Do we know what broke here from history? Was the pool arg added
(breaking binary compat) in the 0.9 tree, or was this backing out a
bugged change?
At 01:00 PM 6/23/2004, [EMAIL PROTECTED] wrote:
>stoddard 2004/06/23 11:00:02
>
> Modified: xlate Tag: APU_0_9_BRANCH xlate.c
> Log:
> Fix build breakage on APR_HAVE_APR_ICONV path
>
> Revision Changes Path
> No revision
> No revision
> 1.17.2.3 +3 -3 apr-util/xlate/xlate.c
>
> Index: xlate.c
> ===================================================================
> RCS file: /home/cvs/apr-util/xlate/xlate.c,v
> retrieving revision 1.17.2.2
> retrieving revision 1.17.2.3
> diff -u -r1.17.2.2 -r1.17.2.3
> --- xlate.c 27 May 2004 21:55:02 -0000 1.17.2.2
> +++ xlate.c 23 Jun 2004 18:00:02 -0000 1.17.2.3
> @@ -178,9 +178,9 @@
> else {
> /* reset the iconv descriptor, since it's now in an undefined
> * state. */
> - apr_iconv_close(convset->ich);
> - convset->ich = apr_iconv_open(convset->topage, convset->frompage);
> - }
> + apr_iconv_close(convset->ich, convset->pool);
> + rv = apr_iconv_open(convset->topage, convset->frompage,
> convset->pool, &convset->ich);
> + }
> }
> #endif /* APU_HAVE_APR_ICONV */
>
>
>
>
