate the data in stream @a src into stream @a dst. It is
> * assumed that @a src is a readable stream and @a dst is a writable stream.
> *
> * If @a eol_str is non-@c NULL, replace whatever bytestring @a src uses to
> @@ -147,6 +150,10 @@
> * Detect only keywords that are no longer than @c SVN_IO_MAX_KEYWORD_LEN
> * bytes, including the delimiters and the keyword itself.
> *
> + * Use @a pool for temporary allocation. For compatibility, if @a pool
> + * is @c NULL, then use a new subpool of the global pool for temporary
> + * allocation, and destroy the subpool before returning.
> + *
why not handle this case in the wrapper isntead (by creating a pool and
destroying it after the call)? Do we need the NULL pool handling in the
new API?
(I'll do a complete review tonight, if I don't fall asleep while going to
bed "just for five minutes", which I did yesterday:-)
Thanks,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon May 9 10:06:06 2005