Req #48532 [Asn]: Allow pg_fetch_all() to index numerically
Edit report at https://bugs.php.net/bug.php?id=48532&edit=1 ID: 48532 Updated by: yohg...@php.net Reported by:strata_ranger at hotmail dot com Summary:Allow pg_fetch_all() to index numerically Status: Assigned Type: Feature/Change Request Package:PostgreSQL related PHP Version:5.2.9 Assigned To:yohgaki Block user comment: N Private report: N New Comment: Why do you want numeric key? Previous Comments: [2009-06-11 15:10:22] strata_ranger at hotmail dot com Description: It could be useful if pg_fetch_all() were extended to support indexing each row's fields numerically (like pg_fetch_row() does) in addition to the current behavior of indexing them associatively. It is possible in some obscure cases that an associative array is unable to contain all columns from the result set, i.e. if the query returns duplicate column names/aliases. Reproduce code: --- --- >From manual page: function.pg-fetch-all --- -- Edit this bug report at https://bugs.php.net/bug.php?id=48532&edit=1
Req #42398 [Asn->Wfx]: pg_(p)connect produces uninformative error messages
Edit report at https://bugs.php.net/bug.php?id=42398&edit=1 ID: 42398 Updated by: yohg...@php.net Reported by:webmaster at touchingvirus dot net Summary:pg_(p)connect produces uninformative error messages -Status: Assigned +Status: Wont fix Type: Feature/Change Request Package:PostgreSQL related Operating System: Windows PHP Version:5.2.3 Assigned To:yohgaki Block user comment: N Private report: N New Comment: This is because PostgreSQL library set these uninformative error message. php > $db = pg_connect('host=192.168.100.50 user=postgres password=dummy'); Warning: pg_connect(): Unable to connect to PostgreSQL server: could not connect to server: Connection refused Is the server running on host "192.168.100.50" and accepting TCP/IP connections on port 5432? in php shell code on line 1 Previous Comments: [2007-08-23 13:45:39] webmaster at touchingvirus dot net Description: When connecting to a pgSQL server (I use 8.2) with a wrong username, password or port in the connection string, the error given is really uninformative to the user and actually suggests a problem with the postgreSQL server: "Unable to connect to PostgreSQL server: server closed the connection unexpectedly This probably means the server terminated abnormally before or while processing the request." The server didn't terminate & the connection was refused/aborted Expected result: To be honest, I expected an error that was similar to that of mySQL - though I know that is because mySQL actually produces good error messages =) e.g. Access denied for: root@localhost (Using Password: YES) The psql binary errors with the expect errors like: psql: FATAL: password authentication failed for user "root" -- Edit this bug report at https://bugs.php.net/bug.php?id=42398&edit=1
Req #24287 [Asn->Csd]: Using user defined Domains
Edit report at https://bugs.php.net/bug.php?id=24287&edit=1 ID: 24287 Updated by: yohg...@php.net Reported by:zive at apollo dot lv Summary:Using user defined Domains -Status: Assigned +Status: Closed Type: Feature/Change Request Package:PostgreSQL related Operating System: Slackware Linux 9 PHP Version:4.3.3RC2-dev Assigned To:yohgaki Block user comment: N Private report: N New Comment: Fix was done already. Previous Comments: [2003-07-22 11:46:11] il...@php.net Normally you should specify the schema along with the table name ([schema].[table_name]) when using pg_(insert/update/convert). Therefor this is not a bug, however I do see the use for this functionality, so I making this report a feature request. [2003-07-22 09:46:06] rorezende at hotmail dot com This occurr because the PHP do not include schema name in pg_insert, pg_update, pg_delete. [2003-06-22 14:24:14] zive at apollo dot lv I tested on newest PHP version 4.3.3RC2-dev and that problesm still exists [2003-06-22 13:35:59] zive at apollo dot lv Description: In PostgreSQL 7.3.2 database I created domain - "public.nauda" (monetary data type) ar Number (8,2). I have table on non-public schema - mySchema.myTable with column 'spec_price' with data type "public.nauda". When I am calling pg_insert(), pg_update or pg_convert, I receive following error: Notice: pg_convert(): Unknown or system data type 'nauda' for 'spec_price'. in xxx.php on line n I am using variable $ids for containing ids (primary key of table) value if I am editing data, and $ids="new" if I am inserting new datas. Reproduce code: --- $data = array('proset_ids'=>$cmbProsetIds, 'spec_price'=>$txtSpecPrice, ... ); if ($ids=="new") pg_insert($db, "tb_product_of_week", $data); else pg_update($db, "tb_product_of_week", $data, array('ids'=>$ids)); $ids = ""; Expected result: I am expecting result that data are inserted/updated in database. Actual result: -- Notice: pg_update(): Unknown or system data type 'nauda' for 'spec_price'. in xxx.php on line n -- Edit this bug report at https://bugs.php.net/bug.php?id=24287&edit=1
[PHP-BUG] Req #65406 [NEW]: Enchant broker plugins are in the wrong place in windows builds
From: bill at zeroedin dot com Operating system: windows PHP version: 5.4.17 Package: Enchant related Bug Type: Feature/Change Request Bug description:Enchant broker plugins are in the wrong place in windows builds Description: As noted in bug #64593 and bug #64397, libenchant.dll now looks for the included broker plugins libenchant_ispell.dll and libenchant_myspell.dll in /lib/enchant instead of . The windows builds provided at windows.php.net have the broker plugin dlls in the directory. Please change the build process to place the enchant broker plugins in the correct location. -- Edit bug report at https://bugs.php.net/bug.php?id=65406&edit=1 -- Try a snapshot (PHP 5.4): https://bugs.php.net/fix.php?id=65406&r=trysnapshot54 Try a snapshot (PHP 5.3): https://bugs.php.net/fix.php?id=65406&r=trysnapshot53 Try a snapshot (trunk): https://bugs.php.net/fix.php?id=65406&r=trysnapshottrunk Fixed in SVN: https://bugs.php.net/fix.php?id=65406&r=fixed Fixed in release: