Bug #63136 [Opn->Nab]: pow() returns NaN where base is negative and exponent is an odd number root
Edit report at https://bugs.php.net/bug.php?id=63136&edit=1 ID: 63136 Updated by: ni...@php.net Reported by:tstobbe at sacbee dot com Summary:pow() returns NaN where base is negative and exponent is an odd number root -Status: Open +Status: Not a bug Type: Bug Package:Math related Operating System: Linux (fc17, 32-bit) PHP Version:5.4.7 Block user comment: N Private report: N New Comment: Real-exponent powers are not defined for negative numbers, see http://en.wikipedia.org/wiki/Exponentiation#Real_exponents_with_negative_bases. You can see this very easily if you remember that a^b is defined as exp(b*ln(a)) and the logarithm is only defined in the range ]0,inf[. Previous Comments: [2012-09-22 02:47:21] tstobbe at sacbee dot com It should be noted that all subsequent fractional pow() issues are related to this. For example: pow(-[any number], 1/5) vs pow(-[any number], 2/5). Likewise I suspect that most other exponent based functions (such as log with an arbitrary base) are probably suspect, but this hasn't been verified. [2012-09-21 22:38:15] tstobbe at sacbee dot com Description: Odd roots (3, 5, etc) of negative numbers passed to pow() result in NaN rather than the actual result. While this makes sense for even numbered roots (sqrt and what-not), it wrecks havoc on any system that actually uses arbitrary roots of negative numbers. Example: -2*-2*-2 = -8; pow(-8, 1/3) should return -2 Test script: --- var_dump(pow(-8, 1/3)); Expected result: float(-2) Actual result: -- float(NAN) -- Edit this bug report at https://bugs.php.net/bug.php?id=63136&edit=1
[PHP-BUG] Req #63137 [NEW]: Allow custom GLOBAL variables
From: valentiny510 at yahoo dot es
Operating system: XP
PHP version: 5.4.7
Package: Variables related
Bug Type: Feature/Change Request
Bug description:Allow custom GLOBAL variables
Description:
I think can be very usefull, sometimes, to be able to make a custom global
variable, like languages, configs, etc.. and working just like $GLOBALS but
only with the value/values assigned. I'm sure more than 75%-80% of the Php
users will be agreed with this change.
Imagine and array with some translation strings:
$lang = array(
'bla' => 'some nonsense talk',
'blabla' => 'More nonsense talks');
and now the 'trick'.. some function to make 'global' the $lang:
make_global( $lang );
or even set in the php.ini some configuration like:
custom_globals = "lang, config, more, etc, bla, bla"
and now we can use it on global scope
function test() { return $lang['bla']; }
is a 'nicer' way than
function test() { return $GLOBALS['lang']['bla']; }
or
function test() { global $lang; return $lang['bla']; }
Test script:
---
-- none --
Expected result:
-- none --
Actual result:
--
-- none --
--
Edit bug report at https://bugs.php.net/bug.php?id=63137&edit=1
--
Try a snapshot (PHP 5.4):
https://bugs.php.net/fix.php?id=63137&r=trysnapshot54
Try a snapshot (PHP 5.3):
https://bugs.php.net/fix.php?id=63137&r=trysnapshot53
Try a snapshot (trunk):
https://bugs.php.net/fix.php?id=63137&r=trysnapshottrunk
Fixed in SVN: https://bugs.php.net/fix.php?id=63137&r=fixed
Fixed in release: https://bugs.php.net/fix.php?id=63137&r=alreadyfixed
Need backtrace: https://bugs.php.net/fix.php?id=63137&r=needtrace
Need Reproduce Script: https://bugs.php.net/fix.php?id=63137&r=needscript
Try newer version: https://bugs.php.net/fix.php?id=63137&r=oldversion
Not developer issue:https://bugs.php.net/fix.php?id=63137&r=support
Expected behavior: https://bugs.php.net/fix.php?id=63137&r=notwrong
Not enough info:
https://bugs.php.net/fix.php?id=63137&r=notenoughinfo
Submitted twice:
https://bugs.php.net/fix.php?id=63137&r=submittedtwice
register_globals: https://bugs.php.net/fix.php?id=63137&r=globals
PHP 4 support discontinued: https://bugs.php.net/fix.php?id=63137&r=php4
Daylight Savings: https://bugs.php.net/fix.php?id=63137&r=dst
IIS Stability: https://bugs.php.net/fix.php?id=63137&r=isapi
Install GNU Sed:https://bugs.php.net/fix.php?id=63137&r=gnused
Floating point limitations: https://bugs.php.net/fix.php?id=63137&r=float
No Zend Extensions: https://bugs.php.net/fix.php?id=63137&r=nozend
MySQL Configuration Error: https://bugs.php.net/fix.php?id=63137&r=mysqlcfg
Req #39150 [Opn->Nab]: PHP allows invalid names for constants and variables
Edit report at https://bugs.php.net/bug.php?id=39150&edit=1
ID: 39150
Updated by: ni...@php.net
Reported by:z_rules55 at hotmail dot com
Summary:PHP allows invalid names for constants and variables
-Status: Open
+Status: Not a bug
Type: Feature/Change Request
-Package:Feature/Change Request
+Package:*General Issues
Operating System: WinXP
PHP Version:5.1.6
Block user comment: N
Private report: N
New Comment:
PHP allows arbitrary variable names using the ${'string'} syntax. The
restrictions only apply to the $name syntax. The same applies to constants and
various other names in PHP, which have a "plain" access method and an access
method based on a string.
Previous Comments:
[2006-10-16 19:40:20] z_rules55 at hotmail dot com
Changing the status of this back to Open because I think it deserves further
consideration.
[2006-10-13 19:12:22] z_rules55 at hotmail dot com
I checked the docs, and I don't see anyone else having submitted the same bug
to the bug database.
If this is not a bug, maybe I should change it to feature request? Why would
PHP allow the creation of variables and constants that can only be accessed
with a work-around instead of directly, as is most natural and intuitive? You
say "read the variable", not "access the superglobal array with the associative
index equal to the variable's name". Likewise, why would you have to call a
function to read a constant that you already know the name of?
Or, change the docs so they make this fact more clear: that you *can* in fact
create variables and constants with supposedly invalid names, but you'll have a
harder time using them later on.
