[PHP] version_compare
I found this code...
if (version_compare(PHP_VERSION, '5.2.0', '>=')) {
$text=filter_var($text, FILTER_SANITIZE_URL);
}
...to be questionable.
Under what conditions would version_compare() return true, yet the filter_var()
be undefined? Because that's what is happening.
Thank you.
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] version_compare
>Personally, I would change that to be
>if ( function_exists('filter_var') ) {
So would I:
*But it's not my code.
*I wish to learn and understand the cause of the problem - not walk around it.
>It means condition (PHP_VERSION >= 5.2.0)
I understand that. There was a second, more relevant, part to my question.
I have found the version of PHP used as reported in the HTTP header responses.
filter_var() is undefined in PHP/5.2.4_p20070914-pl2-gentoo
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] filter_var (was: Re: [PHP] version_compare)
>As Paul pointed out, maybe your version of PHP was built without the >filter_var function compiled in. This is what I have learned about PHP with filter_var() as an illustrative point: Many people who provide elaborations on PHP make too many assumptions or are blatently and woefully incomplete. Statements such as "PHP 5.2.0 or higher is required for this to work" is misleading. The PHP online manual pages for respective functions make no distinction that any particular function or capability is an optional include for a build - considered an "extension" as opposed to, what?, "native code"? I have also learned that when coding my own scripts, I must not only read the manual page for that function, but also any metadata for it and its group, such as the Introduction page, Installing/Configuring, etc. And where, when I finally find it, if I can reason out that I should be looking for it, it may say "The filter extension is enabled by default as of PHP 5.2.0", that I need to be reasonably cautious because some idiot may build PHP 5.2.0+ without this extension. Are the array_*() functions also an optional extension? -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Variable (Class instantiation) collision
I am running into a variable collision. The project I'm developing is NOT guaranteed to be operating on PHP5. Any solution I find should (hopefully) be able to run on PHP4 (yes, I know PHP4 is deprecated). I am building a bridge between two third-party applications. Both instantiate their respective database class assigning it to $db and I cannot change that. So, I am interested in solutions to this. I found a reference to a Packager class and will be looking at it shortly. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] Variable (Class instantiation) collision
>Just to clarify, both packages are instantiating and calling their >respective classes from the $db var, which is in the global scope. >Is this correct? I would say yes to the way you are asking. Take the following two applications. The four respective statements are in each their respective script. Application A: Application B: The bridge project is operating in A and include()'ing the script that is B (as I have not found an API for B). $db in B is colliding with $db in A. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] Variable (Class instantiation) collision
Gentlemen, Thank you. The "Thing1 Thing2" approach worked. >If you have total control over application A which contains the bridge >code, the easiest is to change it to use a different global variable. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Very Large File Splatter
PHP 5.4.4-TS-VC9 on Windows XP SP3 NTFS non-system drive with 18GB free. I dare not try to replicate this. As such, I cannot firmly place the blame on PHP. I have peppered a PHP application with a call to a function which appends-only to a logfile the parameters passed to it. Each pass of the application creates many MB of content. It is conceivable that I ran out of hard drive space. When that which what I was working on seemed to be acting very weird, I rebooted the computer only to see thousands of lines scroll by from Windows repairing the file system. I discovered logfile contents in many dozens of files. The timestamp and filesize of the damaged files were not changed. Only the contents replaced with slices of the logfile. Again, I'm not going to try to 'intentionally' replicate this, so I ask: Has PHP's interface with the NTFS file sub-system ever been reported to splatter a file across the contents of a drive? -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] Very Large File Splatter