https://bugs.php.net/fix.php?id=65406&r=alreadyfixed Need backtrace: https://bugs.php.net/fix.php?id=65406&r=needtrace Need Reproduce Script: https://bugs.php.net/fix.php?id=65406&r=needscript Try newer version: https://bugs.php.net/fix.php?id=65406&r=oldversion Not developer issue:https://bugs.php.net/fix.php?id=65406&r=support Expected behavior: https://bugs.php.net/fix.php?id=65406&r=notwrong Not enough info: https://bugs.php.net/fix.php?id=65406&r=notenoughinfo Submitted twice: https://bugs.php.net/fix.php?id=65406&r=submittedtwice register_globals: https://bugs.php.net/fix.php?id=65406&r=globals PHP 4 support discontinued: https://bugs.php.net/fix.php?id=65406&r=php4 Daylight Savings: https://bugs.php.net/fix.php?id=65406&r=dst IIS Stability: https://bugs.php.net/fix.php?id=65406&r=isapi Install GNU Sed:https://bugs.php.net/fix.php?id=65406&r=gnused Floating point limitations: https://bugs.php.net/fix.php?id=65406&r=float No Zend Extensions: https://bugs.php.net/fix.php?id=65406&r=nozend MySQL Configuration Error: https://bugs.php.net/fix.php?id=65406&r=mysqlcfg
Req #29416 [Wfx]: ob_include function
Edit report at https://bugs.php.net/bug.php?id=29416&edit=1 ID: 29416 User updated by:edwin at cheatah dot nl Reported by:edwin at cheatah dot nl Summary:ob_include function Status: Wont fix Type: Feature/Change Request Package:*General Issues PHP Version:Irrelevant Block user comment: N Private report: N New Comment: Are you serious? This is what you come up with after 9 years of contemplation? Actually I haven't been a web developer for over 5 years now, so I don't need the feature anymore. But you should be ashamed of yourself. First of all for leaving open a feature request this long, but even more so for not understanding the problem and for your worthless answer. Nice demonstration of incompetence. Previous Comments: [2013-08-06 08:01:45] yohg...@php.net $message = ob_include ( './tpl/mail.php' ); You should use $message = file_get_contents(); since it is unrelated to "output". [2004-07-28 01:30:30] edwin at cheatah dot nl Description: I often use something like the following code: ob_start (); include './tpl/mail.php'; $message = ob_get_contents (); ob_end_flush (); I would very much like to see a function of control stucture like this: $message = ob_include ( './tpl/mail.php' ); // or $message = ob_include './tpl/mail.php'; Of course ob_include_once, ob_require and ob_require_once behave similarly. (Creating a custom function in PHP code is NOT possible, for the function's scope would be used.) -- Edit this bug report at https://bugs.php.net/bug.php?id=29416&edit=1
Bug #65404 [Opn->Nab]: OO with screen of death
Edit report at https://bugs.php.net/bug.php?id=65404&edit=1 ID: 65404 Updated by: johan...@php.net Reported by:info at djdb dot be Summary:OO with screen of death -Status: Open +Status: Not a bug Type: Bug Package:Output Control Operating System: win (irrelevant) PHP Version:Irrelevant Block user comment: N Private report: N New Comment: Thank you for taking the time to write to us, but this is not a bug. Please double-check the documentation available at http://www.php.net/manual/ and the instructions on how to report a bug at http://bugs.php.net/how-to-report.php Variables can be reassigned. Previous Comments: [2013-08-06 13:34:19] info at djdb dot be Description: file--1 $myobject = new object1(); class object1{ ... } file--2 $myobject = new object2(); include(file--1) class object2{ ... } so the name of variable of instanties arr the same but no error message is write so i was many time watching wat there was wrong (the file was to big to place them here) Test script: --- file--1 $myobject = new object1(); class object1{ ... } file--2 $myobject = new object2(); include(file--1) class object2{ ... } -- Edit this bug report at https://bugs.php.net/bug.php?id=65404&edit=1
[PHP-BUG] Bug #65404 [NEW]: OO with screen of death
From: info at djdb dot be Operating system: win (irrelevant) PHP version: Irrelevant Package: Output Control Bug Type: Bug Bug description:OO with screen of death Description: file--1 $myobject = new object1(); class object1{ ... } file--2 $myobject = new object2(); include(file--1) class object2{ ... } so the name of variable of instanties arr the same but no error message is write so i was many time watching wat there was wrong (the file was to big to place them here) Test script: --- file--1 $myobject = new object1(); class object1{ ... } file--2 $myobject = new object2(); include(file--1) class object2{ ... } -- Edit bug report at https://bugs.php.net/bug.php?id=65404&edit=1 -- Try a snapshot (PHP 5.4): https://bugs.php.net/fix.php?id=65404&r=trysnapshot54 Try a snapshot (PHP 5.3): https://bugs.php.net/fix.php?id=65404&r=trysnapshot53 Try a snapshot (trunk): https://bugs.php.net/fix.php?id=65404&r=trysnapshottrunk Fixed in SVN: https://bugs.php.net/fix.php?id=65404&r=fixed Fixed in release: https://bugs.php.net/fix.php?id=65404&r=alreadyfixed Need backtrace: https://bugs.php.net/fix.php?id=65404&r=needtrace Need Reproduce Script: https://bugs.php.net/fix.php?id=65404&r=needscript Try newer version: https://bugs.php.net/fix.php?id=65404&r=oldversion Not developer issue:https://bugs.php.net/fix.php?id=65404&r=support Expected behavior: https://bugs.php.net/fix.php?id=65404&r=notwrong Not enough info: https://bugs.php.net/fix.php?id=65404&r=notenoughinfo Submitted twice: https://bugs.php.net/fix.php?id=65404&r=submittedtwice register_globals: https://bugs.php.net/fix.php?id=65404&r=globals PHP 4 support discontinued: https://bugs.php.net/fix.php?id=65404&r=php4 Daylight Savings: https://bugs.php.net/fix.php?id=65404&r=dst IIS Stability: https://bugs.php.net/fix.php?id=65404&r=isapi Install GNU Sed:https://bugs.php.net/fix.php?id=65404&r=gnused Floating point limitations: https://bugs.php.net/fix.php?id=65404&r=float No Zend Extensions: https://bugs.php.net/fix.php?id=65404&r=nozend MySQL Configuration Error: https://bugs.php.net/fix.php?id=65404&r=mysqlcfg
Bug->Req #47664 [Asn->Csd]: get_class returns NULL instead of FALSE.
