{5} Assigned, Active Tickets by Owner (Full Description) (40 matches)

List tickets assigned, group by ticket owner. This report demonstrates the use of full-row display.

anonymous

Ticket Summary Component Milestone Type Created
Description
#1118 files modes in fopen Standard Library 3.0.0 release defect 11/26/03
<?php # vim600:syn=php:
$filename = tempnam("/tmp", "phpt");

$fp = fopen($filename, "w+") or die("can't open $filename for append");
fwrite($fp, "quxbar");
fseek($fp, 3, SEEK_SET);
fread($fp, 1);
fseek($fp, 4, SEEK_SET);
fwrite($fp, '!');
fseek($fp, 0, SEEK_SET);
var_dump(fread($fp, 4095));

ftruncate($fp, 0);
rewind($fp);

fwrite($fp, "barfoo");
fseek($fp, 3, SEEK_SET);
fread($fp, 1);
fwrite($fp, '!');
fseek($fp, 0, SEEK_SET);
var_dump(fread($fp, 4095));

fclose($fp);
unlink($filename);
?>


ADDITIONAL INFORMATION:
File "../phpoo/runtime/ext/standard/./php-files.scm", line 730, character 24221:
#		(else (error "fopen" "invalid file mode" mode)))))))
#		      ^
# *** ERROR:bigloo:fopen
# invalid file mode -- w+
    0. fopen
    1. php-funcall
    2. apply
    3. evaluate (3 times)
    6. evaluate-block
    7. evaluate-default2438
    8. evaluate (2 times)
   10. interpret
   11. php-commandline

#1710 pcc doesn't wait for processes opened with popen to end Standard Library 3.1.0 beta defect 03/29/04
Given the script:

<?php
$foo = popen('tail -f /etc/motd','r');
?>

php never returns because it is waiting for the process
to end or be closed, and pcc returns right away without
waiting for anything.

tim

Ticket Summary Component Milestone Type Created
Description
#2957 string parsing problem Parser/Lexer 3.0.0 release defect 03/28/05
<?php

$class = 'foo';
echo "$class{";

?>


ADDITIONAL INFORMATION:
[weyrick@hikari:tmp]$ php -f bug.php
foo{
[weyrick@hikari:tmp]$ pcc -f bug.php

File "./lexers.scm", character 3633:

*** ERROR:bigloo:parser:
bug.php:4: bug.php:4: parse error (unexpected token `*eoi*') -- #eof-object


#1680 deep recursion causes segfaults, even in simple cases Core 3.1.0 beta defect 03/23/04
<?php
function bla() {
  bla();
};
bla();
?>

That's admittedly a useless example, but it should, perhaps
with pcc -O, turn into a nice tail-recursive loop.


#3065 property with a space won't parse in double quoted string Parser/Lexer 3.0.0 release defect 05/27/05
this produces a parse error in pcc, works in php:

<?php

class foo {
  function foo() {
    $this -> foo = 22;
  }
}

$bar = new foo();
echo $bar -> foo . "\n";
echo "{$bar -> foo}\n";

?>



#3067 parse error on global def, variable variable Parser/Lexer 3.0.0 release defect 06/02/05
using "global ${$foobar}" throws a parse error, should work
the same as "global $$foobar"
 

#3222 we error on $foo[]['bar'][] = 'baz'; but we shouldn't Parser/Lexer 3.0.0 release defect 03/12/06
<?php
class foo {
    function foo($vararray) {
        $this->_tpldata['switch_user_logged_out.'][]['switch_allow_autologin.'][] = $vararray;
    }
}

$foo = new foo("foo");
?>



#3313 number conversion Core 3.0.0 release defect 08/30/06
$test='2507972608';
$test2=$test/1024/1024/1024;
echo "Test =  $test2";

gives : -1.66426849......
it suppouse to give : 2.33573150635


#2615 interpreter creates properties it shouldn't Core 3.0.0 release PHP deviation 11/23/04
<?php

class foo {}
$foo = new foo();
$bar = $foo->width;
$bar = $foo->height;

var_dump($foo);

?>

#3252 different reference behavior Core 3.1.0 beta PHP deviation 05/20/06
not sure if we need to care about this, but this ends up 
producing different results:

<?php
$foo = 1;
$bar =& $foo + 1;
echo "$bar\n";
$bar = "other value";
echo "$foo\n";
?>



#3255 fopen needs to emit a warning when it fails Core 3.0.0 release PHP deviation 05/24/06
like this one:

Warning: fopen(/tmp/asdfasdfasdf): failed to open stream: No such file or directory in /tmp/foo.php on line 3