"If you can show the write portion of the code in your iteration, as well as a sample of the naming convention, it may offer more clues." $dbgMsg = "a diagnostic string maybe 5K in length"; $dbg_fp = fopen(ROOT_DIR.DS."dbg_log.txt", "a"); // Derives a full path fwrite($dbg_fp, $dbgMsg); fclose($dbg_fp); "Did you shutdown the system cleanly, when you recognized this behaviour?" Yes, a clean reboot. Not a reset-switch restart, nor a power-cycle restart. Sample lines from bootex.log: Checking file system on L: The type of the file system is NTFS. One of your disks needs to be checked for consistency. You may cancel the disk check, but it is strongly recommended that you continue. Windows will now check the disk. Attribute record of type 0x80 and instance tag 0x3 is cross linked starting at 0x20b3b for possibly 0x100 clusters. Attribute record of type 0x80 and instance tag 0x3 is cross linked starting at 0x20b3b for possibly 0x100 clusters. Some clusters occupied by attribute of type 0x80 and instance tag 0x3 in file 0x3680 is already in use. Deleting corrupt attribute record (128, $J) from file record segment 13952. Sorting index $O in file 25. The object id in file 0x4d14 does not appear in the object id index in file 0x19. Inserting an index entry into index $O of file 25. The object id in file 0x4d17 does not appear in the object id index in file 0x19. Inserting an index entry into index $O of file 25. The object id in file 0x4d19 does not appear in the object id index in file 0x19. Recovering orphaned file .svn (19642) into directory file 13948. Recovering orphaned file CATEGO~1.TPL (19720) into directory file 28688. Recovering orphaned file categories.tpl (19720) into directory file 28688. Recovering orphaned file CATEGO~1.BAK (19721) into directory file 28688. Recovering orphaned file categories.tpl.bak (19721) into directory file 28688. And many, many more of each section. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Lifetime
I'm curious to know what the lifetime of a constant is. I know that sounds like a stupid question, but the context is this: I have a file I am writing to up to the very last possible micro-second. As such, I know that as PHP is destroying itself after having executed the last statement in the program, things begin to disappear. The Class objects destruct in no discernable order (they have destructors). Any open file handles close. Etc. My question is: During all this destruction, at what point must I no longer expect to have a constant available? That is, when the classes are executing their __destruct() functions, et al, I want to manage something with those destructors. So when does the constant actually-factually disappear? -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Guaranteed Way to Get Error Message Wanted
I have an application running under PHP-5.4.17-TS-VC9 (and .14 as of yesterday) with Aprelium's Abyss X1 v2.8 web server in FastCGI mode on WinXPSP3. An earlier version of this application works. The current version causes a 500 Internal Server Error. There is no entry in PHP's (fully active) error log. I cannot decipher Abyss's logging, so I cannot determine if a clue was reported by Abyss or not. The current version works on a different system (Server 2003, PHP 5.3.5-TS-VC6 (Apache module), Apache 2.2). What I would like to have is a method of getting PHP to report in some undeniable manner, short of total system failure, what it doesn't like about whatever killed it. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Guaranteed Way to Get Error Message Wanted
>It looks like you are running the thread-safe version with >FastCGI, which I understand to be counter to the recommendations. Thank you for the comment. I switched to PHP5.4.17-NTS-VC9, but the application still crashes. And still no clue as to why. So, still looking for that magic method to get PHP to report what's happening on a 500 Internal Server Error when it's (presumably? not sure...) not the server's fault. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Guaranteed Way to Get Error Message Wanted
>Have you got all your extensions updated? I would think so. Just to state the required disclaimers: phpinfo.php with works. Liberally peppering a tracer routine throughout the application shows it is getting executed up until one spot. But there is nothing obviously wrong with the code. Nothing! That's why I need the guaranteed message delivery on why PHP does not like the code. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Class Auto-Assigning to Variable
I have a situation where, for some unknown reason, where each class that
finishes its __contruct{} function, that class gets automatically assigned to a
variable - other than the variable I specify.
Conceptually:
class Hello { private $_world = 'World'; __construct(){} }
$clsHello = new Hello();
echo 'The variable $hello is '.gettype($hello)."\n".print_r($hello,true);
Output:
The variable $hello is object
Hello Object
(
[_world:Hello:private] => World
)
There is no statement in my application that assigns an instance of the class
to another variable, the name being a lowercase variant of the class name.