Edit report at https://bugs.php.net/bug.php?id=47664&edit=1 ID: 47664 Updated by: yohg...@php.net Reported by:wiebe at wiebelt dot nl Summary:get_class returns NULL instead of FALSE. -Status: Assigned +Status: Closed -Type: Bug +Type: Feature/Change Request Package:Class/Object related -Operating System: Ubuntu 8.10 +Operating System: any -PHP Version:5.3.0beta1 +PHP Version:master Assigned To:dmitry Block user comment: N Private report: N New Comment: It is documented as get_class() raise E_WARNING now. I feel E_WARNING is unnecessary for get_class() as it requires is_object() call for error free programming. e.g. factory may fail and return false when something wrong, etc. Anyway, bug is fixed and E_WARNING error is documented. If anyone feel it should not raise E_WARNING, create new entry for it. Closing. Previous Comments: [2009-07-28 13:14:52] der...@php.net The documentation clearly states: "Returns the name of the class of which object is an instance. Returns FALSE if object is not an object." It doesn't say "shows a warning when object is not an object", so this should be fixed according to what PHP 5.2.x and the manual say. [2009-07-28 13:06:48] dmi...@php.net I would prefer to stay the warning, as it indicates a problem in user code that probably need to be fixed. [2009-06-30 12:33:05] der...@php.net I've a patch at: http://php.pastebin.com/m20d61023 [2009-06-23 11:19:37] der...@php.net This isn't fixed, the warning is still there which is a part of the problem here. [2009-03-16 09:53:50] dmi...@php.net This bug has been fixed in CVS. Snapshots of the sources are packaged every three hours; this change will be in the next snapshot. You can grab the snapshot at http://snaps.php.net/. Thank you for the report, and for helping us make PHP better. The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report online at https://bugs.php.net/bug.php?id=47664 -- Edit this bug report at https://bugs.php.net/bug.php?id=47664&edit=1
Req #14615 [Opn->Csd]: PHP log does not contain proper CRLF's
Edit report at https://bugs.php.net/bug.php?id=14615&edit=1 ID: 14615 Updated by: yohg...@php.net Reported by:timvandermolen at hotmail dot com Summary:PHP log does not contain proper CRLF's -Status: Open +Status: Closed Type: Feature/Change Request Package:PHP options/info functions Operating System: Windows PHP Version:* -Assigned To: +Assigned To:yohgaki Block user comment: N Private report: N New Comment: At least 5.5 uses CR/LF under windows. Previous Comments: [2001-12-20 04:52:46] hholz...@php.net now this is strange ... the following snippet from php4/main/main.c (around line 300) does the actual logging and as it does *not* open the logfile in binary mode it should automaticly convert '\n' to whatever the local platform prefers??? log_file = VCWD_FOPEN(PG(error_log), "a"); if (log_file != NULL) { time(&error_time); strftime(error_time_str, 128 , "%d-%b-%Y %H:%M:%S" , php_localtime_r(&error_time, &tmbuf)); fprintf(log_file, "[%s] ", error_time_str); fprintf(log_file, "%s", log_message); fprintf(log_file, "\n"); fclose(log_file); return; } [2001-12-20 04:36:06] mfisc...@php.net Reopen, leave Type: to Feature Request. [2001-12-20 04:28:40] timvandermolen at hotmail dot com "but i'm not really sure what log we are talking about here" I am talking about the PHP log file which can be defined with the directive "error_log = ..." in the PHP.INI file. [2001-12-20 04:23:54] timvandermolen at hotmail dot com There's nothing wrong with the editor. I have used Notepad to view the PHP log. I have also viewed the PHP log with Visual C++ which *converts* (note the word convert) the LF to CRLF. Appearently, CRLF is the preferred format in Windows for new lines above LF. Besides, CRLF is the new line format which practically every editor in Windows supports. And besides everything: CRLF *is* the Windows newline, LF is not. So it would not make sense to use LF as a new line in Windows environments. [2001-12-20 04:23:14] hholz...@php.net well, different systems *have* different default line ending styles, so we might want to write whatever is considered 'native' on a certain platform but i'm not really sure what log we are talking about here, as the default log mechanism (as i know it on Linux/Apache) is to pass log messages to the webserver for logging? The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report online at https://bugs.php.net/bug.php?id=14615 -- Edit this bug report at https://bugs.php.net/bug.php?id=14615&edit=1
Req #47345 [Opn->Csd]: Cloning non-object does not throw exception?
Edit report at https://bugs.php.net/bug.php?id=47345&edit=1 ID: 47345 Updated by: yohg...@php.net Reported by:alt dot e-mail at gmx dot de Summary:Cloning non-object does not throw exception? -Status: Open +Status: Closed Type: Feature/Change Request Package:Class/Object related Operating System: Irrelevant PHP Version:5.3.0beta1 -Assigned To: +Assigned To:yohgaki Block user comment: N Private report: N New Comment: It seems fixed. [yohgaki@dev PHP-5.5]$ ./php-bin t.php PHP Fatal error: __clone method called on non-object in /home/yohgaki/ext/git/oss/php.net/PHP-5.5/t.php on line 28 Previous Comments: [2009-02-09 09:34:07] alt dot e-mail at gmx dot de Addendum: The code for my suggestion should really be: $newColors = array(); foreach ( $colors as $color ) { try { $newColors[] = clone $color; } // do nothing, error in __clone-method or whatever catch ( MySomethingException $e ) {} // Not an object, can't be cloned, "TypeException" would be // the exception PHP throws for non-clonables. // Really do not care what that name is, just that there is one. catch ( TypeException $e ) { $newColors[] = $color; } } [2009-02-09 09:27:01] alt dot e-mail at gmx dot de Description: Since v5.2.5, cloning a non-object results in a fatal error. In my opinion that is the wrong behaviour! It's not really a fatal error (as in DEATH, CORE-DUMP, END-OF-THE-WORLD) to clone a non-object. It's really just an exception (as in OOPS, HELP?, DO-WHAT-NOW?), that should be user-catchable. A practical example: I have an