Re: [PHP-DEV] Question for Zeev, Zak or Georg - mysql_query andpossible integer overflow?
Thanks guys for your feedback. With regard to PEAR, I've had to patch PEAR::DB::mysql to solve another problem I've had with the prepared_queries field growing steadily. I'm sure this is also a problem with the either DB subclasses. I'll post my patch to pear-dev as I don't have a CVS account yet. -Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Question for Zeev, Zak or Georg - mysql_query and possible integeroverflow?
Hi, I have a question for the authors of the mysql extension. I'm sure you gentlemen are very busy, but I'd appreciate your insight if you can spare a moment. I'm developing a stand-alone php application running in an infinite loop from the command line interface. A mysql database is polled continually for new rows to deal with. The same query is executed several times in one second. I've noticed that whether or not a query resource is freed the next query identifier returned from mysql_query() is ++ the last one. I'm sure this is by design and governed by the mysql driver. Naturally I'm concerned about integer overflow when the the application has been running uninterrupted for a very long time. I don't think I can wait around while a test script runs to see what happens after 4,294,967,295 is exceeded though! Can anyone tell me with certainty or hazard a guess what might happen here? I thank you very much for your time. Matt ps - I'm doing this through PEAR::DB::mysql -- Matt Flaherty [EMAIL PROTECTED] Boltblue International Ltd. -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Question for Zeev, Zak or Georg - mysql_query andpossible integer overflow?
Thanks Wez, If all that happens is the query will fail, I can live with that. It's a simple SELECT query called over and over again. If the integer wraps around negative, I'm guessing that would probably have undesirable effects. If not, then I can let it run forever. What I don't want is something like a segfault. I'm considering making the application restart itself periodically to clear out the cobwebs. -Matt On Tue, 2003-03-25 at 14:12, Wez Furlong wrote: Hi Matt, Yes, there is a risk of overflow. From my understanding, the id is signed, so you will hit overflow at 2G rather than 4G resources. This applies to any/all PHP/ZE resources. I'm not sure what happens when it overflows; it seems like the query would fail. You could design your application so that it re-runs itself when the query fails (pcntl_exec(), or one of the other execution functions). --Wez. -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Re: Segmentation violation
J, I will post an issue, but as far as I can tell the error is very intermittent. I don't have a consistent action that I do each time to cause the error. It is just randomly (as far as I can tell) throughout the application. But If you like I'll send some through J Smith wrote: How about providing a short example script that reproduces the segfault? Posting a backtrace is all well and good, but it's not very useful if nobody can see what initiated it. J Matt wrote: Hi, i've never posted to this newsgroup b4 so pls don't shoot me down if this is not protocol. But I've been trying to debug a seg fault trying varying options from various postings with little success. After generating a stacktrace I'm not sure really what to do next to find the problem. If anyone can help me, given the trace below that'd be fantastic. thanks, Matt - [EMAIL PROTECTED] snipped backtrace -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Segmentation violation
Hi, i've never posted to this newsgroup b4 so pls don't shoot me down if
this is not protocol. But I've been trying to debug a seg fault trying
varying options from various postings with little success. After generating
a stacktrace I'm not sure really what to do next to find the problem. If
anyone can help me, given the trace below that'd be fantastic.
thanks,
Matt - [EMAIL PROTECTED]
#0 0x0835036f in ?? ()
#1 0x404090b4 in _zval_dtor (zvalue=0x83545f4) at
/tmp/php-4.3.1/Zend/zend_variables.c:51
#2 0x40402909 in _zval_ptr_dtor (zval_ptr=0xbfffce00) at
/tmp/php-4.3.1/Zend/zend_execute_API.c:289
#3 0x403d56cf in php_var_unserialize (rval=0xbfffce00, p=0xbfffce04,
max=0x8347078 \005,
var_hash=0xbfffce08) at
/tmp/php-4.3.1/ext/standard/var_unserializer.c:762
#4 0x40389785 in ps_srlzr_decode_php (val=0x3d36ab6c Address 0x3d36ab6c
out of bounds,
vallen=10564) at /tmp/php-4.3.1/ext/session/session.c:485
#5 0x40389a57 in php_session_decode (
val=0x8344734
logged_in|s:5:\false\;_env|O:11:\environment\:25:{s:4:\_dbc\;O:11:\patmysqld
bc\:6:{s:6:\dbhost\;s:9:\localhost\;s:6:\dbuser\;s:4:\root\;s:6:\dbpass\;s:11:\thebedotcom
\;s:6:\dbname\;s:12:\dbSupportLog\..., vallen=10564)
at /tmp/php-4.3.1/ext/session/session.c:532
#6 0x40389d0a in php_session_initialize () at
/tmp/php-4.3.1/ext/session/session.c:617
#7 0x4038ac77 in php_session_start () at
/tmp/php-4.3.1/ext/session/session.c:1012
#8 0x4038c479 in zif_session_start (ht=0, return_value=0x8344564,
this_ptr=0x0, return_value_used=0)
at /tmp/php-4.3.1/ext/session/session.c:1453
#9 0x40417816 in execute (op_array=0x833e6b4) at
/tmp/php-4.3.1/Zend/zend_execute.c:1596
#10 0x4040a6b7 in zend_execute_scripts (type=8, retval=0x0, file_count=3)
at /tmp/php-4.3.1/Zend/zend.c:864
#11 0x403e3a9e in php_execute_script (primary_file=0xb420) at
/tmp/php-4.3.1/main/main.c:1573
#12 0x4041befc in php_output_filter (f=0x8198e78, bb=0x819b478)
at /tmp/php-4.3.1/sapi/apache2filter/sapi_apache2.c:465
#13 0x08088e57 in ap_pass_brigade (next=0x8198e78, bb=0x819b048) at
util_filter.c:540
#14 0x080906a8 in default_handler (r=0x8197018) at core.c:3320
#15 0x0807c069 in ap_run_handler (r=0x8197018) at config.c:194
#16 0x0807c6d5 in ap_invoke_handler (r=0x8197018) at config.c:401
#17 0x0806be93 in ap_process_request (r=0x8197018) at http_request.c:288
#18 0x08067429 in ap_process_http_connection (c=0x81909a8) at
http_core.c:293
#19 0x08086949 in ap_run_process_connection (c=0x81909a8) at connection.c:85
---Type return to continue, or q return to quit---
#20 0x08086c5f in ap_process_connection (c=0x81909a8, csd=0x81908d8) at
connection.c:207
#21 0x0807a906 in child_main (child_num_arg=0) at prefork.c:696
#22 0x0807a9c8 in make_child (s=0x80ba178, slot=0) at prefork.c:736
#23 0x0807aae3 in startup_children (number_to_start=5) at prefork.c:808
#24 0x0807aee9 in ap_mpm_run (_pconf=0x80b7b10, plog=0x80efbf0, s=0x80ba178)
at prefork.c:1024
#25 0x08081169 in main (argc=2, argv=0xb7a4) at main.c:643
#26 0x40210082 in __libc_start_main () from /lib/i686/libc.so.6
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] PHP.ini and Virtual Hosts
A possible client wants to do this: - disable errors to browser output - redirect the error_log to a file in my own space - be able to include functions that are in a subfolder in my own space - want to increase the filesize of uploaded items Is there any way to do this without editing the server-wide php.ini file? I'm using CPanel/WHM, RedHat Linux 7.2. Thanks, Matt [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Setting session var to null now wrong?