[2006-10-13 15:52:01] tony2...@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
[2006-10-13 15:47:02] z_rules55 at hotmail dot com
Description:
The PHP manual states that:
A valid variable name starts with a letter or
underscore, followed by any number of letters,
numbers, or underscores.
And that constants must follow the same naming rules as variables. However, PHP
allows me to define variables and constants with an invalid character like a
hyphen or a dollar sign, as well as starting with a number. Those are the only
three invalid names I tested; I assume others could also be used.
Reproduce code:
---
Expected result:
Some kind of error upon defining the constant or variable with the invalid name.
Actual result:
--
PHP happily accepts the define() with the invalid name, as well as the
assignment to the variable variable with the invalid name. Using constant() and
$GLOBALS to read the value of the invalid constant and variable work fine.
Echoing the value of my-definition gives the usual warning of:
Notice: Use of undefined constant my - assumed 'my' in
C:\path\to\script.php on line 14
Notice: Use of undefined constant definition -
assumed 'definition' in C:\path\to\script.php
on line 14
Only when you try to access the invalid constant or variable directly does PHP
throw parse errors.
--
Edit this bug report at https://bugs.php.net/bug.php?id=39150&edit=1
Req #16979 [Opn->Csd]: nested function in functions within a class
Edit report at https://bugs.php.net/bug.php?id=16979&edit=1
ID: 16979
Updated by: ni...@php.net
Reported by:lazybrain at rcn dot com
Summary:nested function in functions within a class
-Status: Open
+Status: Closed
Type: Feature/Change Request
-Package:Feature/Change Request
+Package:*General Issues
Operating System: bsd
PHP Version:4.1.2
-Assigned To:
+Assigned To:nikic
Block user comment: N
Private report: N
New Comment:
Closing as this already works.
Previous Comments:
[2002-08-12 13:57:31] ahristov at icygen dot com
AFAIK this code could work. I've written some code this way but the user has to
guard the definition of the function somehow or on the second call os test1
there will be fatal error.
[2002-05-02 18:29:41] lazybrain at rcn dot com
It would be helpfull to be able to
use nested functions inside a function thats in a class.
For example...
clase something {
function test1() {
do something here
// quick function to make code smaller
function test2() {
do more here
}
}
}
--
Edit this bug report at https://bugs.php.net/bug.php?id=16979&edit=1
Req #18187 [Opn->Csd]: Consequent implementation of array referencing to user method
Edit report at https://bugs.php.net/bug.php?id=18187&edit=1
ID: 18187
Updated by: ni...@php.net
Reported by:bram at totalgsm dot net
Summary:Consequent implementation of array referencing to
user method
-Status: Open
+Status: Closed
Type: Feature/Change Request
-Package:Feature/Change Request
+Package:*General Issues
Operating System: any
PHP Version:4.2.1
-Assigned To:
+Assigned To:nikic
Block user comment: N
Private report: N
New Comment:
Closing as the $callable() syntax now works for array-callbacks too (as of PHP
5.4). Instead of function_exists the function is_callable can be used.
Previous Comments:
[2002-07-05 11:46:28] bram at totalgsm dot net
You support calling user methods using an array in the call_user_func()
function. But related functions of call_user_func() don't support the syntax. I
would like to see that function_exists could check if the user method exists.
The following example code should work:
class Foo {
function bar($what = 'world') {
print 'Hello ' . $what;
}
}
$a_foo = new Foo();
$user_method = array($a_foo, 'bar');
if (!function_exists($user_method))
die('You shouldn\'t see this :)');
$user_method('PHP'); // Should print "Hello PHP" ofcourse.
Furthermore the implementation of the user_method calling in call_user_func is
incorrect for referencing variabeles, which might be pretty usefull. (see
http://bugs.php.net/bug.php?id=17309)
--
Edit this bug report at https://bugs.php.net/bug.php?id=18187&edit=1
Req #18407 [Opn->Csd]: Request of new configuration directive for safe_mode
Edit report at https://bugs.php.net/bug.php?id=18407&edit=1 ID: 18407 Updated by: ni...@php.net Reported by:msteinacher at websource dot ch Summary:Request of new configuration directive for safe_mode -Status: Open +Status: Closed Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: Linux PHP Version:4.2.1 -Assigned To: +Assigned To:nikic Block user comment: N Private report: N New Comment: Closing this as safe mode is no longer supported as of PHP 5.4 Previous Comments: [2002-07-18 05:54:21] msteinacher at websource dot ch I suggest to add a new configuration directive for safe mode to solve the common problem, that a user can't change uploaded files because they have the UID of the user owning the webserver process and not the UID of the user owning the script. I know that I could turn safe_mode off and use only open_basedir instead. Another 'solution' sould be to use safe_mode_gid and put the webserver-user in the same group as the script-owners. But I don't want to do this. Thus my suggestion to add a new directive that could for example be called 'safe_mode_allow_proc_uid'. If this is set to TRUE then PHP should allow the access to files which are owned by the user that owns the script (as it does now) _OR_ files which are owned by the user that owns the webserver process. In other words: Try to implement the patch at http://www.zend.com/lists/php-dev/200201/msg01149.html with the option to enable or disable it. Thanks for listening. Marco -- Edit this bug report at https://bugs.php.net/bug.php?id=18407&edit=1
Req #19236 [Opn->Csd]: Need $php_errorLineNumber $php_errorScript in addition to $php_errormsg
Edit report at https://bugs.php.net/bug.php?id=19236&edit=1 ID: 19236 Updated by: ni...@php.net Reported by:php-bugs at ketan dot org Summary:Need $php_errorLineNumber $php_errorScript in addition to $php_errormsg -Status: Open +Status: Closed Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: FreeBSD PHP Version:4.2.2 -Assigned To: +Assigned To:nikic Block user comment: N Private report: N New Comment: Closing as error_get_last() provides this information as of PHP 5.2. Previous Comments: [2006-04-25 13:10:02] jason_priebe at yahoo dot com I get the point of an error_handler, but this doesn't help when you're eval-ing code that contains a parse error. The parse error doesn't trigger the error_handler. $php_errormsg contains the actual error string, but doesn't tell you what line of the eval-ed code contained the error. The httpd error log contains the offending line number, but AFAICT, I can't get to it inside my PHP code. [2002-09-04 18:13:35] php-bugs at ketan dot org It is half done the way it is now. If we have $php_errormsg, then we should have the other components of the standard php error, the line number and the script name. An error message in isolation is not very useful. [2002-09-04 17:11:50] w...@php.net That's too much magic; you can do this yourself using set_error_handler and having your error handler set up these global vars yourself. [2002-09-04 16:49:35] php-bugs at ketan dot org When using ini configuration directive track_errors = On, $php_errormsg contains the last error message. We need a corresponding variable $php_errorLineNumber (or whatever) that keeps track of where in the file the error occurred, and $php_errorScriptName to tell in which file the error occurred (in case of include()). -- Edit this bug report at https://bugs.php.net/bug.php?id=19236&edit=1