array where the values are strings or objects that can be cast to string with "__toString". If I want to create a copy of that array, I also need to clone the objects. Reproduce code: --- r*255, (int)$this->g*255, (int)$this->b*255); } } // somewhere in your code you create an array of colors $colors = array( "#FFF000", new Color( 0, 0, 1 ), // and many more ); // somewhere else you want to copy that array // this is the current way $newColors = array(); foreach ( $colors as $color ) { if ( is_object( $color ) { try { $newColors[] = clone $color; } // do nothing, error in __clone-method or whatever catch ( MySomethingException $e ) {} } else { $newColors[] = $color; } } // this is the way it should be $newColors = array(); foreach ( $colors as $color ) { try { $newColors[] = clone $color; } catch ( TypeException/*or something*/ $e ) { $newColors[] = $color; } } -- Edit this bug report at https://bugs.php.net/bug.php?id=47345&edit=1
Req #47468 [Opn->Csd]: enable readline for embed sapi
Edit report at https://bugs.php.net/bug.php?id=47468&edit=1 ID: 47468 Updated by: yohg...@php.net Reported by:admin at mateuszjanowski dot pl Summary:enable readline for embed sapi -Status: Open +Status: Closed Type: Feature/Change Request Package:Other web server Operating System: * PHP Version:5.* -Assigned To: +Assigned To:yohgaki Block user comment: N Private report: N New Comment: It seems this issue is fixed. Previous Comments: [2009-04-25 15:38:26] admin at mateuszjanowski dot pl Hi. Yes can i confirm - Paul's patch fixes the problem. Thank you :). [2009-04-25 13:11:17] anil at saog dot net Thank you very much, it works now. Before your solution I tried to change ext/pcntl/config.m4, a line in this file contains "cli" as a function variable; --- PHP_NEW_EXTENSION(pcntl, pcntl.c php_signal.c, $ext_shared, cli) +++ PHP_NEW_EXTENSION(pcntl, pcntl.c php_signal.c, $ext_shared) but i didnt try it, this may also help. [2009-04-25 12:40:22] paul dot biggar at gmail dot com I believe this also applies to ncurses, pcntl, and libedit. Here is a patch to fix it: --- aclocal.m4.orig 2009-04-25 13:11:57.0 +0100 +++ aclocal.m4 2009-04-25 13:24:50.0 +0100 @@ -962,7 +962,7 @@ if test "$3" != "shared" && test "$3" != "yes" && test "$4" = "cli"; then dnl -- CLI static module [PHP_]translit($1,a-z_-,A-Z__)[_SHARED]=no -if test "$PHP_SAPI" = "cgi"; then +if test "$PHP_SAPI" = "cgi" || test "$PHP_SAPI" = "embed"; then PHP_ADD_SOURCES(PHP_EXT_DIR($1),$2,$ac_extra,) EXT_STATIC="$EXT_STATIC $1" else I tested it by building PHP 5.2.6, 5.2.9 and php5.3-200902131730, building with ./configure --enable-embed --with-readline --enable-pcntl --with-ncurses and running nm .libs/libphp5.so | grep readline (and similarly for pcntl and ncurses) which verifies that the readline symbols are available. I also ran the test script from Mateusz through phc, and observed that calls to readline now succeeded. Can the patch be considered for 5.2.10, 5.3 and HEAD? [2009-02-23 14:16:38] paul dot biggar at gmail dot com Here is the context: http://www.phpcompiler.org/lists/phc-general/2009-February/000907.html I would mark this as a bug, not a feature request. The readline functions are not available in the embed SAPI. However, there is no reason that it should be impossible for an embed user to have access to readline. [2009-02-21 17:47:26] admin at mateuszjanowski dot pl Description: Hello, is my first bug report; I'm using php compiler (phc - http://www.phpcompiler.org/) with php CLI [5.2.8]. I meet first problem when i started to use readline extension. I compiled php with '--with readline' is working ok but when i compile my program using phc i get; 'Fatal error: status(): Call to undefined function readline() in rcon.php on line 52'. The phc author tell me to do; "I would recommend filing a bug with PHP, asking for the readline extension to be included in the embed SAPI. It seems like an oversight rather than a feature. In addition, it might be straightforward to create a patch to fix this yourself. I took a quick look at this, but I couldn't find where to do this. You might have more luck." Btw i compiled PHP 5.2.9RC2 version and getting the same error. Mabye you can help me? Greetings Mateusz Janowski. Reproduce code: --- http://mateuszjanowski.pl/rcon.phps Expected result: 0.2 Alpha - www.torvalds.pl Remote Console: IP:PORT... Rcon: Actual result: -- When i run php -f rcon.php is working OK but after compiling nope; 0.2 Alpha - www.torvalds.pl Remote Console: IP:PORT... Fatal error: status(): Call to undefined function readline() in rcon.php on line 52 -- Edit this bug report at https://bugs.php.net/bug.php?id=47468&edit=1
Bug #50283 [Opn->Csd]: allow base in gmp_strval to use full range: 2 to 62, and -2 to -36
Edit report at https://bugs.php.net/bug.php?id=50283&edit=1 ID: 50283 Updated by: yohg...@php.net Reported by:asphp at dsgml dot com Summary:allow base in gmp_strval to use full range: 2 to 62, and -2 to -36 -Status: Open +Status: Closed Type: Bug Package:GNU MP related Operating System: Linux PHP Version:5.2.11 Assigned To:dsp Block user comment: N Private report: N Previous Comments: [2009-11-25 23:38:35] s...@php.net Automatic comment from SVN on behalf of dsp Revision: http://svn.php.net/viewvc/?view=revision&revision=291323 Log: Also allow base 62 for gmp_init as requested in (#50283) [2009-11-24 21:22:40] asphp at dsgml dot com Also, gmp_init needs to be changed. [2009-11-24 20:14:52] asphp at dsgml dot com The docs and checkin comment all say till 62, but the code and the tests actually limit to 36. [2009-11-24 15:08:22] rquadl...@php.net Documented. Revision: http://svn.php.net/viewvc/?view=revision&revision=291267 [2009-11-24 13:35:06] d...@php.net The documentation states that base has to be between 2 and 36. We have to correct this, as we now allow -2 to -36 aswell. The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report online at https://bugs.php.net/bug.php?id=50283 -- Edit this bug report at https://bugs.php.net/bug.php?id=50283&edit=1
Req #33988 [Opn->Wfx]: Locales - convert a string to float