I recently noticed that a commit late on Oct 2nd to session.c makes setting a session variable to null now incorrect and spits back a warning at the end of the page. I did use the ini workaround to get around this, but I am still wondering why it is wrong to do this. I am setting session variables via output from my database which sometime does have a null field (no global vars used). Why (not how to get around it) is doing this wrong? In addition, why should it spit out a warning on every page that uses sessions rather then just the page that tried to set the variable? Just Wondering, --Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] pear's Mail using smtp not working with cvs php
I have made a little test script to see where the hang is occuring instead
of figuring out pear today:
?php
function getmicrotime()
{
list($usec, $sec) = explode( ,microtime());
return ((float)$usec + (float)$sec);
}
$a = getmicrotime();
$fp = fsockopen('mail.haught.org', 25, $errno, $errstr);
$b = getmicrotime();
echo (Connected in .($b-$a). seconds)br;
echo '' .fgets($fp, 2048);
$c = getmicrotime();
echo (Read took .($c-$b). seconds)br;
echo ' EHLO localhost';
fwrite($fp, EHLO localhost\r\n);
$d = getmicrotime();
echo (Write took .($d-$c). seconds)br;
$numlines = 2; //number of lines to read
for ($i; $i$numlines; $i++) {
echo '' .fgets($fp, 2048);
$e = getmicrotime();
echo (Read took .($e-$d). seconds)br;
}
?
The Results:
Sept 22nd
(Connected in 0.0027800798416138 seconds)
220 haught.org ESMTP (Read took 0.0082589387893677 seconds)
EHLO localhost (Write took 0.00019896030426025 seconds)
250-haught.org (Read took 0.0012670755386353 seconds)
250-AUTH LOGIN CRAM-MD5 PLAIN (Read took 0.0014160871505737 seconds)
Sept 23rd
(Connected in 0.0024210214614868 seconds)
220 haught.org ESMTP (Read took 60.012040019035 seconds)
EHLO localhost (Write took 0.00059795379638672 seconds)
250-haught.org (Read took 60.009819984436 seconds)
250-AUTH LOGIN CRAM-MD5 PLAIN (Read took 120.02024698257 seconds)
Oct1st
(Connected in 0.010542035102844 seconds)
220 haught.org ESMTP (Read took 60.008027911186 seconds)
EHLO localhost (Write took 0.00064098834991455 seconds)
250-haught.org (Read took 60.009781002998 seconds)
250-AUTH LOGIN CRAM-MD5 PLAIN (Read took 120.02020800114 seconds)
It appears that fgets is where the problem is occuring. Its taking over a
minute. Can anyone confirm? I tried a server at mail.wvwc.edu and received
the same sort of results.
--
Matt
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] pear's Mail using smtp not working with cvs php
I am trying to use pear's Mail functions to send email to my local smtp server (FreeBSD 4.4 using qmail-smtpd), but it sort of gets hung up when communicating with the smtp server. (I also tried a 2nd smtp server with the same hang). It can take many minutes (way over the script's time out strangely) of the script just setting there going very slowly before I guess it finally times out (no errors are given). Using recordio on the smtp server shows what seems to be a very long time for php to respond to the output of the server. The communication that usually takes under a second, is taking multiple minutes. I was using an Oct 1st checkout, which failed so I backed it out a few times (3:00AM EST each day) with fresh copies to try to find where the problem was introduced: Oct 1 failed Sept 16 works Sept 25 failed Sept 21 works Sept 23 failed Sept 22 works Sept 23 failed (just to double check) I know thats not much help, and I am sorry, but I have not been able to figure out the complexities of the pear classes, so I am a little in the dark as to what function(s) are causing the problems. So could someone in the know could grab the pear scripts and check it out using the smtp option. I would file a bug report, but I have no idea what is going on here (socket problem?). Was there some major changes between the 22nd and the 23rd? -- Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] ldap_start_tls() suggestion
The documentation for ldap_start_tls does not mention that you MUST be using LDAP
protcol version 3. I spent a day looking through google and source
trying to figure out why this was happening. I finally just changed ext/ldap/ldap.c
to force LDAP protocol version 3.
After happening across bug #13278 I realized that I could have just used
ldap_set_option (I neglected to take notice of this function
probably just because I was burnt out from trying to solve this problem.)
Could ldap_start_tls be changed to force protocol version 3 (since it is required) or
could the ldap_start_tls function page be updated
to alert people that this must be done?
The changes I made to get this to work, though sort of a waste of time now that I know
of ldap_set_option are below.
-matt
--- ext/ldap/ldap.c-distThu Jul 18 15:35:00 2002
+++ ext/ldap/ldap.c Thu Jul 18 15:49:52 2002
-1997,15 +1997,22
Start TLS */
PHP_FUNCTION(ldap_start_tls)
{
+ int version = LDAP_VERSION3;
pval **link;
ldap_linkdata *ld;
-
+
if (ZEND_NUM_ARGS() != 1 || zend_get_parameters_ex(1, link) == FAILURE) {
WRONG_PARAM_COUNT;
}
ZEND_FETCH_RESOURCE(ld, ldap_linkdata *, link, -1, ldap link, le_link);
+ //try to set LDAP version to LDAP_VERSION3
+ if (ldap_set_option(ld-link, LDAP_OPT_PROTOCOL_VERSION, version) !=
+LDAP_SUCCESS) {
+ php_error(E_WARNING, LDAP: TLS: Unable to start TLS: %s,
+ ldap_err2string(_get_lderrno(ld-link)));
+ RETURN_FALSE;
+ }
if (ldap_start_tls_s(ld-link, NULL, NULL) != LDAP_SUCCESS) {
php_error(E_WARNING,LDAP: Unable to start TLS: %s,
ldap_err2string(_get_lderrno(ld-link)));
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Re: Can't fiqure out how to fix this error! HELP!
Perhaps it would help if you let everyone know what exactly was on line 13... Celebei [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... This is the error I get and I can't fiqure out how to fix it: Parse error: parse error, unexpected T_ENCAPSED_AND_WHITESPACE, expecting T_STRING or T_VARIABLE or T_NUM_STRING in d:\apache\htdocs\ViewGuestbook.inc on line 13 HELP!! -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Re: unable to pass values between pages
Hi Maria... What version of PHP are you running? If you are running 4.2.0 or newer, the problem you are seeing is by design. From version 4.2.0 onwards, request variables are no longer registered in the global scope. See http://www.php.net/release_4_2_0.php In future, [EMAIL PROTECTED] is the place to post questions about developing WITH PHP - php-dev is for the development OF PHP. Thanks, Matt Maria [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... Dear Freinds, I am trying to pass the value from one page to other. But that value is not displayed by the receiving page. I am using win 2000 server with IIS. Is there anything to be configured in php.ini file? The following is the code: user.php HTML FORM METHOD=POST ACTION=process.php INPUT TYPE=text NAME=usernameBRBR INPUT TYPE=submit value= Submit data /FORM /HTML - process.php HTML ? echo ( Welcome, . $username); ? /HTML -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] XSLT bug
On Thu, Apr 18, 2002 at 12:57:21AM -0400, Ken Egervari wrote: If I have an XML document with 2 applications nodes and the following template in the style sheet xsl:template match=application xsl:number format=a value=position() / xsl:value-of select=name / br / /xsl:template the result is b. the name d. the name2 When I think it should be a. and b .. For some reason, position() is returning the wrong result. Out of curiousity, why are you forcing the xsl:number value? With value=position() I see the same numbering (b, d) as you when I test it with xsltproc (Which doesn't use sablotron). Taking the value attribute out gives the desired a, b listing. So, perhaps it's not a bug, but instead sablotron was doing it wrong beforehand, and has been fixed. When behavior changes, it doesn't always mean that the old behavior was correct. Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] XSLT bug
On Thu, Apr 18, 2002 at 02:09:01PM -0400, Ken Egervari wrote: I know that there is a workaround (which is obvious), but I just wanted to report it. position() is supposed to return the position of the node within the current node list. So i'd expect 1,2,3,4...N. It's just strange why it would behave otherwise. As for doing it this way? I'm just playing around and noticed it. this isn't production code at all. This should've occured to me before sending the first reply, but, eh. Memory doesn't work so well so soon after waking up. It actually is reporting the correct position, if your XML is indendted. For example: root application name=a / application name=b / /root has two text nodes in it at positions 1 and 3. So position() correctly reports the application nodes as being at 2 and 4. The only workaround for that is to take all the whitespace out of your XML: rootapplication name=a /application name=b //root Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Re: Session Oddness
I might be reading it wrong, and feel free to creect me if i am but:
does the code read:
If
a cookie was sent
then
setup SID
else if a cookie wasnt sent
dont setup SID
end if
if (send_cookie) {
smart_str var = {0}; smart_str_appends(var, PS(session_name));
smart_str_appendc(var, '=');
smart_str_appends(var, PS(id));
smart_str_0(var);
REGISTER_STRING_CONSTANT(SID, var.c, 0);
} else {
REGISTER_STRING_CONSTANT(SID, empty_string, 0);
}
surely this is back to front? shouldnt SID be defined when a cookie ISNT
sent ?
