Derick Rethans wrote:
> I'm not happy for two reasons:
> 1. The Test for buffering in core functions with implicit flush on
> (009.phpt) still fails
> 2. I think it's better that somebody who really understands it all (the
> guy who wrote it) has a look at it before you commit things.
>
I'm not happy since you obviously not reading my patch well
and insist to me to revert the patch. Not even have to mention
you don't understand how this thing is working.
Read comment ask specific code to be changed if you can read.
Or the comment is broken hard to be understood? If so, I apologize.
If you understand comment, you should be able to ask to remove
PG(implict_flush) specifically.
When you asking to revert a specific patch, read and comment
about it just asking to revert.
BTW, what kind of OS needs flush() for charactor devices?
--
Yasuo Ohgaki
Index: php4/main/output.c
diff -u php4/main/output.c:1.123 php4/main/output.c:1.124
--- php4/main/output.c:1.123 Mon Sep 30 19:46:43 2002
+++ php4/main/output.c Mon Sep 30 22:43:33 2002
@@ -18,7 +18,7 @@
+----------------------------------------------------------------------+
*/
-/* $Id: output.c,v 1.123 2002/09/30 23:46:43 yohgaki Exp $ */
+/* $Id: output.c,v 1.124 2002/10/01 02:43:33 yohgaki Exp $ */
#include "php.h"
#include "ext/standard/head.h"
@@ -593,6 +593,14 @@
memcpy(target, text, text_length);
target[text_length]=0;
+ /* If implicit_flush is On, send contents to next buffer and return.
+ Both PG() and OG() should be used since we should flush implicitly
+ always when implicit_flush is enabled in php.ini */
+ if (PG(implicit_flush) || OG(implicit_flush)) {
+ php_end_ob_buffer(1, 1 TSRMLS_CC);
+ return;
+ }
+
if (OG(active_ob_buffer).chunk_size
&& OG(active_ob_buffer).text_length >=
OG(active_ob_buffer).chunk_size) {
zval *output_handler = OG(active_ob_buffer).output_handler;
--
PHP CVS Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