Edit report at https://bugs.php.net/bug.php?id=33988&edit=1 ID: 33988 Updated by: yohg...@php.net Reported by:tpc at klub dot chip dot pl Summary:Locales - convert a string to float -Status: Open +Status: Wont fix Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: Linux / Windows PHP Version:4.3.11 Block user comment: N Private report: N New Comment: PHP would not support locale depended number formats in arthritics. Previous Comments: [2005-08-03 22:53:13] tpc at klub dot chip dot pl Description: When I use locales and convert a string to float, I get a truncated value, eg: // comma as a decimal separator setlocale(LC_ALL,'pl_PL'); $b = "1,2"; $c = $b*2; echo $c; // php 4.2 >>> 2,4 // php >= 4.3 >>> 2 (without any warning) In php 4.2 (5.x) you used a strtod function. Now you use our zend_strtod function which ignores locales. This truncating should be reported as a warning. My suggest is: Zend/zend_strtod.c, line: 1290: if (c == ',') zend_error(E_WARNING, "Convert string to double: wrong decimal point, only . (truncate number %s)", s00); Thanks in advance for adding this feature in the new version. Expected result: Loging warning or notice if lose precision. -- Edit this bug report at https://bugs.php.net/bug.php?id=33988&edit=1
Bug #60182 [Opn->Csd]: Implicit call to spl_autoload_call() does not throw exceptions
Edit report at https://bugs.php.net/bug.php?id=60182&edit=1 ID: 60182 Updated by: yohg...@php.net Reported by:pavel dot zubkou at gmail dot com Summary:Implicit call to spl_autoload_call() does not throw exceptions -Status: Open +Status: Closed Type: Bug Package:SPL related Operating System: Debian GNU/Linux PHP Version:5.3.8 -Assigned To: +Assigned To:yohgaki Block user comment: N Private report: N New Comment: It seems fixed at least in 5.4 [yohgaki@dev PHP-5.4]$ ./php-bin -a Interactive shell php > echo 129059460 % 8640; 378000 [yohgaki@dev PHP-5.4]$ ./php-bin getMessage() . "\n"; } try { A::someMethod(); } catch (Exception $e) { echo $e->getMessage() . "\n"; } #0 autoload(A) #1 spl_autoload_call(A) called at [-:12] Test message #0 autoload(A) #1 spl_autoload_call(A) called at [-:18] Test message Previous Comments: [2011-11-07 12:09:09] francois dot dambrine at isen-lille dot fr As I run the test with 5.4b1 the result is : #0 autoload(A) #1 spl_autoload_call(A) called at [/home/isen/essai.php:12] Erreur de segmentation [2011-10-31 22:41:23] pavel dot zubkou at gmail dot com Description: Exceptions thrown from custom autoload function are not thrown outside the spl_autoload_call() function when it is implicitly called on expressions like 'A::someMethod();' and '$b = A::CONSTANT;', i.e. when accessing static methods and constants of the class. Test script: --- getMessage() . "\n"; } try { A::someMethod(); } catch (Exception $e) { echo $e->getMessage() . "\n"; } Expected result: #0 autoload(A) #1 spl_autoload_call(A) called at [/tmp/1.php:12] Test message #0 autoload(A) #1 spl_autoload_call(A) called at [/tmp/1.php:18] Test message Actual result: -- #0 autoload(A) #1 spl_autoload_call(A) called at [/tmp/1.php:12] Test message #0 autoload(A) #1 spl_autoload_call(A) called at [/tmp/1.php:18] PHP Fatal error: Class 'A' not found in /tmp/1.php on line 18 PHP Stack trace: PHP 1. {main}() /tmp/1.php:0 -- Edit this bug report at https://bugs.php.net/bug.php?id=60182&edit=1
Bug #53397 [Opn->Ver]: Modulo is not accurate with double
Edit report at https://bugs.php.net/bug.php?id=53397&edit=1 ID: 53397 Updated by: yohg...@php.net Reported by:mryaggi at hotmail dot com Summary:Modulo is not accurate with double -Status: Open +Status: Verified Type: Bug -Package:Variables related +Package:Scripting Engine problem Operating System: All PHP Version:Irrelevant Block user comment: N Private report: N New Comment: Test with 5.5.1 32bit. This issue exists still. Previous Comments: [2010-11-24 15:12:40] mryaggi at hotmail dot com Description: In C++ : 129059460 % 8640 = 3780 //correct But In PHP : 129059460 % 8640 = 30811200 //wrong However, in PHP : 129059460 - (floor(129059460/8640)*8640) = 3780 //correct! Can modulo opertor be implemented this way when left value > PHP_INT_MAX ? See Patch in attachement Test script: --- echo 129059460 % 8640; echo "\n"; echo 129059460 - (floor(129059460/8640)*8640); Expected result: 3780 3780 Actual result: -- 30811200 3780 -- Edit this bug report at https://bugs.php.net/bug.php?id=53397&edit=1
Req #47890 [Opn->Ver]: Function 'uniqid()' has a design flaw
Edit report at https://bugs.php.net/bug.php?id=47890&edit=1 ID: 47890 Updated by: yohg...@php.net Reported by:bernard dot fouche at kuantic dot com Summary:Function 'uniqid()' has a design flaw -Status: Open +Status: Verified Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: Any -PHP Version:5.2.9 +PHP Version:master Block user comment: N Private report: N New Comment: It might be good to set more_entropy set to true for next PHP. W/o combined lcg, it has too weak uniqueness. Previous Comments: [2009-04-03 16:20:50] bernard dot fouche at kuantic dot com Description: IMHO 'uniqid()' does not provide any warranty about the returned id being really 'uniq': - uniqid() just calls 'gettimeofday()' (from php-5.2.9/ext/standard/uniqid.c): there is no warranty that the underlying operating system is able to really count microseconds. May be it counts by chunks of 10us, or 1ms, etc.. - With multi-core CPU and since AFAIK gettimeofday() is not a 'real' system call any more (on Linux) but reads from shared memory updated by the kernel, chances to have more than one process or thread being able to get the same value are much higher than with older hardware and OS implementations. - NTP may update the system time and whack the microsecond counter at any time. I think that uniqid() should rely on other information, like the process ID, thread ID, and an internal counter... Sure the programmer can use the 'prefix', but then what is the purpose of this function if the programmer has to take care about these details? Reproduce code: --- --- >From manual page: function.uniqid --- -- Edit this bug report at https://bugs.php.net/bug.php?id=47890&edit=1