Matt
On Wed, 2002-03-13 at 20:24, Sascha Schumann wrote:
The current CVS behaviour is the correct and documented one.
- Sascha Experience IRCG
http://schumann.cx/http://schumann.cx/ircg
--
Matt Allen
Technical Director
Investigation Marketplace
0413 777 771
[EMAIL PROTECTED]
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Re: Session Oddness
Ok then, no problems. The thing is, last night in stalled the latest snap and SID was not being popuulated, even though session cookies were DEFINATLY off, accoring to a phpinfo anyway. very strange. Thanks, Matt On Wed, 2002-03-13 at 23:52, Sascha Schumann wrote: On 13 Mar 2002, Matt Allen wrote: I might be reading it wrong, and feel free to creect me if i am but: Yes, you are reading it wrong. does the code read: Note the tempus of the variable send_cookie. - Sascha Experience IRCG http://schumann.cx/http://schumann.cx/ircg -- Matt Allen Technical Director Investigation Marketplace 0413 777 771 [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Session Oddness
Guys,
Somebody hsa done something screwy with sessions, ther WEREN'T defining SID
when a cookie wasnt set. heres a diff:
--- php4-200203120300/ext/session/session.c Wed Mar 13 02:51:51 2002
+++ tmp/php4-200203120300/ext/session/session.c Sat Mar 9 11:33:29 2002
-940,14 +940,7
smart_str_0(var);
REGISTER_STRING_CONSTANT(SID, var.c, 0);
} else {
-smart_str var = {0};
-
-smart_str_appends(var, PS(session_name));
-smart_str_appendc(var, '=');
-smart_str_appends(var, PS(id));
-smart_str_0(var);
-REGISTER_STRING_CONSTANT(SID, var.c, 0);
- /* REGISTER_STRING_CONSTANT(SID, empty_string, 0); */
+ REGISTER_STRING_CONSTANT(SID, empty_string, 0);
}
PS(session_status) = php_session_active;
it had me scratchin my head for a while
Matt
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Suggest: Have PHP Session cookies-enabled check use any cookie
Hi, I don't know if this has been discussed before, but I have a suggestion about how PHP checks whether cookie support is enabled. Here is my understanding of the current setup: On the first PHP page encountered, all links will be turned into GET style URLS with the session id attached because it does not yet know whether cookies are enabled. At the same time, it sends a cookie with the session id back to the client. On the second page view, the session id cookie is sent back to the server and thus the server knows cookies are enabled, and no longer has to modify the links. I've seen a fair number of people complaining about the appearance of the PHPSESSID tacked on the end of all URLs on the first page, and some people doing work-arounds like automatically redirecting back to the home page. I agree with them that it would be nice not to have the URL modification happen, but I don't like the redirect workaround either. So, I have a suggestion: To determine whether cookies are enabled, PHP could simply check to see whether *any* cookie was sent on the page request. If it were set up like this, we could then set a dummy cookie with a long lifetime on their first visit to the site. Then, on all future visits to the site, PHP would discover even on the first page that cookies are enabled. The result would be the PHPSESSID var would only be stuck on the URLs the very first time a user visits the site, and would never appear again (as long as cookies are enabled). If you fear that this is less safe than actually checking for the PHPSESSID variable, it seems that this could at least be an option that could be enabled in php.ini. What do you think? Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] Re: Suggest: Have PHP Session cookies-enabled check use any cookie
I think there has been some confusion as to what I was suggesting for people with cookies disabled. What I am suggesting would have the following effect: (1) If user has cookies disabled: PHP sessions would end up acting as before, they would convert URLs into GET style URLs (2) If user has cookies enabled: PHP sessions would act as before except would not end up rewriting URLs on the first page view. The change to the PHP code should be small: when seeing if cookies are enabled simply look for any cookie sent by the client rather than just the PHPSESSID cookie. Matt Matt Richardson [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... Hi, I don't know if this has been discussed before, but I have a suggestion about how PHP checks whether cookie support is enabled. Here is my understanding of the current setup: On the first PHP page encountered, all links will be turned into GET style URLS with the session id attached because it does not yet know whether cookies are enabled. At the same time, it sends a cookie with the session id back to the client. On the second page view, the session id cookie is sent back to the server and thus the server knows cookies are enabled, and no longer has to modify the links. I've seen a fair number of people complaining about the appearance of the PHPSESSID tacked on the end of all URLs on the first page, and some people doing work-arounds like automatically redirecting back to the home page. I agree with them that it would be nice not to have the URL modification happen, but I don't like the redirect workaround either. So, I have a suggestion: To determine whether cookies are enabled, PHP could simply check to see whether *any* cookie was sent on the page request. If it were set up like this, we could then set a dummy cookie with a long lifetime on their first visit to the site. Then, on all future visits to the site, PHP would discover even on the first page that cookies are enabled. The result would be the PHPSESSID var would only be stuck on the URLs the very first time a user visits the site, and would never appear again (as long as cookies are enabled). If you fear that this is less safe than actually checking for the PHPSESSID variable, it seems that this could at least be an option that could be enabled in php.ini. What do you think? Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-DEV] Suggest: Have PHP Session cookies-enabled check use any cookie
This is an interesting idea -- simply having the old session cookie stay around longer. On the server side, you'd probably want to keep track of the time the session was created and if it's an old session then change the session id (I don't know if that's easy or hard). I think for security it is good for each session to have its own id. Otherwise, if I read your cookies once, I could use your session indefinitely. I think it would be easier if the PHP code was modified to look for any cookie, but if that doesn't happen then I may look into doing something like this instead. Marcus Börger [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... The good idee is to set a cookie...and if so not to modify anything. What you missed is that you need not get any cookie if you did not write one yourself (but that does not affect your idea). We could simply leave an old session-cookie and reread it - that shouldn't be a security problem. Perhaps we should then take some thoughts about the security of the session cookies. However it is important that we must rewrite the urls for cases where no cookie was received. That makes your idea no solutions for peoples which claim this because for example thier JavaScript code does not work anymore when session cookies rewrite their urls. (My own hp would have the same problem...but i do not change that) marcus At 22:34 10.03.2002, you wrote: Hi, I don't know if this has been discussed before, but I have a suggestion about how PHP checks whether cookie support is enabled. Here is my understanding of the current setup: On the first PHP page encountered, all links will be turned into GET style URLS with the session id attached because it does not yet know whether cookies are enabled. At the same time, it sends a cookie with the session id back to the client. On the second page view, the session id cookie is sent back to the server and thus the server knows cookies are enabled, and no longer has to modify the links. I've seen a fair number of people complaining about the appearance of the PHPSESSID tacked on the end of all URLs on the first page, and some people doing work-arounds like automatically redirecting back to the home page. I agree with them that it would be nice not to have the URL modification happen, but I don't like the redirect workaround either. So, I have a suggestion: To determine whether cookies are enabled, PHP could simply check to see whether *any* cookie was sent on the page request. If it were set up like this, we could then set a dummy cookie with a long lifetime on their first visit to the site. Then, on all future visits to the site, PHP would discover even on the first page that cookies are enabled. The result would be the PHPSESSID var would only be stuck on the URLs the very first time a user visits the site, and would never appear again (as long as cookies are enabled). If you fear that this is less safe than actually checking for the PHPSESSID variable, it seems that this could at least be an option that could be enabled in php.ini. What do you think? Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php -- PHP Development Mailing List http://www.php.net/ To unsubscribe, visit: http://www.php.net/unsub.php
[PHP-DEV] PHP for Win32 without MySQL?