Req #18407 [Csd->Wfx]: Request of new configuration directive for safe_mode
Edit report at https://bugs.php.net/bug.php?id=18407&edit=1 ID: 18407 Updated by: ni...@php.net Reported by:msteinacher at websource dot ch Summary:Request of new configuration directive for safe_mode -Status: Closed +Status: Wont fix Type: Feature/Change Request Package:*General Issues Operating System: Linux PHP Version:4.2.1 Assigned To:nikic Block user comment: N Private report: N Previous Comments: [2012-09-22 09:18:58] ni...@php.net Closing this as safe mode is no longer supported as of PHP 5.4 [2002-07-18 05:54:21] msteinacher at websource dot ch I suggest to add a new configuration directive for safe mode to solve the common problem, that a user can't change uploaded files because they have the UID of the user owning the webserver process and not the UID of the user owning the script. I know that I could turn safe_mode off and use only open_basedir instead. Another 'solution' sould be to use safe_mode_gid and put the webserver-user in the same group as the script-owners. But I don't want to do this. Thus my suggestion to add a new directive that could for example be called 'safe_mode_allow_proc_uid'. If this is set to TRUE then PHP should allow the access to files which are owned by the user that owns the script (as it does now) _OR_ files which are owned by the user that owns the webserver process. In other words: Try to implement the patch at http://www.zend.com/lists/php-dev/200201/msg01149.html with the option to enable or disable it. Thanks for listening. Marco -- Edit this bug report at https://bugs.php.net/bug.php?id=18407&edit=1
Req #61222 [Opn->Dup]: For... else construct
Edit report at https://bugs.php.net/bug.php?id=61222&edit=1
ID: 61222
Updated by: ni...@php.net
Reported by:jeroenvandenenden at gmail dot com
Summary:For... else construct
-Status: Open
+Status: Duplicate
Type: Feature/Change Request
Package:Arrays related
PHP Version:Irrelevant
Block user comment: N
Private report: N
New Comment:
Closing as duplicate of https://bugs.php.net/bug.php?id=26411.
Previous Comments:
[2012-03-25 17:57:54] phristen at yahoo dot com
There are already bugs for foreach { } else { } and while { } else { }, both
highly rated.
I think this should be all consolidated into a single bug or an RFC, or it will
never get done.
[2012-03-01 15:42:43] jeroenvandenenden at gmail dot com
Of course instead of
for ($val in $array) {
I meant
foreach ($array as $val) {
[2012-03-01 15:27:36] jeroenvandenenden at gmail dot com
Description:
It would be nice to have a for...else construct like you see in other
languages.
This makes the code more compact and readable and it avoids the necessity to
combine an if and a for loop.
Test script:
---
for ($val in $array) {
echo $val.'';
} else {
echo 'No values found';
}
Expected result:
In the case of an empty array, the else part would be called. Without this
construct the code would be as follows.
if (count($array) > 0) {
for ($val in $array) {
echo $val.'';
}
} else {
echo 'No values found';
}
--
Edit this bug report at https://bugs.php.net/bug.php?id=61222&edit=1
Req #21555 [Opn->Dup]: Possible extensions to alternative control syntax
Edit report at https://bugs.php.net/bug.php?id=21555&edit=1 ID: 21555 Updated by: ni...@php.net Reported by:gabe at bravenet dot com Summary:Possible extensions to alternative control syntax -Status: Open +Status: Duplicate Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: All PHP Version:4.3.0 Block user comment: N Private report: N New Comment: Closing as duplicate of https://bugs.php.net/bug.php?id=26411. (This bug is older, but the other one has more discussion.) Previous Comments: [2003-01-09 15:48:14] gabe at bravenet dot com The alternative control structure syntax (see http://php.benscom.com/manual/en/control-structures.alternative-syntax.php) is very useful for templating purposes. However, I miss the smarty-style 'else' clauses: foreachelse, whileelse etc. I would like to be able to do this: // if array has items, this is repeated once per item // we get here if $array has no items, then exit ...or this... // repeated while condition is met // if condition is never met we go here once and exit. I think these additions would be fairly useful. -- Edit this bug report at https://bugs.php.net/bug.php?id=21555&edit=1
Req #40218 [Opn->Dup]: Add Else clause to while
Edit report at https://bugs.php.net/bug.php?id=40218&edit=1
ID: 40218
Updated by: ni...@php.net
Reported by:jbailey at raspberryginger dot com
Summary:Add Else clause to while
-Status: Open
+Status: Duplicate
Type: Feature/Change Request
-Package:Feature/Change Request
+Package:*General Issues
Operating System: Linux
PHP Version:5.2.0
Block user comment: N
Private report: N
New Comment:
Closing as duplicate of https://bugs.php.net/bug.php?id=26411.
Previous Comments:
[2007-01-24 15:08:50] didou at keliglia dot com
I like the idea, but using the "else" keyword would break a lot of apps using
something like this:
if (X)
while (foo) {
}
else
no X
whileelse ?
[2007-01-24 14:31:11] jbailey at raspberryginger dot com
Grr.. This bug tracking system ate my comment. So here's a much shorter
version of what I wrote:
Use cases:
Retrieving from a database (especially since PDO doesn't have a numrows
options):
while ($foo = $db->sql_fetchrow($result)) {
// Do something
} else {
// No rows found
}
This could also be useful for ArrayObjects and foreach:
$i = $myArray->getIterator();
foreach ($i as $item) {
// Do something
} else {
// Empty array
}
[2007-01-24 14:18:57] jbailey at raspberryginger dot com
Description:
Right now checks have to be done separately for when a while will execute zero
times, which means code duplication and constructs that have to be kept in
sync. I would like to see an else clause to while that is executed when the
while loop goes zero times.
Reproduce code:
---
while (0) {
echo "I ran";
} else {
echo "I didn't run.";
}
Expected result:
I didn't run.