Would there be a PHP function that would do this as a side-effect?
I am more interested in learning what is happening as opposed to rolling back
to a previous version. (A backup copy functions fine. A file compare does not
reveal any likely suspects.)
PHP5.4.17-NTS-VC9 (Windows XP-SP3)
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] Class Auto-Assigning to Variable
Second go around:
I have a situation where, for some unknown reason, where each class that
finishes its __contruct{} function, that class gets automatically assigned to a
variable - other than the variable I specify.
Conceptually (a little bit better on the conceptualizing):
class Hello {
private $_world = 'World';
function __construct(){}
}
$clsHello = new Hello();
echo 'The variable $hello is '.gettype($hello)."\n".print_r($hello,true);
Output:
The variable $hello is object
Hello Object
(
[_world:Hello:private] => World
)
There is no statement in my application that assigns an instance of the class
to another variable, the name being a lowercase variant of the class name.
Would there be a PHP function that would do this as a side-effect?
I am more interested in learning what is happening as opposed to rolling back
to a previous version. (A backup copy functions fine. A file compare does not
reveal any likely suspects.)
PHP5.4.17-NTS-VC9 (Windows XP-SP3)
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] Class Auto-Assigning to Variable
>I cannot replicate this. I don't expect anyone to be able to replicate this behavior. The example shows an extraordinarily stripped-down sequence of statements that informs what should work, but do to some unknown agent, which, therefore, cannot be included in the example, produces unexpected output. This conceptual example is not a 'sample' or 'excerpt' of the actual application. There is much, much more code. None of it is an apparent likely suspect in causing this behavior. I am hoping for a, "Oh, yeah! I've seen that happen before. It's caused by..." Let us focus on the central question: Would there be a PHP function that would do [what the example describes] as a side-effect? -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] Class Auto-Assigning to Variable
> $hello = $clsHello; If that conceptual statement (or any occurance of the conceptual $hello) were in the code, then my (really good) Find feature of my code editor would have found it. > There are only a few variables that get assigned as side effects of > functions, but they have very specific names, and none of them are > $hello (but I'm guessing that's not the actual variable name) I have two dozen classes in this application. In every case, there will be a variable, the name of which is a lowercase variant of the class name, to which is assigned an instance of the class, when the class's construct() function completes. The example informs you of this. > Somewhere in your code there is something that is assigning to $hello. > Find everything that's doing that and look at each instance in detail. I have. Many times. What I am looking for is a side-effect -- something not obvious. (This application does not use an eval() function where some obfuscated string manipulation would play this out.) > I would say for definite that it's some of the surrounding code, Exactly. No sooner and no later than precisely when the class's construct() function ends, and control is given to the next statement after the one that instantiated that class. > probably something similar to this: Let's explore this statement: > $GLOBALS[strtolower(get_class($this))] = $this; May I infer that the declaration of $GLOBALS['hello'] will, at the same time, also create $hello (without a statement declaring such)? This: http://php.net/manual/en/reserved.variables.globals.php implies the opposite direction. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP] Class Auto-Assigning to Variable
>Your example does _not_ show this, it works
>as expected and throws a notice, from which can be inferred there is other
>code doing this
Or relevant code having a side-effect not currently realized.
>Is your class maybe inheriting from another one and that contains the code
>causing this issue?
No.
>>May I infer that the declaration of $GLOBALS['hello'] will, at the same
>>time, also create $hello (without a statement declaring such)?
>
>The $GLOBALS array is what's known as a super global. Elements within the
>array are reflected as global variables and vice-versa.
Let's refine the conceptual example:
class Hello {
private $_world = 'World';
function __construct(){}
}
$GLOBALS['hello'] = new Hello();
echo 'The variable $hello is '.gettype($hello)."\n".print_r($hello,true);
Can we then postulate that the value of $GLOBALS['hello'] will also be revealed
in $hello, even though $hello was never formally declared?
I know that $GLOBALS['hello'] has a universal scope, and $hello (wherever it
comes from) needs to be global-ized inside functions.