Has anyone attempted to build a version of PHP on Win32 without MySQL? In config.w32.h there is the following couple of lines: /* set to enable mysql */ #define HAVE_MYSQL 1 ... and if you set this equal to 0, the compile progresses normally until you hit php_mysql.c. (VC++ gives up after a couple of hundred errors... mostly undeclared identifier errors, which are most likely caused by a header not being included somewhere along the way.) - Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] Re: [PHP] PHP XML
On Fri, Dec 07, 2001 at 06:43:31PM -0500, benjamin yates wrote: Most people never mix XML and PHP in the same file. Let's not inconvenience the masses for the benefit of the few. -Rasmus Yet another good reason to get rid of them. :) --Jani please never get rid of them... ?php is terrible !! so much typing... and i thought php largely about rapid development :) Three extra characters? Surely there are more significant issues in development to fret over than this. if it's ever depricated then i'll definatly write a pre-processor of some kind... and i wonder why no one has yet anyway. or at least why there isn't a way to create the same as a #define macro... man would that be nice. define() Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] multiple inheritance ext
Where would such a public discussion take place? And does anybody have any idea on the time line of all of this(not impatient, just curious...)? - Matt Markus Fischer wrote: On Fri, Dec 07, 2001 at 06:26:32PM -0700, Blake Barnett wrote : http://www.zend.com/engine2/ZendEngine-2.0.pdf Yep, and don't forget to quote the important thing: Note: Multiple Inheritance functionality may or may not be a part of the Zend Engine 2.0. Whether it is added to the engine or not will be publicly discussed. Although there were discussions, there was no decision. So, saying 'will have' is still plain wrong. - Markus -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] Extension checkins?
On Fri, Nov 02, 2001 at 12:11:17AM +, Jonathan Gillette wrote: Hi. I'm aware of an extension that would fit quite nicely into the current set. What's the protocol for proposing a entry into the extension repository? Mentioning what it does may be a good start.. Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug #12565: array_map with NULL callback causes crash
From: [EMAIL PROTECTED]
Operating system: Mac OS X 10.0.4
PHP version: 4.0.6
PHP Bug Type: Reproducible crash
Bug description: array_map with NULL callback causes crash
When used with NULL as the callback function, array_map
causes a SIGBUS.
Example: this crashes
?php
$a=array(1,2,3,4);
$b=array('a','b','c','d');
$c=array_map(NULL,$a,$b);
print_r(
$c);
?
But this works:
?php
function foo($x,$y)
{
return array($x,$y);
}
$
a=ar
r
a
y(1,2,
3,4);
$
b=array('a','b','c',
'd
');
$c=array_map('foo
',
$a,$b);
print_r($c)
;
?
configure flags: --w
ith-apxs
Apache (1.3.14) l
og:
[Sat Aug
4
1
6
:
54:19 2001] [notice] child pid 18915 exit signal Bus erro
r (10)
[Sat Aug 4 16:
54:20 2001] [notice] child pid 18916 exit signal Bus erro
r (10)
[Sat Aug 4 16:
54:20 2001] [error] (22)Invalid argument: getsockname
gd
b backtrace:
Program received signal EXC_BAD_ACCESS, Could not access
memory.
0x00c1ffac in _efree (ptr=0xbfffdf08) at zend_alloc.c:226
226 REMOVE_POINTER_FROM_LIST(p);
(gdb) bt
#0 0x00c1ffac in _efree (ptr=0xbfffdf08) at
zend_alloc.c:226
#1 0x00c8fc98 in php_if_array_map (ht=3, return_value=
0x63e13c, this_ptr=0x0, return_value_used=13834456) at
array.c:2999
#2 0x00c33384 in execute (op_array=0x5a079c) at ./
zend_execute.c:1504
#3 0x00c1fbe4 in zend_execute_scripts (type=-1073750264,
file_count=3) at zend.c:752
#4 0x00c55a5c in php_execute_script (primary_file=
0xb558) at main.c:1206
#5 0x00c52b18 in apache_php_module_main (r=0xbfffdf08,
display_source_mode=0) at sapi_apache.c:89
#6 0x00c515b8 in send_php (r=0x5990b4,
display_source_mode=0, filename=0x0) at mod_php4.c:536
#7 0x00c515f4 in send_parsed_php (r=0xbfffdf08) at
mod_php4.c:547
#8 0xc700 in ap_invoke_handler ()
#9 0x00016780 in process_request_internal ()
#10 0x000167fc in ap_process_request ()
#11 0x5ec4 in child_main ()
#12 0x6080 in make_child ()
#13 0x61ec in startup_children ()
#14 0x67f0 in standalone_main ()
#15 0x7074 in main ()
#16 0x2268 in _start ()
#17 0x20a8 in start ()
#18 0x in ?? ()
--
Edit bug report at: http://bugs.php.net/?id=12565edit=1
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug #12565 Updated: array_map with NULL callback causes crash
ID: 12565
User updated by: [EMAIL PROTECTED]
Reported By: [EMAIL PROTECTED]
Status: Open
Bug Type: Reproducible crash
Operating System: Mac OS X 10.0.4
PHP Version: 4.0.6
New Comment:
I messed up the formatting, sorry
Previous Comments:
[2001-08-04 12:04:38] [EMAIL PROTECTED]
When used with NULL as the callback function, array_map
causes a SIGBUS.
Example: this crashes
?php
$a=array(1,2,3,4);
$b=array('a','b','c','d');
$c=array_map(NULL,$a,$b);
print_r(
$c);
?
But this works:
?php
function foo($x,$y)
{
return array($x,$y);
}
$
a=ar
r
a
y(1,2,
3,4);
$
b=array('a','b','c',
'd
');
$c=array_map('foo
',
$a,$b);
print_r($c)
;
?
configure flags: --w
ith-apxs
Apache (1.3.14) l
og:
[Sat Aug
4
1
6
:
54:19 2001] [notice] child pid 18915 exit signal Bus erro
r (10)
[Sat Aug 4 16:
54:20 2001] [notice] child pid 18916 exit signal Bus erro
r (10)
[Sat Aug 4 16:
54:20 2001] [error] (22)Invalid argument: getsockname
gd
b backtrace:
Program received signal EXC_BAD_ACCESS, Could not access
memory.