Actual result:
--
Parse error: syntax error, unexpected T_ELSE in /tmp/test.php on line 5
--
Edit this bug report at https://bugs.php.net/bug.php?id=40218&edit=1
Req #27612 [Opn->Dup]: "else" statement for "foreach" construct
Edit report at https://bugs.php.net/bug.php?id=27612&edit=1
ID: 27612
Updated by: ni...@php.net
Reported by:camka at email dot ee
Summary:"else" statement for "foreach" construct
-Status: Open
+Status: Duplicate
Type: Feature/Change Request
-Package:Feature/Change Request
+Package:*General Issues
Operating System: Any
PHP Version:Irrelevant
Block user comment: N
Private report: N
New Comment:
Closing as duplicate of https://bugs.php.net/bug.php?id=26411.
Previous Comments:
[2004-03-16 04:05:39] camka at email dot ee
Description:
"else" statement would be very useful and comfortable when the array_expression
in foreach construct is empty. There'd be no need to calculate count(array)
every time you need to execute some code in case of empty array.
Smarty has such syntax and it really helps.
Reproduce code:
---
$count = count($array);
foreach($array as $arr)
{
...
}
if(!$count)
{
...
}
===
foreach($array as $arr)
{
...
}
else //may be [foreachelse] or [elseforeach]
{
...
}
--
Edit this bug report at https://bugs.php.net/bug.php?id=27612&edit=1
Req #46240 [Opn->Dup]: Build in foreach else support
Edit report at https://bugs.php.net/bug.php?id=46240&edit=1
ID: 46240
Updated by: ni...@php.net
Reported by:kjarli at gmail dot com
Summary:Build in foreach else support
-Status: Open
+Status: Duplicate
Type: Feature/Change Request
Package:Scripting Engine problem
Operating System: *
PHP Version:5.2.6
Block user comment: N
Private report: N
New Comment:
Closing as duplicate of https://bugs.php.net/bug.php?id=26411.
Previous Comments:
[2011-12-22 20:41:46] jason at valdron dot ca
I completely disagree about the onFail section.
The foreach else would be useful if there is no item. As a shortcut to
if(count($elements) == 0).
[2011-07-12 12:13:44] dinumarina at yahoo dot com
Opposing general consensus, I think adding such feature would be semantically
ambiguous. The foreach function does not fail, it has types it can handle
(assoc
array, object) and that it can't handle (scalar). A function can always return
a
result that may be valid for foreach but not the expected format, so I think
data sanitizing is best done dilligently. This is just a lazy-man's hack for a
non-issue. For the function completion status (onFail), there is already such a
construct: try {throw()} catch(){} which already supersets the imagined onFail
implementation. Only thing I would hindsight: it would have been SOOO nice if
php had a false/null/na result convention, as well as an error/exception
convention and it would actually stick by it. Each module signals completion,
computability, singular cases and errors as it well pleases.
[2011-02-18 01:20:24] ijrbiz at gmail dot com
Highly agreed with adding a foreach :: else statement, this request would be
very
practical for improved coding structure and follows logical language syntax
nicely.
foreach ($items as $item) {
echo $item; // Manage each item, ...
} else {
echo 'No items present.'; // Manage no items found, ...
}
[2011-02-04 02:30:00] pedro at worcel dot com
foreachelse seems ok to me. Onfail is... weird. :)
[2010-12-20 13:40:54] rick dot sketchy at gmail dot com
I have to agree with the OP. foreachelse (or something similar) is really
needed. I do like the suggestion posted by cerlestes at googlemail dot com:
foreach($arr as $var)
{
doCode();
}
onFail
{
failHandling();
}
A fail handler would prove useful,however I can see it may have some
limitations, in which case an else option on the foreach would be satisfactory.
Whilst we're at it, it may as well be added to while statements too.
The remainder of the comments for this report are too long. To view
the rest of the comments, please view the bug report online at
https://bugs.php.net/bug.php?id=46240
--
Edit this bug report at https://bugs.php.net/bug.php?id=46240&edit=1
Req #21913 [Opn->Csd]: formatting for sci notation
Edit report at https://bugs.php.net/bug.php?id=21913&edit=1
ID: 21913
Updated by: ni...@php.net
Reported by:sprice at wisc dot edu
Summary:formatting for sci notation
-Status: Open
+Status: Closed
Type: Feature/Change Request
-Package:Feature/Change Request
+Package:*General Issues
Operating System: Mac OS X.2.3
PHP Version:4.3.0
-Assigned To:
+Assigned To:nikic
Block user comment: N
Private report: N
New Comment:
Closing as scientific notation is already available via %e, %E (and %g, %G for
mixed).
Previous Comments:
[2003-01-27 19:30:52] sprice at wisc dot edu
I humbly request that scientific notation formatting (ex. 7.16e-11) is added to
the function sprintf(). Currently I am using a function supplied by another
user:
function sci($x, $d=-1) {
$e=floor(log10($x));
$x*=pow(10,-$e);
$fmt=($d>=0)?".".$d:"";
$e=($e>=0)?"+".sprintf("%02d",$e):"-".sprintf("%02d",-$e);
return sprintf("%".$fmt."fe%s",$x,$e);
}
and can be found in the comments of this page:
http://www.php.net/manual/en/function.sprintf.php
This function seems to be working fine for the time being, but this seems like
somethine many other people would use and an addition to PHP itself would be
better.