Req #26586 [Opn->Csd]: cross referenced objects
Edit report at https://bugs.php.net/bug.php?id=26586&edit=1 ID: 26586 Updated by: yohg...@php.net Reported by:marc dot daumas at urssaf dot fr Summary:cross referenced objects -Status: Open +Status: Closed Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: WinXP PHP Version:Irrelevant -Assigned To: +Assigned To:yohgaki Block user comment: N Private report: N New Comment: It seems working now. Previous Comments: [2012-10-27 23:52:40] will dot skates at ntlworld dot com Please Close [2003-12-10 11:45:20] marc dot daumas at urssaf dot fr Description: It becomes impossible to compare objects as soon as they have cross references. if an object A reference an object B that reference A it becomes impossible to compare objects with == or === operators. I understood your "reference" system and actualy I know that's not really a bug but an issue, but at least the === operator must deal with "already seen" objects. I just wonder if PHP5 will become a near to real object language or if that behavior will persist. Follow a basic example of Parent/Child object structure that can't be implemented in PHP4. I think that if it's impossible to work properly with instances, PHP5 will not ritch the rank of an OOP language despite all that is said about it. Best regards -Marc Reproduce code: --- name=$name; $this->children=array(); } Function addChild(&$child){ if(!isset($this->children[$child->name])){ $this->children[$child->name]=&$child; $child->setParent($this); } } } Class ChildObject { var $name; var $parent; Function &ChildObject($name){ $this->name=$name; $this->parent=null; } Function setParent(&$parent){ if($this->parent!==$parent){ if($this->parent!=null){ $this->parent->removeChild($this); } $this->parent=&$parent; if($parent!=null){ $parent->addChild($this); } } } } $p=&new ParentObject("parent"); $c=&new ChildObject("child"); $c->setParent($p); ?> Expected result: something other than Fatal error: Nesting level too deep - recursive dependency? in C:\Program Files\Apache Group\Apache2\htdocs\ngn2\test.php on line 29 Actual result: -- Fatal error: Nesting level too deep - recursive dependency? in C:\Program Files\Apache Group\Apache2\htdocs\ngn2\test.php on line 29 -- Edit this bug report at https://bugs.php.net/bug.php?id=26586&edit=1
Req #29416 [Opn->Wfx]: ob_include function
Edit report at https://bugs.php.net/bug.php?id=29416&edit=1 ID: 29416 Updated by: yohg...@php.net Reported by:edwin at cheatah dot nl Summary:ob_include function -Status: Open +Status: Wont fix Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues PHP Version:Irrelevant Block user comment: N Private report: N New Comment: $message = ob_include ( './tpl/mail.php' ); You should use $message = file_get_contents(); since it is unrelated to "output". Previous Comments: [2004-07-28 01:30:30] edwin at cheatah dot nl Description: I often use something like the following code: ob_start (); include './tpl/mail.php'; $message = ob_get_contents (); ob_end_flush (); I would very much like to see a function of control stucture like this: $message = ob_include ( './tpl/mail.php' ); // or $message = ob_include './tpl/mail.php'; Of course ob_include_once, ob_require and ob_require_once behave similarly. (Creating a custom function in PHP code is NOT possible, for the function's scope would be used.) -- Edit this bug report at https://bugs.php.net/bug.php?id=29416&edit=1
Req #14354 [Opn->Fbk]: sybase_query returns 1 regardless of delete success
Edit report at https://bugs.php.net/bug.php?id=14354&edit=1 ID: 14354 Updated by: yohg...@php.net Reported by:mheumann at sei dot cl Summary:sybase_query returns 1 regardless of delete success -Status: Open +Status: Feedback Type: Feature/Change Request Package:Sybase-ct (ctlib) related Operating System: Linux Suse 8.1 PHP Version:4.3.2RC4-dev Block user comment: N Private report: N New Comment: Please try using this snapshot: http://snaps.php.net/php5.4-latest.tar.gz For Windows: http://windows.php.net/snapshots/ Previous Comments: [2003-05-20 11:55:56] mheumann at sei dot cl Hi, the problem persists even in the current snapshot 4.3.2RC4-dev. Since I've got the feeling, this issue might not be fully understood, I've written a complete test script that will demonstrate this. Meanwhile, we have used --with-sybase-ct and are running Sybase 12.5. Here's the script (please copy and paste it and edit the database info at the beginning. The required tables and triggers are created at the start and deleted at the end.): PHP Bug Test script "; sybase_query( "create table Test_Reference ". "(". "RefId numericnot null,". "UserId numericnot null,". "DummyData varchar(255) null,". "constraint PK_TESTREF primary key (RefId, UserId)". ")", $iConnId ) or die( "Error in create table Test_Reference." ); echo "Table Test_Reference created."; sybase_query( "if exists (select 1". "from sysobjects". "where id = object_id('td_test_users')". "and type = 'TR')". " drop trigger td_test_users", $iConnId ) or die( "Error in drop trigger td_test_users." ); sybase_query( "create trigger td_test_users on Test_Users for delete as \n". " begin\n". "declare\n". " @numrows int,\n". " @errnoint,\n". " @errmsg varchar(255)\n". "\n". "select @numrows = @@rowcount\n". "if @numrows = 0\n". " return\n". "/* Cannot delete parent Test_Users if children still exist in Test_Reference */\n". "if exists (select 1\n". " from Test_Reference t2, deleted t1\n". " where t2.UserId = t1.UserId)\n". " begin\n". " select @errno = 30006,\n". " @errmsg = 'Children still exist in Test_Reference. Cannot delete parent Test_Users.'