0x00c1ffac in _efree (ptr=0xbfffdf08) at zend_alloc.c:226
226 REMOVE_POINTER_FROM_LIST(p);
(gdb) bt
#0 0x00c1ffac in _efree (ptr=0xbfffdf08) at
zend_alloc.c:226
#1 0x00c8fc98 in php_if_array_map (ht=3, return_value=
0x63e13c, this_ptr=0x0, return_value_used=13834456) at
array.c:2999
#2 0x00c33384 in execute (op_array=0x5a079c) at ./
zend_execute.c:1504
#3 0x00c1fbe4 in zend_execute_scripts (type=-1073750264,
file_count=3) at zend.c:752
#4 0x00c55a5c in php_execute_script (primary_file=
0xb558) at main.c:1206
#5 0x00c52b18 in apache_php_module_main (r=0xbfffdf08,
display_source_mode=0) at sapi_apache.c:89
#6 0x00c515b8 in send_php (r=0x5990b4,
display_source_mode=0, filename=0x0) at mod_php4.c:536
#7 0x00c515f4 in send_parsed_php (r=0xbfffdf08) at
mod_php4.c:547
#8 0xc700 in ap_invoke_handler ()
#9 0x00016780 in process_request_internal ()
#10 0x000167fc in ap_process_request ()
#11 0x5ec4 in child_main ()
#12 0x6080 in make_child ()
#13 0x61ec in startup_children ()
#14 0x67f0 in standalone_main ()
#15 0x7074 in main ()
#16 0x2268 in _start ()
#17 0x20a8 in start ()
#18 0x in ?? ()
Edit this bug report at http://bugs.php.net/?id=12565edit=1
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug #12535: We'd like support for output buffering and here syntax
From: [EMAIL PROTECTED] Operating system: Linux PHP version: 4.0.6 PHP Bug Type: Feature/Change Request Bug description: We'd like support for output buffering and here syntax Hello, For a long time I've wished that PHP supported the here string syntax, i.e in Perl: print EOF; html body font face=Arial size=2This is great for outputting HTML/font /body /html EOF or better yet... $string = EOS; Hi, this is a multi-line string. EOS The second example is a must considering that PHP does not support output buffering, which is another feature that I believe is essential. It would be lovely to flush an output buffer and then issue an HTTP header like a cookie or a redirect. Let me just say that I love PHP, but this is one aspect in which it falls short of (forgive me) Active Server Pages. Matt Flaherty -- Edit bug report at: http://bugs.php.net/?id=12535edit=1 -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug #12535 Updated: We'd like support for output buffering and here syntax
ID: 12535 User updated by: [EMAIL PROTECTED] Reported By: [EMAIL PROTECTED] Status: Bogus Bug Type: Feature/Change Request Operating System: Linux PHP Version: 4.0.6 New Comment: Wow, thanks for the quick feedback. And sorry. I looked hard but did not find these features. I stand corrected. Previous Comments: [2001-08-02 13:55:36] [EMAIL PROTECTED] Heredoc is supported, see the manual on strings. You need three , in stead of 2, and it works. And output buffering is supported all the way since PHP 4, AFAIK... See http://www.php.net/ob_start [2001-08-02 13:53:01] [EMAIL PROTECTED] Hello, For a long time I've wished that PHP supported the here string syntax, i.e in Perl: print EOF; html body font face=Arial size=2This is great for outputting HTML/font /body /html EOF or better yet... $string = EOS; Hi, this is a multi-line string. EOS The second example is a must considering that PHP does not support output buffering, which is another feature that I believe is essential. It would be lovely to flush an output buffer and then issue an HTTP header like a cookie or a redirect. Let me just say that I love PHP, but this is one aspect in which it falls short of (forgive me) Active Server Pages. Matt Flaherty Edit this bug report at http://bugs.php.net/?id=12535edit=1 -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] [PATCH] More error handling changes
Hi All,
To carry on from the XMLRPC Error patch, here is another one that makes
PHP bork on parse errors as well as others from an INI directive called
break_on_parse.
The reason for this is if xmlrpc_errors is on, and PHP throws 2 errors,
2 xmlrpc error packets will be returned, making certain xmlrpc clients
(rightfully) spew becasue they have recieved 2 xmlrpc fault packets (xml
documents).
Cheerio,
Matta
--
Matt Allen
Technical Director
Investigation Marketplace
0413 777 771
[EMAIL PROTECTED]
Index: main.c
===
RCS file: /repository/php4/main/main.c,v
retrieving revision 1.383
diff -u -r1.383 main.c
--- main.c 23 Jul 2001 14:05:26 - 1.383
+++ main.c 25 Jul 2001 10:24:24 -
@@ -202,6 +202,7 @@
STD_PHP_INI_BOOLEAN(enable_dl,1,
PHP_INI_SYSTEM, OnUpdateBool, enable_dl,
php_core_globals, core_globals)
STD_PHP_INI_BOOLEAN(expose_php, 1,
PHP_INI_SYSTEM, OnUpdateBool, expose_php,
php_core_globals, core_globals)
STD_PHP_INI_BOOLEAN(html_errors, 1,
PHP_INI_SYSTEM, OnUpdateBool, html_errors,
php_core_globals, core_globals)
+ STD_PHP_INI_BOOLEAN(break_on_parse, 0,
+PHP_INI_SYSTEM, OnUpdateBool, break_on_parse,
+ php_core_globals, core_globals)
STD_PHP_INI_BOOLEAN(xmlrpc_errors,0,
PHP_INI_SYSTEM, OnUpdateBool, xmlrpc_errors,
php_core_globals, core_globals)
STD_PHP_INI_ENTRY(xmlrpc_error_number,0,PHP_INI_ALL,
OnUpdateInt,xmlrpc_error_number,php_core_globals,
core_globals)
STD_PHP_INI_BOOLEAN(ignore_user_abort,0,PHP_INI_ALL,
OnUpdateBool, ignore_user_abort,
php_core_globals, core_globals)
@@ -277,7 +278,7 @@
time_t error_time;
PLS_FETCH();
- /* Try to use the specified logging location. */
+
if (PG(error_log) != NULL) {
#ifdef HAVE_SYSLOG_H
if (!strcmp(PG(error_log), syslog)) {
@@ -451,7 +452,6 @@
}
/* no break - intentionally */
case E_ERROR:
- /*case E_PARSE: the parser would return 1 (failure), we can bail out
nicely */
case E_COMPILE_ERROR:
case E_USER_ERROR:
if (module_initialized) {
@@ -459,6 +459,13 @@
return;
}
break;
+ case E_PARSE:
+ if (PG(break_on_parse)) {
+ if (module_initialized) {
+ zend_bailout();
+ return;
+ }
+ }
}
/* Log if necessary */
Index: php_globals.h
===
RCS file: /repository/php4/main/php_globals.h,v
retrieving revision 1.68
diff -u -r1.68 php_globals.h
--- php_globals.h 17 Jul 2001 16:46:07 - 1.68
+++ php_globals.h 25 Jul 2001 10:24:25 -
@@ -127,7 +127,7 @@
zend_bool xmlrpc_errors;
long xmlrpc_error_number;
-
+ zend_bool break_on_parse;
zend_bool modules_activated;
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug #12354 Updated: getImageSize() fails for some JPEGs
ID: 12354 User updated by: [EMAIL PROTECTED] Reported By: [EMAIL PROTECTED] Old Status: Open Status: Closed Bug Type: GetImageSize related Operating System: linux Old PHP Version: 4.0.6 PHP Version: 4.0.5 New Comment: My initial report gave the wrong version -- this is indeed fixed in 4.0.6. Sorry for the misdirection; I thought one of my two test systems had already been upgraded. Previous Comments: [2001-07-24 22:54:07] [EMAIL PROTECTED] works fine for me with 4.0.6 release. [2001-07-24 20:12:15] [EMAIL PROTECTED] ... and checking the image you posted, it works fine for me with the current getimagesize() in CVS. And I don't think it has changed since 4.0.6 [2001-07-24 20:04:05] [EMAIL PROTECTED] Are you absolutely sure you are using PHP 4.0.6? This was a bug in 4.0.5. [2001-07-24 19:57:23] [EMAIL PROTECTED] JPEGs saved directly from some digital cameras do not contain whatever data getImageSize relies on to produce an array of size and type data. Opening such an image and re-saving (even as JPEG) within an image editor (photoshop or equivalent) will 'fix' the image, allowing getImageSize() to read it. Here's an image that causes getImageSize() to fail: http://sitefoundry.com/php/DSC00021.JPG Here's the same image after opening/saving; getImageSize() copes fine with this one: http://sitefoundry.com/php/DSC00021fixed.jpg Here's the output of ImageMagick's 'identify' utility for both images: DSC00021fixed.jpg 640x480 DirectClass 81kb JPEG 1s DSC00021.JPG[1] 640x480 DirectClass 88kb JPEG 2s Here's the output of 'file': DSC00021fixed.jpg: JPEG image data, JFIF standard DSC00021.JPG: JPEG image data Edit this bug report at http://bugs.php.net/?id=12354edit=1 -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug #12354: getImageSize() fails for some JPEGs
From: [EMAIL PROTECTED] Operating system: linux PHP version: 4.0.6 PHP Bug Type: GetImageSize related Bug description: getImageSize() fails for some JPEGs JPEGs saved directly from some digital cameras do not contain whatever data getImageSize relies on to produce an array of size and type data. Opening such an image and re-saving (even as JPEG) within an image editor (photoshop or equivalent) will 'fix' the image, allowing getImageSize() to read it. Here's an image that causes getImageSize() to fail: http://sitefoundry.com/php/DSC00021.JPG Here's the same image after opening/saving; getImageSize() copes fine with this one: http://sitefoundry.com/php/DSC00021fixed.jpg Here's the output of ImageMagick's 'identify' utility for both images: DSC00021fixed.jpg 640x480 DirectClass 81kb JPEG 1s DSC00021.JPG[1] 640x480 DirectClass 88kb JPEG 2s Here's the output of 'file': DSC00021fixed.jpg: JPEG image data, JFIF standard DSC00021.JPG: JPEG image data -- Edit bug report at: http://bugs.php.net/?id=12354edit=1 -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] [PEAR-PATCH] XMLRPC Pear Patch
Hi All,
Attached is a small patch tp pear/XML/RPC.php.
