Re: [mfischer@php.net: Bug #12465 Updated: posix_* issuing Warnings, no error code.]

From: Date: Fri, 01 Mar 2002 09:59:36 +0000
Subject: Re: [mfischer@php.net: Bug #12465 Updated: posix_* issuing Warnings, no error code.]
References: 1  Groups: php.dev 
Request: Send a blank email to php-dev+get-80547@lists.php.net to get a copy of this message
Markus Fischer wrote:
    I'm forwarding this to php-dev@ (so people doing filtering
    don't miss it ;) and let's see what opinion others have about
    this.
    - Markus
Even if I bogusified(?) it, +1 for return FALSE and set error level to E_NOTICE :) -- Yasuo Ohgaki
------------------------------------------------------------------------ Subject: Bug #12465 Updated: posix_* issuing Warnings, no error code. From: mfischer@php.net Date: 28 Feb 2002 22:48:58 -0000 To: php-bugs@lists.php.net
 ID:               12465
 Updated by:       mfischer@php.net
 Reported By:      herp@wildsau.idv.uni-linz.ac.at
-Status:           Bogus
+Status:           Open
 Bug Type:         POSIX related
Operating System: Linux
 PHP Version:      4.0.6
New Comment: I agree with Herbert here. It's pretty worthless for the function to generate this verbose error messages during runtime. It's mimic is excatly what the C version does (function-wise). Does the C version do any output itself if it encounters an error (e.g. posix_getpwuid) ? No. Why? Because a NULL return value is a valid return value. It's not a php_error() nor a E_WARNING. Instead, the extension should be re-written to a) return false (the PHP-way), b) store the errno in a thread-global contect variable and c) provide a function to cleanly retrieve the exacty error message (currently, you would habe to catch $php_errmsg (or whatever it was) and parse it out. Short: re-opening ;) Previous Comments: ------------------------------------------------------------------------ [2002-02-28 14:48:28] herp@wildsau.idv.uni-linz.ac.at It?s still nonsense to write an error-message! stat() *IS STILL USED* for checking the existence of files. Why do I have to *suppress* error-messages?! PHP should not *generate* them in the first! *If* you choose to implement stat() and name it after the C-function, then stat() should behave as closely as possible like its C-equivalent. ------------------------------------------------------------------------ [2002-02-04 02:46:54] yohgaki@php.net It returns false. You need to get rid of error messages with @... ------------------------------------------------------------------------ [2001-07-30 09:24:17] herp@wildsau.idv.uni-linz.ac.at hi, I tried to use some of the posix_* functions to work with user-accounts on the system, like "posix_getpwnam()" and "posix_getpwuid()". I expected to get an error-code back (like Failed or FALSE) for pwnames or pwuids which do not exist in /etc/passwd. Instead, PHP will write a warning message in my html-output: : Warning: posix_getpwuid(4711) failed with 'Success' in : /data/home/webmaster/admin/admin.php : on line 197 and, what I think is strange, will "fail with ?Success?". Could you please modify the posix_getpw* functions in a way that they 1) do not write strange warning-messages and 2) return a NULL-Value or FALSE, where the unix getpw*(3) will return NULL (just like documented in the man-page) thanks in advance, herbert rosmanith herp@wildsau.idv.uni-linz.ac.at ------------------------------------------------------------------------


« previous php.dev (#80547) next »