\n". " goto error\n". " end\n". "\n". "return\n". "\n". "/* Errors handling */\n". "error:\n". "raiserror @errno @errmsg\n". "rollback transaction\n". "end\n", $iConnId ) or die( "Error in create trigger td_test_users." ); echo "Trigger td_test_users created."; echo "Insert new user test_noref:"; $Result = sybase_query( "insert into Test_Users (Username,Password)". " values ('test_noref','test_noref')", $iConnId ) or die( "Error inserting test_noref." ); if( $Result ) echo "...Success - "; $Result = sybase_query( "SELECT LastId=@@identity", $iConnId ); $Row = sybase_fetch_array( $Result ); $iLastId = $Row["LastId"]; sybase_free_result( $Result ); echo "ID is $iLastId"; echo "Insert new user test_with_ref:"; $Result = sybase_query( "insert into Test_Users (Username,Password)". " values ('test_with_ref','test_with_ref')", $iConnId ) or die( "Error inserting test_with_ref." ); if( $Result ) echo "...Success - "; $Result = sybase_query( "SELECT LastId=@@identity", $iConnId ); $Row = sybase_fetch_array( $Result ); $iLastId2 = $Row["LastId"]; sybase_free_result( $Result ); echo "ID is $iLastId2"; echo "Insert reference for user test_with_ref:"; $Result = sybase_query( "insert into Test_Reference (RefId, UserId, DummyData)". " values (1,$iLastId2,'This is the reference')", $iConnId ) or die( "Error inserting reference." ); if( $Result ) echo "...Success."; $Result = sybase_query( "select UserId, Username, Password from Test_Users", $iConnId ) or die( "Error selecting from Test_Users." ); echo "Test_Users:"; echo "UserIdUsernamePassword"; while( $Row = sybase_fetch_array( $Result ) ) { echo "".$Row["UserId"]."";
Req #42196 [Opn->Csd]: unserialize artichecture problem
Edit report at https://bugs.php.net/bug.php?id=42196&edit=1 ID: 42196 Updated by: yohg...@php.net Reported by:joungkyun at gmail dot com Summary:unserialize artichecture problem -Status: Open +Status: Closed Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: Linux PHP Version:5CVS-2007-08-03 (CVS) -Assigned To: +Assigned To:yohgaki Block user comment: N Private report: N New Comment: Already implemented at least 5.4 Previous Comments: [2007-08-04 16:42:10] joungkyun at gmail dot com I thinks that php has auto casting type. So, on 32bit systems php get 64bit interger, first convert casting type to double. sorry for my pool english. [2007-08-04 14:08:04] j...@php.net First of all, exactly why do you pass data like this between machines? Secondly: OF course you can't pass huge integers to 32bit system. So this is expected behaviour. What do you suggest happens in this case? Try this on 32bit system: echo (int) 4294967297; [2007-08-03 09:51:46] joungkyun at gmail dot com Description: There is serialized data that made with 64bit integer on 64 bit system. On 32bit machine, if unserialize this data, you have result of integer overflow. Reproduce code: --- On, 64bit system echo serialize (4294967296 + 1); => result : "i:4294967297;" On, 32bit system echo unserilaize ("i:4294967297;"); => result : 1 Expected result: On, 32bit system echo unserilaize ("i:4294967297;"); => result : 4294967297 Actual result: -- On, 32bit system echo unserilaize ("i:4294967297;"); => result : 1 -- Edit this bug report at https://bugs.php.net/bug.php?id=42196&edit=1
[PHP-BUG] Bug #65399 [NEW]: Parse error when trying to access class constant from another class property
From: filippovdaniil at gmail dot com Operating system: Debian Wheezy 7.0 PHP version: 5.5.1 Package: Scripting Engine problem Bug Type: Bug Bug description:Parse error when trying to access class constant from another class property Description: Workaround: $obj = $b->obj; echo $obj::foo; Test script: --- obj = new a; } } $b = new b; echo $b->obj::foo; // Parse error! Expected result: test Actual result: -- Parse error: syntax error, unexpected '::' (T_PAAMAYIM_NEKUDOTAYIM) -- Edit bug report at https://bugs.php.net/bug.php?id=65399&edit=1 -- Try a snapshot (PHP 5.4): https://bugs.php.net/fix.php?id=65399&r=trysnapshot54 Try a snapshot (PHP 5.3): https://bugs.php.net/fix.php?id=65399&r=trysnapshot53 Try a snapshot (trunk): https://bugs.php.net/fix.php?id=65399&r=trysnapshottrunk Fixed in SVN: https://bugs.php.net/fix.php?id=65399&r=fixed Fixed in release: https://bugs.php.net/fix.php?id=65399&r=alreadyfixed Need backtrace: https://bugs.php.net/fix.php?id=65399&r=needtrace Need Reproduce Script: https://bugs.php.net/fix.php?id=65399&r=needscript Try newer version: https://bugs.php.net/fix.php?id=65399&r=oldversion Not developer issue:https://bugs.php.net/fix.php?id=65399&r=support Expected behavior: https://bugs.php.net/fix.php?id=65399&r=notwrong Not enough info: https://bugs.php.net/fix.php?id=65399&r=notenoughinfo Submitted twice: https://bugs.php.net/fix.php?id=65399&r=submittedtwice register_globals: https://bugs.php.net/fix.php?id=65399&r=globals PHP 4 support discontinued: https://bugs.php.net/fix.php?id=65399&r=php4 Daylight Savings: https://bugs.php.net/fix.php?id=65399&r=dst IIS Stability: https://bugs.php.net/fix.php?id=65399&r=isapi Install GNU Sed:https://bugs.php.net/fix.php?id=65399&r=gnused Floating point limitations: https://bugs.php.net/fix.php?id=65399&r=float No Zend Extensions: https://bugs.php.net/fix.php?id=65399&r=nozend MySQL Configuration Error: https://bugs.php.net/fix.php?id=65399&r=mysqlcfg
Req #35289 [Opn->Csd]: Win32 symlink unimplemented - could be easily
Edit report at https://bugs.php.net/bug.php?id=35289&edit=1 ID: 35289 Updated by: yohg...@php.net Reported by:scott dot pascoe at gmail dot com Summary:Win32 symlink unimplemented - could be easily -Status: Open +Status: Closed Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: Windows XP PHP Version:5.0.5 -Assigned To: +Assigned To:yohgaki Block user comment: N Private report: N New Comment: Implemented since 5.3.0. http://www.php.net/manual/en/function.symlink.php Previous Comments: [2005-11-19 05:39:38] scott dot pascoe at gmail dot com Description: It is noted in the documentation that 'symlink' is unimplemented on the Win32 platform. It could be easily implemented for Windows 2000 and above OS's. I found that PostgresSQL implements them in the file src/port/dirmod.c. They refer to a codeproject article: http://www.codeproject.com/W2K/junctionpoints.asp as a reference. Reproduce code: --- mkdir('c:\test'); symlink('c:\junk', 'c:\test'); Expected result: it would work on the Windows Platform Actual result: -- error unimplemented function -- Edit this bug report at https://bugs.php.net/bug.php?id=35289&edit=1