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Parent Limits?
The following works (three parents):
include("../../../includes/ini.inc.php");
require_once("../../../includes/ini.inc.php");
The following works (four parents):
include("../../../../includes/ini.inc.php");
The following does not work (four parents):
require_once("../../../../includes/ini.inc.php");
That is, require_once() with four parents has been reported to have been
disabled for security purposes. But apparently, not disabled is include() with
four parents.
The device or setting which is causing the problem is not known to me. It may
be just the four parents (and include() is getting the data from this file
elsewhere) or it may be the combination.
It has been reported this is not a limit within open_basedir. So, what is it?
What module or php.ini setting would have this effect?
Thank you.
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Re: Parent Limits?
With all kind respect to Richard Buskirk and Daniel Brown (thank you for
responding), their replies did not actually answer my question.
My question is: What module or php.ini setting would render inoperative a
directory traversal of X parents?
My original post follows.
The following works (three parents):
include("../../../includes/ini.inc.php");
require_once("../../../includes/ini.inc.php");
The following works (four parents):
include("../../../../includes/ini.inc.php");
The following does not work (four parents):
require_once("../../../../includes/ini.inc.php");
That is, require_once() with four parents has been reported to have been
disabled for security purposes. But apparently, not disabled is include() with
four parents.
The device or setting which is causing the problem is not known to me. It may
be just the four parents (and include() is getting the data from this file
elsewhere) or it may be the combination.
It has been reported this is not a limit within open_basedir. So, what is it?
What module or php.ini setting would have this effect?
Thank you.
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] PEAR Mail $obj->send()
A client has: PHP 5.3 on Win7x64 running a local web app that needs to send mail. (This app was once hosted on a linux-based hosted space.) Apache 2.2 is installed but apparently not being used. I think the IIS service is actually the web server that is engaged. During troubleshooting a wide range of problems, I discovered that the PEAR Mail module needed to be installed. So I installed PEAR (the PEAR Installer) and the Mail module with all dependencies. The PEAR_ENV was added. The system was rebooted. A test php script instantiates the Mail class and the script proceeds fine until the send() method is called. I get a browser with "Waiting for localhost" for more than 60 seconds. (I used die(); to trace the script. Instead of 'auth'=> true, I used 'auth' => "PLAIN" as suggested by a user comment on the Mail documentation page.) I believe all the parameters are correct. The actual web app works - except emailing. I added a firewall rule allowing outbound port 465 just for kicks. I can double-check for the PHP timeout setting but would PHP timeout in this case (waiting for a socket??)? Any suggestions? -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Re: PEAR Mail $obj->send()
>>Instead of 'auth'=> true, I used 'auth' => "PLAIN" as
>> suggested by a user comment on the Mail documentation page.)
$obj = Mail::factory('smtp',
array ('host' => $host,
'port' => $port,
'auth' => true,
'username' => $username,
'password' => $password));
(Variables are set. $host = "smtp.gmail.com", $port = "465")
>Do you have a mail server running on localhost?
No.
>If you do not have a local MTA then you need to change the
>params so it uses a remote mail server.
It does.
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Wrong POSTFIELDS Posted
I have a script that accepts four POST variables. Three are used and five more are added for a total of eight keys and their urlencode() values all strung together in the proper format. Then cURL is initialized with the field string given to: curl_setopt($ch, CURLOPT_POST,8); curl_setopt($ch, CURLOPT_POSTFIELDS, $string); But that's not the string arriving at cURL's target URL! What's arriving at the target is exactly the POST array that this script received in the first place. Here's what I don't understand... The cURL request is coming from this script on domainA, but the target address is also domainA. Does everybody play nice in this situation? PHP 5.3.5 Apache 2.2 Note: CURLOPT_POST evaluates to true, regardless that I use an integer. All other curl* commands are set. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
[PHP] Re: ASP to PHP
I've done a site from Classic ASP (no .net) to non-oop PHP. -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