--
Edit this bug report at https://bugs.php.net/bug.php?id=21913&edit=1
Bug #61141 [Com]: curl_multi_select returns -1
Edit report at https://bugs.php.net/bug.php?id=61141&edit=1 ID: 61141 Comment by: amoo_miki at yahoo dot com Reported by:amoo_miki at yahoo dot com Summary:curl_multi_select returns -1 Status: Assigned Type: Bug Package:cURL related Operating System: Windows 7 x64 PHP Version:5.3.10 Assigned To:pierrick Block user comment: N Private report: N New Comment: PHP 5.3.17 and 5.4.7 too suffer from this bug. Previous Comments: [2012-06-06 20:57:57] ihate at spam dot com Always returns -1 on Windows 7 64bit. Tried on 5.4 and 5.4.3. It worked a couple of months ago, perhaps some Win7 update changed something. [2012-05-30 02:48:04] larue...@php.net Pierrick, could you please look at this? :) [2012-05-29 11:52:09] amoo_miki at yahoo dot com I can confirm that PHP 5.4.3 (cli) (built: May 8 2012 00:23:27) suffers from the same bug. [2012-05-25 12:08:45] kulminaator at gmail dot com Running PHP 5.4.0 on Windows 7 has the same issue, curl_multi_select always returns -1 PHP Version => 5.4.0 System => Windows NT MSA-3644048 6.1 build 7601 (Windows 7 Enterprise Edition Service Pack 1) i586 cURL support => enabled cURL Information => 7.24.0 Age => 3 Features AsynchDNS => Yes Debug => No GSS-Negotiate => Yes IDN => No IPv6 => Yes Largefile => Yes NTLM => Yes SPNEGO => No SSL => Yes SSPI => Yes krb4 => No libz => Yes CharConv => No Protocols => dict, file, ftp, ftps, gopher, http, https, imap, imaps, ldap, pop3, pop3s, rtsp, scp, sftp, smtp, smtps, telnet, tftp Host => i386-pc-win32 SSL Version => OpenSSL/0.9.8t ZLib Version => 1.2.5 libSSH Version => libssh2/1.3.0 [2012-04-03 16:26:35] bompus at gmail dot com Related to 60790 and 61240 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=61141 -- Edit this bug report at https://bugs.php.net/bug.php?id=61141&edit=1
Req #63137 [Com]: Allow custom GLOBAL variables
Edit report at https://bugs.php.net/bug.php?id=63137&edit=1
ID: 63137
Comment by: valentiny510 at yahoo dot es
Reported by:valentiny510 at yahoo dot es
Summary:Allow custom GLOBAL variables
Status: Open
Type: Feature/Change Request
Package:Variables related
Operating System: XP
PHP Version:5.4.7
Block user comment: N
Private report: N
New Comment:
I "flashed" the source code and founded the "zend_register_auto_global" and Im
sure it can be done in the "php_startup_auto_globals" function or some other
place, php_variables.c or zend.c just like:
zend_register_auto_global("GLOBALS", sizeof("GLOBALS") - 1, 1,
php_auto_globals_create_globals TSRMLS_CC);
but with our custom variable from php.ini for example.. just use
ZEND_INI_ENTRY("custom_globals" ... and done ! :D
I know it can be done easy but now depend of you guys if you want to implement
it or not.
Previous Comments:
[2012-09-22 08:23:36] valentiny510 at yahoo dot es
Description:
I think can be very usefull, sometimes, to be able to make a custom global
variable, like languages, configs, etc.. and working just like $GLOBALS but
only with the value/values assigned. I'm sure more than 75%-80% of the Php
users will be agreed with this change.
Imagine and array with some translation strings:
$lang = array(
'bla' => 'some nonsense talk',
'blabla' => 'More nonsense talks');
and now the 'trick'.. some function to make 'global' the $lang:
make_global( $lang );
or even set in the php.ini some configuration like:
custom_globals = "lang, config, more, etc, bla, bla"
and now we can use it on global scope
function test() { return $lang['bla']; }
is a 'nicer' way than
function test() { return $GLOBALS['lang']['bla']; }
or
function test() { global $lang; return $lang['bla']; }
Test script:
---
-- none --
Expected result:
-- none --
Actual result:
--
-- none --
--
Edit this bug report at https://bugs.php.net/bug.php?id=63137&edit=1
Req #22958 [Opn->Wfx]: more returned elements from debug_backtrace() function
Edit report at https://bugs.php.net/bug.php?id=22958&edit=1 ID: 22958 Updated by: ni...@php.net Reported by:ondrej at kmit dot sk Summary:more returned elements from debug_backtrace() function -Status: Open +Status: Wont fix Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: any PHP Version:4.3.0 Block user comment: N Private report: N New Comment: Adding such information would add significant overhead to function calls. If you want to have profiling information, use a profiling extension (e.g. XDebug or xhprof). Previous Comments: [2003-03-30 05:12:39] ondrej at kmit dot sk Add enter/exit high resolution timestamp for function, memmory usage to debug_buacktrace output. -- Edit this bug report at https://bugs.php.net/bug.php?id=22958&edit=1
Req #23346 [Opn->Wfx]: stripslashes extention
Edit report at https://bugs.php.net/bug.php?id=23346&edit=1 ID: 23346 Updated by: ni...@php.net Reported by:php at dries dot net Summary:stripslashes extention -Status: Open +Status: Wont fix Type: Feature/Change Request -Package:Feature/Change Request +Package:*General Issues Operating System: all PHP Version:4.3.0 Block user comment: N Private report: N New Comment: Closing as magic quotes were dropped in PHP 5.4. Previous Comments: [2003-04-25 10:38:20] php at dries dot net I think it would be usefull to extend the stripslashes function so it would check for magic_quotes on or of. So, for example: stripslashes($string) would always strip slashes while stripslashes($string, 1) would only strip slashes when magic_quotes is on -- Edit this bug report at https://bugs.php.net/bug.php?id=23346&edit=1
[PHP-BUG] Bug #63138 [NEW]: dozens of implicit declarations in custom configure tests