Bug #65367 [ReO->Csd]: Segmentation fault when mixing =& and =
Edit report at https://bugs.php.net/bug.php?id=65367&edit=1 ID: 65367 Updated by: larue...@php.net Reported by:mbecc...@php.net Summary:Segmentation fault when mixing =& and = -Status: Re-Opened +Status: Closed Type: Bug Package:Reproducible crash Operating System: Any PHP Version:5.5.1 Assigned To:laruence Block user comment: N Private report: N New Comment: thanks for the ssh access, it is helpful. fixed in: http://git.php.net/?p=php- src.git;a=commit;h=a831499b4a1029118dc45375e62af42043110ade Previous Comments: [2013-08-06 05:53:18] mbecc...@php.net Yes, I've used a fresh git clone. [2013-08-06 03:02:53] larue...@php.net did you built it frome a fresh work dir? [2013-08-05 14:50:51] mbecc...@php.net I have upgraded PHP 5.4 to latest-git on a new machine. With the patch applied I now see many test runs consistently fail with a segafult. Reverting to 5.4.17 fixes the segfault. Backtrace is: #0 0x009beb33 in zend_std_object_get_class (object=0x7fffef535cd0) at /root/compile/php-src/Zend/zend_object_handlers.c:1500 zobj = 0x7fff0021 #1 0x0098dd98 in zend_get_class_entry (zobject=0x7fffef535cd0) at /root/compile/php-src/Zend/zend_API.c:238 No locals. #2 0x00a17121 in ZEND_INIT_METHOD_CALL_SPEC_CV_CONST_HANDLER (execute_data=0x77fa1ea0) at /root/compile/php-src/Zend/zend_vm_execute.h:29282 opline = 0x70a34228 function_name = 0x70a35058 function_name_strval = 0x77f97d50 "setFileNameProtection" function_name_strlen = 21 #3 0x009c6513 in execute (op_array=0x1446f00) at /root/compile/php-src/Zend/zend_vm_execute.h:410 ret = 0 execute_data = 0x77fa1ea0 nested = 1 '\001' original_in_execution = 0 '\000' #4 0x0098ca9f in zend_execute_scripts (type=8, retval=0x0, file_count=3) at /root/compile/php-src/Zend/zend.c:1315 files = {{gp_offset = 40, fp_offset = 48, overflow_arg_area = 0x7fffae40, reg_save_area = 0x7fffad80}} i = 1 file_handle = 0x7fffd1e0 orig_op_array = 0x0 orig_retval_ptr_ptr = 0x0 orig_interactive = 0 #5 0x00902ff4 in php_execute_script (primary_file=0x7fffd1e0) at /root/compile/php-src/main/main.c:2497 realfile = "/home/atlassian/bamboo/xml-data/build-dir/AP-RET-P53P/tests/run.php\000\000\000\000\000\021", '\000' , "P\301\377\377\377\177\000\000\336U\225\000\000\000\000\000\234\066\336\367\377\177\000\000\000\020$\001\000\000\000\000\016\000\000\000\000\000\000\000\260\302\377\377\377\177\000\000-\000\000\000\000\000\000\000fII\"\000\000\000\000\240>\336\367\377\177\000\000\000\000\000\000\000\000\000\000&\000\000\000\000\000\000\000%%\211\000\000\000\000\000\030\255\231\365\377\177\000\000\214\236\231\365\377\177\000\000"... __orig_bailout = 0x7fffd2f0 __bailout = {{__jmpbuf = {0, -26362260470167, 4380576, 140737488348720, 0, 0, -263622602725482883, 263621642691976829}, __mask_was_saved = 0, __saved_mask = {__val = {0, 0, 140737314399616, 140737488343184, 0, 140737488343200, 4380576, 140737488348720, 0, 0, 9431409, 140737488344000, 140737488349319, 19186208, 287762808856, 21253568 prepend_file_p = 0x0 append_file_p = 0x0 prepend_file = {type = ZEND_HANDLE_FILENAME, filename = 0x0, opened_path = 0x0, handle = {fd = 0, fp = 0x0, stream = {handle = 0x0, isatty = 0, mmap = {len = 0, pos = 0, map = 0x0, buf = 0x0, old_handle = 0x0, old_closer = 0x0}, reader = 0x0, fsizer = 0x0, closer = 0x0}}, free_filename = 0 '\000'} append_file = {type = ZEND_HANDLE_FILENAME, filename = 0x0, opened_path = 0x0, handle = {fd = 0, fp = 0x0, stream = {handle = 0x0, isatty = 0, mmap = {len = 0, pos = 0, map = 0x0, buf = 0x0, old_handle = 0x0, old_closer = 0x0}, reader = 0x0, fsizer = 0x0, closer = 0x0}}, free_filename = 0 '\000'} old_cwd = 0x7fffae60 "" use_heap = 0 '\000' retval = 0 [2013-08-02 16:24:26] larue...@php.net fixed in http://git.php.net/?p=php- src.git;a=commitdiff;h=ce9169e360701ea3b1ab2366171c24d4de5e78e3 [2013-08-02 07:29:59] mbecc...@php.net Yes, the patch fixes the issue as far as I can tell. Well done! The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report onli
Req #64460 [Com]: ob_start implicit flush content
Edit report at https://bugs.php.net/bug.php?id=64460&edit=1 ID: 64460 Comment by: valentiny510 at yahoo dot es Reported by:valentiny510 at yahoo dot es Summary:ob_start implicit flush content Status: Not a bug Type: Feature/Change Request Package:Output Control PHP Version:Irrelevant Assigned To:mike Block user comment: N Private report: N New Comment: wait.. mike.. what .. ?? Previous Comments: [2013-07-30 19:16:53] m...@php.net Thank you for taking the time to write to us, but this is not a bug. Please double-check the documentation available at http://www.php.net/manual/ and the instructions on how to report a bug at http://bugs.php.net/how-to-report.php Quote: "While output buffering is active..." Output buffering will be disabled at request shutdown of course. [2013-03-20 01:09:19] valentiny510 at yahoo dot es P.S. I found something related Request #23877 "ob_implicit_flush does not work". Damn, a 10 years bug/request again... Rasmus I think is time to upgrade the team not the documentation :D. Just joking.. But seriously guys, you cannot allow to happend thinks like this, I perfectly understand is a lot of work sometimes but 10 years... is something.. too much I remember times when some of my clients calls me at midnight to fix their sites... what a happy days.. [2013-03-20 00:48:54] valentiny510 at yahoo dot es Description: I write this like a Request but I think is more a bug. According to documentation of ob_start: "While output buffering is active no output is sent from the script (other than headers), instead the output is stored in an internal buffer." but is not true. implicit_flush = Off or ob_implicit_flush( false ) are also useless. Or change the documentation or change the function behaviour to act according to the documentation. When somebody active the buffering it expect to see the blank page until explicit call some flush functions right ? At least this said the docs. Test script: --- https://bugs.php.net/bug.php?id=64460&edit=1