[email protected] wrote:
> Author: ylavic
> Date: Wed Jun 11 12:50:29 2014
> New Revision: 1601877
>
> URL: http://svn.apache.org/r1601877
> Log:
> mod_sed: Reuse ctx->bb in sed_response_filter() and be safe with its
> reentrance. The single return point helps to not duplicate cleanup code.
>
> Modified:
> httpd/httpd/trunk/modules/filters/mod_sed.c
>
> Modified: httpd/httpd/trunk/modules/filters/mod_sed.c
> URL:
> http://svn.apache.org/viewvc/httpd/httpd/trunk/modules/filters/mod_sed.c?rev=1601877&r1=1601876&r2=1601877&view=diff
> ==============================================================================
> --- httpd/httpd/trunk/modules/filters/mod_sed.c (original)
> +++ httpd/httpd/trunk/modules/filters/mod_sed.c Wed Jun 11 12:50:29 2014
> @@ -351,20 +349,19 @@ static apr_status_t sed_response_filter(
> status = sed_eval_buffer(&ctx->eval, buf, bytes, ctx);
> }
> if (status != APR_SUCCESS) {
> - clear_ctxpool(ctx);
> - return status;
> + break;
> }
> }
> apr_bucket_delete(b);
> }
> }
> - status = flush_output_buffer(ctx);
> - if (status != APR_SUCCESS) {
> - clear_ctxpool(ctx);
> - return status;
> + if (status == APR_SUCCESS) {
> + status = flush_output_buffer(ctx);
> }
> if (!APR_BRIGADE_EMPTY(ctx->bb)) {
> - status = ap_pass_brigade(f->next, ctx->bb);
> + if (status == APR_SUCCESS) {
> + status = ap_pass_brigade(f->next, ctx->bb);
> + }
> apr_brigade_cleanup(ctx->bb);
> }
> clear_ctxpool(ctx);
>
>
>
We have a small change in logic here. Maybe it makes sense. Previously we did
not cleanup ctx->bb if flush_output_buffer
returned != APR_SUCCESS. Now we do (only if not empty of course).
Regards
Rüdiger