From: lullol at yopmail dot com Operating system: Linux PHP version: 5.4Git-2012-09-22 (Git) Package: Compile Failure Bug Type: Bug Bug description:dozens of implicit declarations in custom configure tests Description: basically all custom (i.e. written by php coders rather than autoconf authors) configure checks use exit() but fail to include which defines this function. they use strcmp without including etc etc so if your distro toolchain has -Werror-implicit-function-declaration baked in (which is basically a very sane feature to use, because it will prevent broken code generation on platforms where sizeof(int) != sizeof(void*), like amd64), all these tests will fail, and either cause total build failure, cause duplicated code to get pulled in, or build without some optional packages which are misdetected as not existing. examples include: (using the cache variable names of those tests) _cv_have_broken_glibc_fopen_append ac_cv_what_readdir_r ac_cv_crypt_blowfish ac_cv_crypt_md5 ac_cv_crypt_SHA256 ac_cv_crypt_SHA512 ac_cv_crypt_des ac_cv_crypt_ext_des ac_cv_pwrite ac_cv_pread the broken "what readdir_r" check will cause "old readdir_r" support code get compiled, which fails to build because it is using an invalid prototype for POSIX readdir_r. to find all such issues automatically, it is recommended to use "CC=gcc -Werror- implicit-function-declaration" ./configure, then grep config.log for "error: implicit declaration" additionally, musl libc should be used for this instead of glibc, because the latter has numerous namespace issues. for example on glibc, including will automatically cause to get pulled in, thus hiding the issue with exit(). another approach is to manually go through all the custom tests, analyse which functions are used and add the headers that define it according to the POSIX or C spec. -- Edit bug report at https://bugs.php.net/bug.php?id=63138&edit=1 -- Try a snapshot (PHP 5.4): https://bugs.php.net/fix.php?id=63138&r=trysnapshot54 Try a snapshot (PHP 5.3): https://bugs.php.net/fix.php?id=63138&r=trysnapshot53 Try a snapshot (trunk): https://bugs.php.net/fix.php?id=63138&r=trysnapshottrunk Fixed in SVN: https://bugs.php.net/fix.php?id=63138&r=fixed Fixed in release: https://bugs.php.net/fix.php?id=63138&r=alreadyfixed Need backtrace: https://bugs.php.net/fix.php?id=63138&r=needtrace Need Reproduce Script: https://bugs.php.net/fix.php?id=63138&r=needscript Try newer version: https://bugs.php.net/fix.php?id=63138&r=oldversion Not developer issue:https://bugs.php.net/fix.php?id=63138&r=support Expected behavior: https://bugs.php.net/fix.php?id=63138&r=notwrong Not enough info: https://bugs.php.net/fix.php?id=63138&r=notenoughinfo Submitted twice: https://bugs.php.net/fix.php?id=63138&r=submittedtwice register_globals: https://bugs.php.net/fix.php?id=63138&r=globals PHP 4 support discontinued: https://bugs.php.net/fix.php?id=63138&r=php4 Daylight Savings: https://bugs.php.net/fix.php?id=63138&r=dst IIS Stability: https://bugs.php.net/fix.php?id=63138&r=isapi Install GNU Sed:https://bugs.php.net/fix.php?id=63138&r=gnused Floating point limitations: https://bugs.php.net/fix.php?id=63138&r=float No Zend Extensions: https://bugs.php.net/fix.php?id=63138&r=nozend MySQL Configuration Error: https://bugs.php.net/fix.php?id=63138&r=mysqlcfg
Bug #61240 [Opn->Nab]: curl_multi_select cause cpu load 100%
Edit report at https://bugs.php.net/bug.php?id=61240&edit=1
ID: 61240
Updated by: pierr...@php.net
Reported by:niniwzw at qq dot com
Summary:curl_multi_select cause cpu load 100%
-Status: Open
+Status: Not a bug
Type: Bug
Package:cURL related
Operating System: window xp
PHP Version:5.3.10
-Assigned To:
+Assigned To:pierrick
Block user comment: N
Private report: N
New Comment:
This is not a bug. This behaviour is due to the same reason as bug #61141 (cf
my
comment on this bug).
Previous Comments:
[2012-04-05 02:01:06] niniwzw at qq dot com
the best solution work without the curl_multi_select may be this:
http://lxr.php.net/";);
curl_setopt($ch1, CURLOPT_HEADER, 0);
curl_setopt($ch2, CURLOPT_URL, "http://www.php.net/";);
curl_setopt($ch2, CURLOPT_HEADER, 0);
//create the multiple cURL handle
$mh = curl_multi_init();
//add the two handles
curl_multi_add_handle($mh,$ch1);
curl_multi_add_handle($mh,$ch2);
$active = null;
$exec_count = 0;
do {
$mrc = curl_multi_exec($mh, $active);
usleep(1);
$exec_count++;
} while ($mrc == CURLM_CALL_MULTI_PERFORM || ($active && $mrc == CURLM_OK));
//close the handles
curl_multi_remove_handle($mh, $ch1);
curl_multi_remove_handle($mh, $ch2);
curl_multi_close($mh);
var_dump($exec_count);
?>
I scan the source code of curl_multi_select, there was no change from php
5.2.17 to 5.3.10, may be it's the problem of libcurl(curl_multi_fdset, fuction)
[2012-04-03 16:27:47] bompus at gmail dot com
Related to 60790 and 61141
[2012-03-02 08:51:47] niniwzw at qq dot com
Description:
---
>From manual page: http://www.php.net/function.curl-multi-exec
---
In the document of curl_multi_exec, has a example in the document.
it can run in php 5.2, but php 5.3 and php 5.4 cause cpu load 100%.
the example is:
http://lxr.php.net/";);
curl_setopt($ch1, CURLOPT_HEADER, 0);
curl_setopt($ch2, CURLOPT_URL, "http://www.php.net/";);
curl_setopt($ch2, CURLOPT_HEADER, 0);
//create the multiple cURL handle
$mh = curl_multi_init();
//add the two handles
curl_multi_add_handle($mh,$ch1);
curl_multi_add_handle($mh,$ch2);
$active = null;
//execute the handles
do {
$mrc = curl_multi_exec($mh, $active);
} while ($mrc == CURLM_CALL_MULTI_PERFORM);
while ($active && $mrc == CURLM_OK) {
if (curl_multi_select($mh) != -1) {
do {
$mrc = curl_multi_exec($mh, $active);
} while ($mrc == CURLM_CALL_MULTI_PERFORM);
}
}
//close the handles
curl_multi_remove_handle($mh, $ch1);
curl_multi_remove_handle($mh, $ch2);
curl_multi_close($mh);
?>
--
Edit this bug report at https://bugs.php.net/bug.php?id=61240&edit=1
Bug #60790 [Opn->Fbk]: "curl_multi_select" executed first time always one second
Edit report at https://bugs.php.net/bug.php?id=60790&edit=1
ID: 60790
Updated by: pierr...@php.net
Reported by:leonchuk at gmail dot com
Summary:"curl_multi_select" executed first time always one
second
-Status: Open
+Status: Feedback
Type: Bug
Package:cURL related
Operating System: windows (XP,7)
PHP Version:5.3.9
Block user comment: N
Private report: N
New Comment:
This behavior is probably related to Bug #61141. Can you try this script
$ch1 = curl_init();
curl_setopt_array($ch1, array(CURLOPT_URL=>"http://lxr.php.net/";,
CURLOPT_HEADER=>0, CURLOPT_RETURNTRANSFER=>1));
$mh = curl_multi_init();
curl_multi_add_handle($mh,$ch1);
$active = null;
do $mrc = curl_multi_exec($mh, $active);
while ($mrc == CURLM_CALL_MULTI_PERFORM);
$start_time = microtime(true);
while ($active && $mrc == CURLM_OK) {
if (curl_multi_select($mh) == -1) {
usleep(100);
}
do $mrc = curl_multi_exec($mh, $active);
while ($mrc == CURLM_CALL_MULTI_PERFORM);
}
curl_multi_remove_handle($mh, $ch1); curl_multi_close($mh);
And let me know if you still have your problem ?