It fixes 2 things:
1. There was a print statement left in the client class
2. xml_rpc_encode had new XML_RPC_VAL instead of new XML_RPC_VALUE.
Cheerio,
Matta
--
Matt Allen
Technical Director
Investigation Marketplace
0413 777 771
[EMAIL PROTECTED]
Index: RPC.php
===
RCS file: /repository/php4/pear/XML/RPC.php,v
retrieving revision 1.10
diff -u -r1.10 RPC.php
--- RPC.php 13 Jul 2001 17:45:43 - 1.10
+++ RPC.php 22 Jul 2001 01:52:47 -
@@ -509,7 +509,7 @@
strlen($msg-payload) . \r\n\r\n .
$msg-payload;
-print($op);
+//print($op);
if (!fputs($fp, $op, strlen($op))) {
$this-errstr=Write error;
@@ -1092,7 +1092,7 @@
global $XML_RPC_Struct;
$type = gettype($php_val);
- $XML_RPC_val = new XML_RPC_val;
+ $XML_RPC_val = new XML_RPC_value;
switch($type) {
case array:
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug #12079: SEGFAULT when compiled with APXS and QTDOM
From: [EMAIL PROTECTED] Operating system: Linux RH 7.1 PHP version: 4.0CVS-2001-07-12 PHP Bug Type: DOM XML related Bug description: SEGFAULT when compiled with APXS and QTDOM httpd wont start, segfaults stright away. qtdom makes it do it. It works fine in 4.0.6RC1 configure line: ./configure --with-apxs=/www/bin/apxs --with-qtdom=yes --disable-pear --without-mysql no make errors backtrace: Starting program: /www/bin/httpd -X Program received signal SIGSEGV, Segmentation fault. 0x400fb344 in chunk_alloc (ar_ptr=0x401a4f00, nb=16) at malloc.c:2781 2781malloc.c: No such file or directory. in malloc.c (gdb) bt #0 0x400fb344 in chunk_alloc (ar_ptr=0x401a4f00, nb=16) at malloc.c:2781 #1 0x400fb13a in __libc_malloc (bytes=8) at malloc.c:2714 #2 0x08075c32 in ap_hook_register_I () at eval.c:41 #3 0x4021bc69 in ssl_io_register () from /www/libexec/libssl.so #4 0x40212ebf in ssl_hook_AddModule () from /www/libexec/libssl.so #5 0x0805508f in ap_add_module () at eval.c:41 #6 0x0805519d in ap_add_loaded_module () at eval.c:41 #7 0x08050060 in load_module () at eval.c:41 #8 0x08055650 in invoke_cmd () at eval.c:41 #9 0x08056016 in ap_handle_command () at eval.c:41 #10 0x080560a9 in ap_srm_command_loop () at eval.c:41 #11 0x08056716 in ap_process_resource_config () at eval.c:41 #12 0x0805704b in ap_read_config () at eval.c:41 #13 0x08060d88 in standalone_main () at eval.c:41 #14 0x080616cf in main () at eval.c:41 #15 0x40098177 in __libc_start_main (main=0x8061314 main, argc=2, ubp_av=0xb9dc, init=0x804f334 _init, fini=0x80984c0 _fini, rtld_fini=0x4000e184 _dl_fini, stack_end=0xb9cc) at ../sysdeps/generic/libc-start.c:129 (gdb) -- Edit bug report at: http://bugs.php.net/?id=12079edit=1 -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] XMLRPC Error Returns Patch
Hi All,
Attached is a patch that does the following:
- Allows *all* PHP Errors to be returned as a Valid XMLRPC Fault Packet (ie including
parse errors).
- Allows the user to turn these errors on and off in php.ini
- Allows the user to set the faultCode (in the XMLRPC packet) via php.ini
Why is this important?
When using XMLRPC, anything at all that gets returned from the server needs to be a
valid XMLRPC packet, either a Fault or a proper payload, if PHP dies with any sort of
error or warning the XMLRPC client dies becasue it cant grok the returned payload.
This limits PHP ever being used as an XMLRPC server. With these patches the user can
set the faultCode that gets returned to the client, so if im making an XMLRPC server
that anyone can build a client that will access it I can make a statement like If you
get a faulCode of 666, then something broke on the server, bail out nicely and let me
know.
I was originally thinking that maybe we should write a generice Errorhandler, ie let
it be set in the php.ini, the reason for this is with more and more web services
coming out (XMLRPC,SOAP, etc) each one will need a specific error format. This might
be something that could be discussed, but XMLRPC has been around the longest and alot
of these other protocols are based on it, so here it is :)
I didnt change the php.ini-dist as I didnt know the protocol, ie, since most people
wont want this feature do we leave it out?
Cheers,
Matta
--
Matt Allen
Technical Director
Investigation Marketplace
0413 777 771
[EMAIL PROTECTED]
Index: main/main.c
===
RCS file: /repository/php4/main/main.c,v
retrieving revision 1.371
diff -r1.371 main.c
204a205,206
STD_PHP_INI_BOOLEAN(xmlrpc_errors,0,
PHP_INI_SYSTEM, OnUpdateBool, xmlrpc_errors,
php_core_globals, core_globals)
STD_PHP_INI_ENTRY(xmlrpc_error_number, NULL,PHP_INI_ALL,
OnUpdateString,xmlrpc_error_number,
php_core_globals, core_globals)
397c399
char *error_format;
---
char error_format[1024];
399,401c401,411
error_format = PG(html_errors) ?