#1943 xor(??) problem Core 3.0.0 release defect 06/11/04
runs interperted ok, but complains when you try to compile

<?
    if ((isset($tid)) && (!isset($pid))) {
        singlecomment($tid, $sid, $mode, $order, $thold);
    } elseif (($mainfile) xor (($pid==0) AND (!isset($pid)))) {
        Header("Location: modules.php?name=$module_name&file=article&sid=$sid&mode=$mode&order=$order&thold=$thold");
    } else {
        if(!isset($pid)) $pid=0;
        DisplayTopic($sid, $pid, $tid, $mode, $order, $thold);
    }
?>

#2037 var_dump() can't tell if its argument is a reference Core 3.0.0 release defect 07/25/04
var_dump can tell if it was passed a reference or not:

<?php

$foo = 2;
var_dump(&$foo);
var_dump($foo);

?>




#2069 problems with file Core 3.0.0 release defect 08/06/04
the following test script sucks up alot of memory and doesn't return with-in a reasonable amount of time (I killed the process after 5+ minutes). This is probably related to bug #0002066

<?
$host = "tenkan.org";
$page = "/";

$lines = file("http://$host/$page");
foreach ($lines as $line) {
        echo $line;
}
?>

#2428 Including the same file file via a differnt path Core 3.0.0 release defect 10/04/04
If you include the same file via a different path, pcc doesn't recognize that it's the same file which causes it to be included again evein if it's an include_once.

This was making pcc think that a function was getting re-defined when it shouldn't have been.


<---------include.php----------->
<?
function foo() {
        echo "in foo\n";
}
?>
<------------------------------->


<---------test.php-------------->
<?
$pwd = getcwd();

include('include.php');
include_once($pwd . '/include.php');

foo();

?>
<------------------------------->


#2757 type isn't a tight bound Core 3.0.0 release defect 01/18/05
the type of $qp should be number, but it's (any number)


<?php
    function executeEmulateQuery()
    {
        $qp =  1;

        for ($i == 0; $i < $qp; $i++) {
            if ($a != $b) {
                $c = 2;
            } 
        }
    }


?>


#2811 numeric hash keys should be signed int Core 3.0.0 release defect 01/30/05
<?php

$foo[2147483647] = "foo";

$foo[] = "bar";

$foo[2147483649] = "baz";

$foo[] = "bling";

var_dump($foo);

$bar[1] =  "at one";
$bar["1"] =  "at one";
var_dump($bar);

?>

#3019 when there's a #! line at the top of the file, our line number counts are off by one Core 3.0.0 release defect 05/11/05
you can see this problem in Phpmole.php which comes with phpmole.

#3078 regex inside quoted string parsed as string index Parser/Lexer 3.0.0 release defect 06/09/05
echo "$foo{2,20}"
php: {2,20}
pcc: (error)


ADDITIONAL INFORMATION:
Original source via phpgroupware package:

phpgroupware/email/inc/class.mail_msg_base.inc.php:3907: parse error
          (unexpected token `comma') -- comma

#3081 segfault in mktime Core 3.1.0 beta defect 06/16/05
$ts_new = mktime($hour, $min, $sec, $month, $day, $year);

bug was reported on windows

haven't verified it

#3142 array access with double quote, in a double quoted string Parser/Lexer 3.0.0 release defect 11/17/05
right now this fails:

echo ("pages[".$pageCount++."] = new realmPage(\"{$pageInfo["pageTitle"]}\",\"{$pageInfo["pageID"]}\"); \n");

but this is ok:

echo ("pages[".$pageCount++."] = new realmPage(\"{$pageInfo[pageTitle]}\",\"{$pageInfo[pageID]}\"); \n");

as is this:

echo ("pages[".$pageCount++."] = new realmPage(\"{$pageInfo['pageTitle']}\",\"{$pageInfo['pageID']}\"); \n");


#855 strtotime Standard Library 3.1.0 beta enhancement 10/10/03
support is there but minimal

#3195 add -W switch Core 3.1.0 beta wishlist 02/08/06
it would enable warnings for undefined functions at compile time, for one thing.


#2810 easy to get pathological hash behavior when low-order bits don't change Core None 01/30/05
This script takes 38 seconds or so in both zend php and phpoo, since we just bit-and numerical hash keys with a mask, so when the low-order bits don't change, they all end up in the same bucket.  Using the fnv-1a hash for numerical keys makes us take only 1.2 seconds, but at the cost of also taking 1.2 seconds when the low-order bits _do_ change, when  php takes only 0.6.  Since we can't afford to be slower, and this is a somewhat rare case, I'll leave this out for now.  But the code for the numerical hash is in php-hash-helper.c, should it ever come up.