Previous Comments:
[2012-04-03 16:27:11] bompus at gmail dot com
Possibly related to 61141 and 61240
[2012-03-24 07:28:59] qiwei dot mo at gmail dot com
I have the same problem white them.
Env: win7 + php5.4.0
Use the code as curl_multi-exec
http://cn2.php.net/manual/zh/function.curl-multi-exec.php
cause the final error: Fatal error: Maximum execution time of 30 seconds
exceeded. But it be runing well in php5.3.6
In PHP5.4 , when i replace the under code
---
$active = null;
// æ§è¡æ¹å¤çå¥æ
do {
$mrc = curl_multi_exec($mh, $active);
} while ($mrc == CURLM_CALL_MULTI_PERFORM);
while ($active && $mrc == CURLM_OK) {
if (curl_multi_select($mh) != -1) {
do {
$mrc = curl_multi_exec($mh, $active);
} while ($mrc == CURLM_CALL_MULTI_PERFORM);
}
}
--
With
do {
curl_multi_exec($hMultiCurl, $iRunning);
} while ($iRunning > 0);
-
the programe run ok
So the question is the curl_multi_select bug , cause the while infinite loop.
[2012-02-28 17:04:37] bompus at gmail dot com
I've noticed this as well. The function should return as soon as it is finished
I
believe. This could be a CURL issue or a PHP issue. I'm interested to hear what
the dev's say since I've had all kinds of issues with curl_multi_select in
newer
PHP versions. I keep having to revert to 5.3.6.. I've tried 5.3.10 and using
curl_multi_select within a loop checking for != -1 and it causes an infinite
loop
since it always returns -1 on Windows 7 x64/x86, PHP 5.3.10
[2012-01-18 13:18:08] leonchuk at gmail dot com
Description:
In Windows (tested XP and w7) function "curl_multi_select" executed first time
always one second (equal to default parameter - $timeout)
---
>From manual page:
>http://www.php.net/function.curl-multi-init#refsect1-function.curl-multi-init-examples
---
Test script:
---
$ch1 = curl_init();
curl_setopt_array($ch1, array(CURLOPT_URL=>"http://lxr.php.net/";,
CURLOPT_HEADER=>0, CURLOPT_RETURNTRANSFER=>1));
$mh = curl_multi_init();
curl_multi_add_handle($mh,$ch1);
$active = null;
do $mrc = curl_multi_exec($mh, $active);
while ($mrc == CURLM_CALL_MULTI_PERFORM);
$start_time = microtime(true);
while ($active && $mrc == CURLM_OK) {
if (curl_multi_select($mh) != -1) { #after first run = 1 sec
echo (microtime(true)-$start_time)."".PHP_EOL;
do $mrc = curl_multi_exec($mh, $active);
while ($mrc == CURLM_CALL_MULTI_PERFORM);
}
}
curl_multi_remove_handle($mh, $ch1); curl_multi_close($mh);
--
Edit this bug report at https://bugs.php.net/bug.php?id=60790&edit=1
Bug #62641 [Opn->Nab]: curl_multi times out
Edit report at https://bugs.php.net/bug.php?id=62641&edit=1 ID: 62641 Updated by: pierr...@php.net Reported by:jimi_barka at yahoo dot com Summary:curl_multi times out -Status: Open +Status: Not a bug Type: Bug Package:cURL related Operating System: Windows 7 PHP Version:5.4.5 Block user comment: N Private report: N New Comment: Please do not submit the same bug more than once. An existing bug report already describes this very problem. Even if you feel that your issue is somewhat different, the resolution is likely to be the same. Thank you for your interest in PHP. This bug is a duplicate of Bug #61141 Previous Comments: [2012-07-23 15:08:11] jimi_barka at yahoo dot com Description: Running the sample script provided here: http://php.net/manual/en/function.curl-multi-init.php results in a timeout. This bug is not present in version 5.3.8, but does appear in 5.4.0 and in later versions, including version 5.4.5. PHP info (excerpt): System Windows NT 6.1 build 7601 (Windows 7 Business Edition Service Pack 1) i586 Build Date Jul 18 2012 22:21:56 CompilerMSVC9 (Visual C++ 2008) Architecturex86 Configure Command cscript /nologo configure.js "--enable-snapshot-build" "--disable-isapi" "--enable-debug-pack" "--without-mssql" "--without-pdo-mssql" "--without-pi3web" "--with-pdo-oci=C:\php-sdk\oracle\instantclient10\sdk,shared" "--with-oci8=C:\php-sdk\oracle\instantclient10\sdk,shared" "--with-oci8-11g=C:\php-sdk\oracle\instantclient11\sdk,shared" "--enable-object-out-dir=../obj/" "--enable-com-dotnet=shared" "--with-mcrypt=static" "--disable-static-analyze" "--with-pgo" Apache Version Apache/2.2.22 (Win32) PHP/5.4.5 Binaries from http://windows.php.net/downloads/releases/php-5.4.5-Win32-VC9-x86.zip Test script: --- http://php.net/manual/en/function.curl-multi-init.php Example #1 curl_multi_init() example Expected result: Display contents of http://lxr.php.net/ and http://www.php.net/ Actual result: -- Script times out with a message: Fatal error: Maximum execution time of 30 seconds exceeded in E:\www\curl.php on line 26 -- Edit this bug report at https://bugs.php.net/bug.php?id=62641&edit=1
Bug #62036 [Opn->Fbk]: CURLOPT_RETURNTRANSFER returns both boolean and response
Edit report at https://bugs.php.net/bug.php?id=62036&edit=1
ID: 62036
Updated by: pierr...@php.net
Reported by:ken dot depelchin at gmail dot com
Summary:CURLOPT_RETURNTRANSFER returns both boolean and
response
-Status: Open
+Status: Feedback
Type: Bug
Package:cURL related
Operating System: N/A
PHP Version:5.3.13
Block user comment: N
Private report: N
New Comment:
Thank you for this bug report. To properly diagnose the problem, we
need a short but complete example script to be able to reproduce
this bug ourselves.