br\nb%s/b: %s in b%s/b on line
b%d/bbr\n
: \n%s: %s in %s on line %d\n;
---
if (PG(html_errors)) {
strcpy(error_format , br\nb%s/b: %s in b%s/b oni line
b%d/bbr\n);
} else if (PG(xmlrpc_errors)) {
snprintf(error_format, 1023 , ?xml
version=\1.0\?methodResponsefaultvaluestructmembernamefaultCode/namevalueint%d/int/value/membermembernamefaultString/namevaluestring%%s:%%s
in %%s on line
%%d/string/value/member/struct/value/fault/methodResponse,PG(xmlrpc_error_number));
} else {
strcpy (error_format , \n%s: %s in %s on line
%d\n);
}
Index: main/php_globals.h
===
RCS file: /repository/php4/main/php_globals.h,v
retrieving revision 1.63
diff -r1.63 php_globals.h
125a126,129
zend_bool xmlrpc_errors;
short xmlrpc_error_number;
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] Bug #11611 Updated: nl2br() outputting invalid br tags
On Fri, Jun 22, 2001 at 03:18:20PM +0200, lenar wrote: Shouldn't there be an optional flag to nl2br to change the behavior of function to what it used to be. Just there's no point in br / like tags when the rest of your code is just generating HTML compliant output, not XHTML. Ok, somebody can always use something like str_replace(\n, br, $text) to get the old functionality. What if the rest of your code is generating XHTML? It's better to assume the more strict syntax, especially considering that nobody has yet found (to my knowledge) a browser where br / doesn't work. Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] Re: [PHP-QA] 4.0.6
Wez; Is there another test suite other than the run-test.php script? (Which does run on Win32: TEST RESULT SUMMARY = Number of tests: 165 Tests skipped: 66 ( 40%) Tests failed: 22 ( 22%) Tests passed: 77 ( 78%) = Skipped 0 extensions. V:\php-4.0.5RC8 ) - Matt Wez Furlong [EMAIL PROTECTED] 05/02/01 12:02PM On 2001-05-02 15:43:57, Andi Gutmans [EMAIL PROTECTED] wrote: At 03:38 PM 5/2/2001 +0100, Wez Furlong wrote: Seriously though, win32 is particular hard to do automated testing. Maybe we could use cygwin for running the test-suite under win32 and at least be able to use standard *nix tools? The nature of PHP on Windows is that it's really best to compile it with Visual C++. It uses native threading support and ISAPI support. I don't think it's a good or feasible idea to move to cygwin. I agree about the compiling part, but cygwin isn't just a compiler - you get bash, sed, perl, awk and all those unix tools. I'm suggesting that perhaps the test suite could be run using those tools on a win32 platform. --Wez. -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Re: [PHP-QA] 4.0.6
On Wed, May 02, 2001 at 06:58:57PM +0100, James Moore wrote: I would be very against this.. to me it seems silly, the current QA Team will have to spend 90% of their time running through the (maybe hundreds) of reports rather than testing. It makes more sense to me to try and attract more people who know what they are doing to the QA Team rather than having a fairly (maybe more :)) disorderly group of people testing from people who do not really know what they are doing.. You have tens of thousands of people testing releases today. What's the difference? The big difference is during a release process is the time scale. There are likley to be more bugs in an RC as well as people reporting bugs more rigerously (As well as probably reporting lots of bogus/dup bugs, which are very tedious to trawl through). Exactly. While it would be nice to think that broadening the audience of the testing cycle would yield more bug finding, I fear it would inundate the bug database with only marginally useful bug reports. The number of bug reports that boil down to My script doesn't work! would certainly increase. Reports that extension X doesn't work, but provide no substantial information would increase, and so on. I don't see inviting this wider audience as providing enough beneficial information to justify the work of clearing away the less useful reports. Normally I test RC1 massively then if there are problems I check for them in later RC's where people have said they have been fixed (or its decided that the bug should be fixed before the release). This time this didnt work for the single reason Phanto was unresposible and commited a huge (700 line commit) to RC7 and DIDNT test it. I asked him (as I asked sascha too) to when we decided to have RC8 (I think I cc'd the list) to test his changes throughly as I would not have time due to real work. Now Phanto obviously didnt do this, maybe someone should have caught it but I feel that by not testing Phanto invalidated a lot of hard work by the rest of the team to make 4.0.6 stable. This sort of thing simply shouldn't happen, imho. If the QA team's word is to mean anything, they need to have the authority to say No, that needs to be removed. If they don't have that authority, then how can they be expected to say We have tested RC8, and it is ready to be released.? I am certainly pissed off that this has happened as a lot of people put a lot of work into making sure 4.0.5 was stable and the problem here is not the testing but the developers commiting unneeded stuff to the RC branch. I feel we should only have x people commiting to the branch and if somthing is commited as late as the COM stuff was its up to the developer to test throughly otherwise its their head on the block. I can understand why it would be frustrating to see what could amount to an undermining of the testing cycle, but I don't think shifting blame to the developers is the best solution. This will only lead to animosity over the quality of releases. What I'd like to see instead is for the QA team to own the release cycle. That is, as soon as RC's start being rolled, QA has the right to determine if a commit shouldn't go in. Without that control, it seems to me that their hands are being tied. Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] Re: [PHP-QA] 4.0.6
On Wed, May 02, 2001 at 09:03:00PM +0200, Hartmut Holzgraefe wrote: Matt McClanahan wrote: I don't see inviting this wider audience as providing enough beneficial information to justify the work of clearing away the less useful reports. right now we invite this wider audience the day we release a 'release' and again and again we end up with a .pl1 James already addressed this in the top of his post that I responed to, and quoted. Release cycles are supposed to happen in a much shorter period of time than the lifetime of an actual release. Given that releases can't expect to tackle all known bugs, the goal becomes hitting a reasonable number of bugs, such that the release cycle doesn't take too long, while still making good progress on the number of outstanding bugs. i just want to get into some process that shifts labels here making what we call 'release' now the final release candidate so that we end up with a release that deserves this name without having an .pl1 attached if you do not want to see useless bug reports you should not release at all Such bugs are inevitable, against releases. However there's no reason that such bugs have to be inevitable against release candidates when the people testing them are all (theoretically) knowledgable about how to submit a good bug report. Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] RE: [PHP-QA] RC8
Liz; http://myrile.madriver.k12.oh.us/technology/php/php-4.0.5RC8-win32-ts.zip Contains ISAPI, CGI, and Apache/Win32 builds. The only extension I've included is the MS-SQL one. - Matt Liz [EMAIL PROTECTED] 04/26/01 12:32PM If someone can compile up a windows set, and mail me the base DLLs+EXE and the MS-SQL-DLL one, I'd be eternally grateful, as it took me about 3 hours to compile it over citrix and then SMS to a machine with the stuff to do so. I'd like the CGI + ISAPI version with MS-SQL, its for IIS, but I dont use any other extensions at the moment.. I'll then trash it a bit.. Thanks Liz -- PHP Quality Assurance Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] Compiling PHP as a NLM
Davin; I've been wanting to do this for a while (compile PHP as an Apache/NetWare module). If you need help, please let me know. I don't have a NetWare compiler, but I've been needing a reason to get one. The flags that are being passed into GCC control the way it compiles (In order... All warnings, optimize to level two, include debug symbols, use /usr/nwsdk/include as a include path, don't use the standard includes, don't use the built-in functions, and pack structs). If you're not using GCC (or are you?), you'll need to set those flags in your compiler. - Matt Davin Thompson [EMAIL PROTECTED] 04/25/01 01:01PM I am currently attempting to compile PHP as an NLM. Unfortunately,my programming skills fall short of the task, but I am wading my way throughit. What I am currently stuck on is where in the php4 configure script(if this is the right place to do it at) should I add the following tohave it passed to gcc as it compiles it? CFLAGS = -Wall -O2 -g -I/usr/nwsdk/include/ -nostdinc -fno-builtin -fpack-struct Im not sure what other problems I may run into down the road, but Ihave made it this far at least.. Any help anyone could providewould be immensely appreciated Thanks, Davin Thompson [EMAIL PROTECTED] PHP Development Mailing List To unsubscribe, e-mail: [EMAIL PROTECTED] additional commands, e-mail: [EMAIL PROTECTED] contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Bug 8352 should be closed
Bug #8352 is invalid, the reporter misunderstood how MySQL's decimal field works. The behavior of mysql_field_len is correct. The SQL line 'price decimal(9,2)' calls for 9 digits, two of which are after the decimal point. http://www.php.net/bugs.php?id=8352edit=1 Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Notes on four MySQL bugs