<?php

$inc = 1024;
for ($i=0; $i<($inc * 100000); $i+=$inc) {
	$foo[$i] = "on";
}
for ($i=0; $i<($inc * 100000); $i+=$inc) {
	$baz[$i] =  $foo[$i];
}
print $i . "\n";
print count($foo) . "\n";

?>



#1910 defbuiltin optional arg default values don't get eval'd Generator 3.1.0 beta enhancement 06/01/04
Given a php-builtin defined like this:

(define *scheme-var* 10)
(defbuiltin (some-builtin (opt-arg1 *scheme-var*))
   opt-arg1))

The following PHP script:

<?php echo some-builtin()."\n"; ?>

produces the _string_:

*scheme-var*

because the default value gets quoted before being evaluated and then when it eventually gets evaluated, it evaluates to the symbol *scheme-var* and not the value of *scheme-var*. 


ADDITIONAL INFORMATION:
changing compiler/generate.scm line 947 from:

,(let ((default (optional-formal-param-default-value p)))

to:
	  
,(let ((default (eval (optional-formal-param-default-value p))))

fixes the problem and doesn't break anything else.

Also, changing line 1552 (or so) in php-runtime.scm from:

((quote ?FOO) FOO)

to:

((quote ?FOO) (eval FOO))

did not fix the problem.


#2944 inline C/assembly Core wishlist 03/17/05
request to allow inline c and/or assembly code in php

weyrick

Ticket Summary Component Milestone Type Created
Description
#3482 Function return error between compiled and interpreted scripts Core 3.0.0 release defect 03/27/07

# test.php <?php

$include='test2.php'; include($include); class c{

function d($pp){

return $pp->b();

}

} $y=new c(); echo $y->d($x);

?>

#test2.php <?php

class a{

function b(){

return 'qq';

}

}

$x=new a();

?>

pcc test.php ./test Runtime error in file test.php on line 6: Type pair' expected, bstring' provided

pcc test.php test2.php ./test qq

If you replace in test2.php return 'qq'; with $x='qq';return $x; it works. So it all comes down to constant return values not being returned properly from a runtime interpreted class to a compiled class.


#3572 microserver: dynamic-load: Can't find library libphp-std_u-3.0c.so Microserver 2.9.7 defect 04/12/08

LD_LIBRARY_PATH=/usr/local/lib PCC_CONF=/usr/local/etc/pcc.conf /usr/local/lib/libphp-std_u-3.0c.so and /usr/local/etc/pcc.conf exist. I didn't edit pcc.conf. There's a string in it: (library "./" "/usr/local/lib")

[root@VMLinux wtorrent]# ./myapp -p 8080
<br /><b>Fatal error:</b> Extension libphp-std_u-3.0c.so didn't load because:
dynamic-load: Can't find library libphp-std_u-3.0c.so
You may wish to remove this extension from /usr/local/etc/pcc.conf if it exists. in unknown on line 0<br />

if I copy libphp-std_u-3.0c.so the same directory where myapp it starts. But in browser I get this error:

Fatal error: Unable to identify class or object: SQLiteDatabase in lib/cls/Web.cls.php on line 47
Stack trace:
File index.php line 29: getClass(ListT)
File lib/cls/Web.cls.php line 96: __construct
File cls/rtorrent.cls.php line 40: __construct

snapshot 20080412 configure params were: --prefix=/usr/local --with-fcgi --with-sockets --with-pcre --with-xml --with-sqlite3


#3561 move object visibility checks to object runtime Runtime 3.0.0 release task 03/06/08

EOM


#3306 2 reports of file uploads not working in microserver Microserver 3.0.0 release defect 08/09/06
eom

#3309 session segfaults and failures Standard Library 3.0.0 release defect 08/09/06
one report of session segfault, another test case of failure (microserver):

page1.php

<?php
session_start();
echo 'Welcome to page #1';
$_SESSION['favcolor'] = 'green';
$_SESSION['animal'] = 'cat';
// Works if session cookie was accepted
echo '<br /><a href="page2.php">page 2</a>';
?>

page2.php

<?php
session_start();
echo 'Welcome to page #2<br />';
echo 'favcolor=' . $_SESSION['favcolor'] . '<br>'; // green
echo 'animal=' . $_SESSION['animal'] .'<br>' ; // cat
echo '<br /><a href="page1.php">page 1</a>';
?>

and of course a suitable mhttpd_startup.inc



#3466 Screwed up POST data parsing Build System 3.0.0 release defect 02/04/07

This one's really nasty. If you POST in an multipart/form-data form, and this form contains a value that is a URL, then the URL will be cut up at the point of the first ampersand, and subsequent parameter segments are absorbed into the top-level POST request.

i.e. If I passed in... <input type="hidden" name="url" value="http://example/foo?a=1&b=1"> then I'd get url=http://example/foo?a=1 b=1 ...