A proper reproducing script starts with ,
is max. 10-20 lines long and does not require any external
resources such as databases, etc. If the script requires a
database to demonstrate the issue, please make sure it creates
all necessary tables, stored procedures etc.
Please avoid embedding huge scripts into the report.
The script work for me if I do a json_decode of $result. Can you give me the
exact script you're using to get your problem (if it's not this one) but also
the
expected and actual result.
Previous Comments:
[2012-05-15 11:18:33] ken dot depelchin at gmail dot com
Description:
When calling a API and converting the response to an associative array, the
returned response can not be converted because of the extra boolean in the
response. Only viewable when exit($response)
The options of the cURL operation are set using the @curl_setopt_array($ch,
$array).
When checked for the header, it does not indicate any problem and it outputs
the correct content (with content length)
When dumped the boolean is dumped with the correct response.
Test script:
---
// check if curl is available
if(!function_exists('curl_init')) throw new Exception('This method requires
cURL (http://php.net/curl), it seems like the extension isn\'t installed.');
$url =
'http://ws.audioscrobbler.com/2.0/?method=album.getbuylinks&artist=radiohead&album=in%20rainbows&country=united%20kingdom&api_key=b25b959554ed76058ac220b7b2e0a026&format=json';
// set options
// $options[CURLOPT_URL] = (string) $url;
// $options[CURLOPT_USERAGENT] = 'PlonkHTTP '. self::version;
// $options[CURLOPT_FOLLOWLOCATION] = true;
// $options[CURLOPT_RETURNTRANSFER] = true;
// $options[CURLOPT_TIMEOUT] = 20;
// $options[CURLOPT_HEADER] = true;
// init
$curl = @curl_init();
// @curl_setopt_array($curl, $options);
curl_setopt($curl,CURLOPT_URL, (string) $url);
curl_setopt($curl,CURLOPT_USERAGENT,'PlonkHTTP');
curl_setopt($curl,CURLOPT_RETURNTRANSFER, true);
// execute
$response = @curl_exec($curl);
// fetch errors
$errorNumber = @curl_errno($curl);
$errorMessage = @curl_error($curl);
// close
@curl_close($curl);
Expected result:
The script will result in a decent response on a local webserver. In a
production environment, it gives the extra boolean when outputted.
Actual result:
--
http://ws.audioscrobbler.com/2.0/?method=album.getbuylinks&artist=radiohead&album=in%20rainbows&country=united%20kingdom&api_key=b25b959554ed76058ac220b7b2e0a026&format=json
--
Edit this bug report at https://bugs.php.net/bug.php?id=62036&edit=1
Bug #63132 [Opn->Csd]: Generators do not back up argument type stack
Edit report at https://bugs.php.net/bug.php?id=63132&edit=1 ID: 63132 Updated by: ni...@php.net Reported by:ni...@php.net Summary:Generators do not back up argument type stack -Status: Open +Status: Closed Type: Bug Package:Scripting Engine problem PHP Version:5.4Git-2012-09-21 (Git) Block user comment: N Private report: N New Comment: Automatic comment on behalf of nikic Revision: http://git.php.net/?p=php-src.git;a=commit;h=a31fa55b44bcb342c00e9ab2f4a851d054897a39 Log: Fixed bug #63132 Previous Comments: [2012-09-21 14:53:18] ni...@php.net Description: As reported by dmitry in https://bugs.php.net/bug.php?id=63066 there currently are issues with yield interruptions in nested method calls. The following code: foo($obj->foo(yield)); } $g1 = gen(new A); $g1->current(); $g2 = gen(new B); $g2->current(); $g1->next(); Will output: Called A::foo Called B::foo [Fri Sep 21 16:49:14 2012] Script: '/home/nikic/dev/php-src/t26.php' /home/nikic/dev/php-src/Zend/zend_vm_execute.h(949) : Freeing 0xB77579F0 (20 bytes), script=/home/nikic/dev/php-src/t26.php === Total 1 memory leaks detected === The issue is that generators currently only back up the argument stack (the main VM stack), but don't back up the argument type stack EG(arg_types_stack). This stack contains object and fbc. So when two generators with nested method calls are intermixed the VM mixes up the different objects/fbcs (and if a generator is closed during the nested method call the object leaks). -- Edit this bug report at https://bugs.php.net/bug.php?id=63132&edit=1
Bug #63138 [Com]: dozens of implicit declarations in custom configure tests
Edit report at https://bugs.php.net/bug.php?id=63138&edit=1 ID: 63138 Comment by: lullol at yopmail dot com Reported by:lullol at yopmail dot com Summary:dozens of implicit declarations in custom configure tests Status: Open Type: Bug Package:Compile Failure Operating System: Linux PHP Version:5.4Git-2012-09-22 (Git) Block user comment: N Private report: N New Comment: oops, _cv_have_broken_glibc_fopen_append doesn't belong to this group, it is broken in its own special way, in that it is triggered whenever you don't use glibc. Previous Comments: [2012-09-22 13:26:28] lullol at yopmail dot com Description: basically all custom (i.e. written by php coders rather than autoconf authors) configure checks use exit() but fail to include which defines this function. they use strcmp without including etc etc so if your distro toolchain has -Werror-implicit-function-declaration baked in (which is basically a very sane feature to use, because it will prevent broken code generation on platforms where sizeof(int) != sizeof(void*), like amd64), all these tests will fail, and either cause total build failure, cause duplicated code to get pulled in, or build without some optional packages which are misdetected as not existing. examples include: (using the cache variable names of those tests) _cv_have_broken_glibc_fopen_append ac_cv_what_readdir_r ac_cv_crypt_blowfish ac_cv_crypt_md5 ac_cv_crypt_SHA256 ac_cv_crypt_SHA512 ac_cv_crypt_des ac_cv_crypt_ext_des ac_cv_pwrite ac_cv_pread the broken "what readdir_r" check will cause "old readdir_r" support code get compiled, which fails to build because it is using an invalid prototype for POSIX readdir_r. to find all such issues automatically, it is recommended to use "CC=gcc -Werror- implicit-function-declaration" ./configure, then grep config.log for "error: implicit declaration" additionally, musl libc should be used for this instead of glibc, because the latter has numerous namespace issues. for example on glibc, including will automatically cause to get pulled in, thus hiding the issue with exit(). another approach is to manually go through all the custom tests, analyse which functions are used and add the headers that define it according to the POSIX or C spec. -- Edit this bug report at https://bugs.php.net/bug.php?id=63138&edit=1