Some more bug comments. They all look like they can be closed to me: 8634: User is expecting the default link identifier to switch back to $db after he closed $db2, which isn't how it works. Not a bug. 9455: I don't know what we're supposed to do with this one. Should be solved in a support forum (php-general, php-db, php-install...), since it's probably not a bug. 9734: Don't see what the problem is. A delete query won't return an error if there were no rows to delete. 9839: The function does check that mysql_fetch_assoc's second parameter is one of MYSQL_*, but it doesn't check the type of variable being converted to a long. Could add an (*arg2)-type == IS_LONG check, but would this be inconsistent with behavior in much of the other code that accepts this sort of type conversion? I'd just leave it as it is, since it's not hurting anything. Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] 4.0.5 RC7 and TRANS_SID under Win32
Hello... I'm banging around on PHP 4.0.5 RC7 Win32 (Apache module), and I'm having great difficultly getting the TRANS_SID feature to work. I have one web application which depends heavily on it, and it's unusable under 4.0.5RC7. Has anyone else been experiencing this difficultly? - Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] Re: [PHP-QA] RC7 is out
Zeev; Compiled on Win32 as an Apache module just fine. One tiny problem... main\php_version.h still says it's version RC6. It took me a minute to realize that I hadn't made a mistake and copied files from the wrong directory... =) PHP Version 4.0.5RC7 System Windows NT 5.0 build 2195 Build Date Apr 19 2001 Server API Apache I will bang on it tomorrow to see if it breaks. - Matt Zeev Suraski [EMAIL PROTECTED] 04/19/01 10:02PM I rolled RC7 - if there are no surprises (there'd better not be! :), it can finally go out early next week. Zeev -- Zeev Suraski [EMAIL PROTECTED] http://www.zend.com/ -- PHP Quality Assurance Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] PHP 4.0.5RC3
Bad news... Compile failure on Win32. Configuration: php4dllts - Win32 Release_TS_inline Compiling... . . . basic_functions.c V:\php-4.0.5RC3\ext\standard\basic_functions.c(2489) : error C2065: 'core_globals' : undeclared identifier V:\php-4.0.5RC3\ext\standard\basic_functions.c(2489) : error C2223: left of '-safe_mode' must point to struct/union bcmath.c . . . Error executing cl.exe. php4isapi.dll - 2 error(s), 32 warning(s) I don't have time right now to dig into it, but I'll look later. - Matt Zeev Suraski [EMAIL PROTECTED] 03/28/01 10:01AM RC3 is out - http://www.php.net/distributions/php-4.0.5RC3.tar.gz Zeev -- Zeev Suraski [EMAIL PROTECTED] CTO co-founder, Zend Technologies Ltd. http://www.zend.com/ -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] PHP 4.0.5RC3
Okay, so I have a little bit of time. The error was introduced between versions 1.315.2.1 and 1.315.2.2. If I comment out the line in question it builds (with crypt() support!) and runs as ISAPI: PHP Version 4.0.5RC3 System Windows NT 5.0 build 2195 Build Date Mar 28 2001 Server API ISAPI I'll bang on it more this evening and see if it breaks. - Matt "Matt White" [EMAIL PROTECTED] 03/28/01 02:53PM Bad news... Compile failure on Win32. Configuration: php4dllts - Win32 Release_TS_inline Compiling... . . . basic_functions.c V:\php-4.0.5RC3\ext\standard\basic_functions.c(2489) : error C2065: 'core_globals' : undeclared identifier V:\php-4.0.5RC3\ext\standard\basic_functions.c(2489) : error C2223: left of '-safe_mode' must point to struct/union bcmath.c . . . Error executing cl.exe. php4isapi.dll - 2 error(s), 32 warning(s) I don't have time right now to dig into it, but I'll look later. - Matt Zeev Suraski [EMAIL PROTECTED] 03/28/01 10:01AM RC3 is out - http://www.php.net/distributions/php-4.0.5RC3.tar.gz Zeev -- Zeev Suraski [EMAIL PROTECTED] CTO co-founder, Zend Technologies Ltd. http://www.zend.com/ -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] PHP 4.0.5RC3
Zeev; I may be jumping the gun here, but I found 4.0.5 RC4 on the CVSweb server and grabbed it. It built just fine. Looks like it's fixed. PHP Version 4.0.5RC4 System Windows NT 5.0 build 2195 Build Date Mar 28 2001 Server API ISAPI - Matt Zeev Suraski [EMAIL PROTECTED] 03/28/01 18:32 PM Should be fixed... At 22:02 28/3/2001, Matt White wrote: Okay, so I have a little bit of time. The error was introduced between versions 1.315.2.1 and 1.315.2.2. If I comment out the line in question it builds (with crypt() support!) and runs as ISAPI: PHP Version 4.0.5RC3 System Windows NT 5.0 build 2195 Build Date Mar 28 2001 Server API ISAPI I'll bang on it more this evening and see if it breaks. - Matt "Matt White" [EMAIL PROTECTED] 03/28/01 02:53PM Bad news... Compile failure on Win32. Configuration: php4dllts - Win32 Release_TS_inline Compiling... . . . basic_functions.c V:\php-4.0.5RC3\ext\standard\basic_functions.c(2489) : error C2065: 'core_globals' : undeclared identifier V:\php-4.0.5RC3\ext\standard\basic_functions.c(2489) : error C2223: left of '-safe_mode' must point to struct/union bcmath.c . . . Error executing cl.exe. php4isapi.dll - 2 error(s), 32 warning(s) I don't have time right now to dig into it, but I'll look later. - Matt Zeev Suraski [EMAIL PROTECTED] 03/28/01 10:01AM RC3 is out - http://www.php.net/distributions/php-4.0.5RC3.tar.gz Zeev -- Zeev Suraski [EMAIL PROTECTED] CTO co-founder, Zend Technologies Ltd. http://www.zend.com/ -- PHP Development Mailing List http://ww.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED] -- Zeev Suraski [EMAIL PROTECTED] CTO co-founder, Zend Technologies Ltd. http://www.zend.com/ -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
RE: [PHP-DEV] 4.0.5 RC2 Win32 Binaries?
Daniel; No, I suppose not for the RC. For the final release, yes. =) - Matt Daniel Beulshausen [EMAIL PROTECTED] 03/23/01 03:12PM At 19:30 03.04.2001 +0100, James Moore wrote: Any chance in getting the Win32 4.0.5 RC2 binaries built? The 4.0.5 RC1 binaries did not include crypt() support, so I could not fully test them on our test server. Ill pop a build up later although I dont know if I can get crypt support in there.. Ill try is crypt that crucial? it has to be patched in, see http://www.php4win.de/article.php3 most php.net win32 releases support it, so you can expect that 4.0.5 will come with it too... daniel /*-- daniel beulshausen - [EMAIL PROTECTED] using php on windows? http://www.php4win.de -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
Re: [PHP-DEV] PHP 4.0 Bug #1179 Updated: PAM support
On Sat, Feb 10, 2001 at 07:07:19PM +0100, Derick Rethans wrote: On 10 Feb 2001 [EMAIL PROTECTED] wrote: ID: 1179 Updated by: jimw Reported By: [EMAIL PROTECTED] Status: Open Bug Type: Feature/Change Request Assigned To: Comments: nobody has written this extension, apparently. refiled against 4.0. There is a PAM extension, but it's not in the main CVS. I don't have the URL anymore too, but we did some things with it before. There's a set of PAM bindings, php_pam, available. However, it doesn't yet support methods of authentication that require root privilages. http://freshmeat.net/projects/php_pam/ ftp://ftp.netexpress.net/pub/pam/php_pam.newest.tgz Matt -- PHP Development Mailing List http://www.php.net/ To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
[PHP-DEV] PHP 4.0 Bug #8689: pg_Connect() seems to do some type of caching that doesn't quite work
From: [EMAIL PROTECTED]
Operating system: Linux (2.2.7)
PHP version: 4.0.2
PHP Bug Type: PostgreSQL related
Bug description: pg_Connect() seems to do some type of caching that doesn't quite work
?
function dummy($x) {
$database = pg_Connect("dbname=template1 port=5432");
$querystring = "set datestyle to 'ISO'";
$result = pg_exec ($database, $querystring);
echo "database = $database br";
// note 1: by commenting out the following line the bug disappears
$x = pg_NumRows($result);
// note 2: by commenting out the following line the bug disappears
$y = pg_NumRows($result);
// note 3: by un-commenting following line the bug disappears
// pg_close($database);
return 0;
}
// note 4: if the return-value is ignored by commenting out the following
// line (only the first line: $x =)
$x =
dummy("12345");
echo "database = $database br";
// note 5: a change in the connect-string, like an additional space
// at the end of the connect-string, makes the bug disappear
$database = pg_Connect("dbname=template1 port=5432");
$querystring = "set datestyle to 'ISO'";
$result = pg_exec ($database, $querystring);
echo "database = $database br";
?
Comment:
You can easily reproduce the bug by copy/paste-ing the above code in a file
and check it out in a browser.
By following the instruction in any one (try one at a time!) of the notes 1 to 5 in
the above source you can get the bug to dissappear, which doesn't make a lot
of sense in most cases (except from note 3).
It seems to me, that pg_Connect() decides it has that connection already made
(based on a cmpstr() of the two connect-strings), but the database-ressource
has actually gone out of scope (but only under obscure circumstances, see notes...).
Therefore you will get an error like this:
Warning: 1 is not a valid PostgreSQL link resource in
/www/auszweiterhand/docs/connect_test.php on line 33
Even though it should be valid at this point.
Thanks for looking into it!
regards
Matt
--
Edit Bug report at: http://bugs.php.net/?id=8689edit=1
--
PHP Development Mailing List http://www.php.net/
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