#3465 memoized parse: cache name collision Interpreter 3.0.0 release defect 02/04/07

The interpretor seems to be cacheing the environmental change of frontend files that have exactly the same code and worked upon exactly the same state variables - and thus not executing the files if the result is cached already.

I wrote that in very general terms as I don't know exactly what it is doing. So explained loosely from my point of view...

My system has lots of files sitting in different directories that are identical. These files should execute differently because they execute as a function of the directories they are in. They are all frontends to my system, and the directory that they are in specifies what aspect of the system is called up.


#3502 POST and GET arguments passing doesn't work FastCGI 3.0.0 release defect 11/29/07

It looks like there is a problem in POST and GET arguments passing. If I try to compile and then run web page that contains form, I'm not able to get POST and GET parameters in PHP code.

For example let's have a page like this:

<html> <head> <title>Test Page</title> <link rel="stylesheet" type="text/css" href="stylesheet.css"> </head> <body> <?php echo "GET PARAM ".$HTTP_GET_VARSInputText? ?> <?php echo "GET PARAM ".$_GETOK? ?> <?php echo "Text_start: ".$InputText? ;?> <form action="index.php" method="get" name="Form1"> <input name="InputText?" type="text" value="text" /> <button value="OK" onClick="document.form[0].submit()"></button> <input name="OK" type="submit" value="OK" /> </form> <?php echo "Text_end: ".$InputText? ;?> </body>

I compile the PHP code like a "Online Web Application" to a FastCGI and then I deploy a compiled web application on the Apache web server (Apatche is configurated acording to:http://code.roadsend.com/pcc-manual/Deploying-on-Apache.html#Deploying-on-Apache)

Then I'm able to run the web pages compiled to FastCGI. But there is a problem in POST and GET arguments passing. There are no values in POST and GET passed parameters.

So code:

<?php echo "GET PARAM ".$HTTP_GET_VARSInputText? ?> <?php echo "GET PARAM ".$_GETOK? ?> <?php echo "Text_start: ".$InputText? ;?>

writes only:

GET PARAM GET PARAMText_start:

Even though there should be some values when I call the page with parameters "index.php?InputText?=text&OK=OK"


#3517 bug-id-0001217.php regression: var-compare with objects Runtime 3.0.0 release defect 12/30/07

we still get object comparison wrong, with respect to greater and less than.


#3530 microserver $_SERVER vars Microserver 3.0.0 release defect 01/30/08

microserver doesn't setup HTTP_HOST or SERVER_NAME. Also, on the first load, $_SERVER contains all of the _ENV vars, but on subsequent loads it contains only a few setup by microserver.

This can be see by sticking a var_dump($_SERVER) in the index page, viewing first page load then reloading once.


#3600 pdo extension no longer functions compiled Library Generation/Use 2.9.7 defect 06/17/08

at least with bigloo 3.1a, the PDO extension does not initialize correctly when used for compiled applications (works ok interpreted). there appears to be a problem with initialization order, as pdo_init.php can't find pdo_mysql.php which is compiled into the same library.

the tests/basic.php returns:

File "php-errors.scm", character 6103:
*** ERROR:error-handler:
couldn't find include file pdo_mysql.php (include path was ./) -- error-handler
    0.php-error, php-errors.scm:176
    1.find-include, include.scm:200
    2.do-include, include.scm:145
    3.php-require, include.scm:79

quick investigation shows php-pdo-make-lib_s.c does the init calls in the right order.


#3594 Reference inconsistencies Core 3.0.0 release defect 05/27/08

- unset works improperly and inconsistent for references - referencing uninitialized vars creates circular references and breaks $_SESSION (see the attached php)


#3597 Property access: visibility check should be succeeded by __get PHP Compatability 2.9.7 defect 06/10/08

In Zend, when an attempt is made to access an unaccessible private or protected property, the request is passed to get or set overload functions, if provided; to me, this makes sense. In PCC an error is thrown.


#3490 replace libwebserver with bigloo hop web server Microserver 3.1.0 beta wishlist 08/13/07

this has been discussed briefly with Manuel, and needs some work to detach it from the hop code before it can be used generically


Note: See TracReports for help on using and creating reports.