| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Comments:
<0> is there any way to check that data has been sent to the browser, not including headers, without starting a buffer?
<1> at what other point in history have you been able to express _completely_ the wrong emotion at the slip of a finger?!
<2> argh, why does file_exists($filename) where file name is Something 'Jojgoj.extension fail?
<3> dools: Not many ppl on here are carving Kanji, not much is needs to remember =)
<2> It works with special character free filenames
<1> Bungopolis: what data do you mean?
<0> dools page content
<3> duanear try specifying the FULL path maybe?
<1> duanear: use pastebin.com to put an example up... and what do you mean by fail? what error do you gt?
<1> Bungopolis: what type of page content?
<0> dools if it's html the first part of the content would be <html>
<2> Jymm: I ddo use fullpath. The example was just to illustrate
<2> By fail I mean it returns false even though the file does exist.
<3> duanear: And we are mind readers?
<0> dools web server sends the headers, then the "content". there's a headers_sent() function, of course, but I wonder if there's a way to check if any content has been sent by the current page
<0> is there a default buffer I can loo at?
<0> *look
<2> This works: ls /path/"filename"
<4> I'm planning on building a script to troubleshoot network problems, can someone give me some ideas on how to implement this? I'm gonna have a series of yes or no questions and after each question i want the script to spit out a sentence on the page depending on the answer.
<2> but file_exists($path.'/'.$filename) returns false
<2> :(
<1> Bungopolis: what type of request? POST? GET?
<0> dools I'm talking about data the server is sending, not the browser
<0> I want to check if something prior to the check has sent some content
<1> Bungopolis: what is the error you are getting?
<2> I am so stumped.
<0> dools I am not getting an error
<2> Should I be escaping something?
<1> Bungopolis: okay, what is it you are trying to achieve (i mean apart from finding out if content has been sent)
<1> duanear: put a test case in pastebin
<0> nothing in particular
<1> Bungopolis: well then you're fine!!
<0> echo "something"; if ( content_has_been_sent() ) // true
<1> sounds like homework to me ...
<0> <?php if ( content_has_been_sent() ) // false ?>
<0> no it's not "homework"
<0> i just dont see the point in explaining what I want to do
<0> ob_start(); echo "something"; iif (ob_get_content != '') // true
<0> thats basically what I want to do, but I dont want to have to start a buffer first
<1> Bungopolis: because there clearly is not a content_has_been_sent function, so if you give me an idea of what you are trying to achive then it better enables me to lend a helping hand
<0> I'm presuming its impossible, I'm wondering if you know something different
<2> dools: http://pastebin.com/645608
<0> alright I have a cl*** which generates javascript that uses xml http request. I want the same cl*** to transparantly act as the processor for the xml http request, but if the user of the cl*** has already sent some content, the xml sent back by the processor is obviously going to be invalid
<0> so I want to check that the user is not sending content first, if they have I can throw an error about it
<0> does that make sense
<0> by "user" i meant user of the cl***
<0> not the end-user
<0> I can use headers_sent, because in most cases headers_sent == true means that content has been sent, since most of the time we dont need to send headers explicitally
<0> but in theory the user could be sending a header and still want to have the cl*** acting as the xml http request processor
<0> so that's not a good enough proxy
<0> but if you're right, dools, that there is absolutely no way to check that content has been sent to the browser without starting an output buffer first, then I dont need any further help
<1> Bungopolis: i presume that you only are averse to starting an output buffer at the very beginning of your script? ie. if you had a solution using ob_start local to the function that response to the xml http request that would be okay?
<5> what's the best way to convert serialized byte array to a string?
<0> dools I am averse to it because tihs is a cl*** I am releasing publically and that's an unintuitive for a user of the cl*** to do
<0> and anyway, if I am able to tell the user they need to start an ouptut buffer for me, i might as well just tell them to make sure the cl*** processes xml http request at the beginning of the script
<1> Bungopolis: ob_start() can be used inside another ob_start(), so what i'm saying is that if ob_start(); ob_get_content(); ob_end_clean(); were used entirely locally within that cl*** it would present no problems in terms of the interfae
<1> interface
<0> dools thats not the question, the question is has the user ALREADY sent some content, becuase if they have the cl*** CANNOT send the xml http request response
<0> it has to be the only data that is sent
<1> Bungopolis: now this is a pretty big hack, but if you call session_start() and content output has already begun then you get a notice, so if you, locally, call ob_start(); session_start(); $result = ob_get_contents(); ob_end_clean(); then parse $result, you will see a notice that session_start() could not send the sesssion cache limiter because output has already begun
<6> i have 4.4.2 set up to run as CGI via .htaccess. unfortunately it seems that the server will only start one php.cgi process at a time. if another request comes in, it blocks.
<0> ah I see
<1> Bungopolis: or perhaps session_cache_limiter() is better
<0> that's interesting
<6> any way to add more concurrency?
<1> Bungopolis: play around with it, but i think session_cache_limiter() will cause the notice as well
<0> so at least internally php does actually keep track of whether content has been sent
<1> Bungopolis: it realise on the user having E_NOTICE in their error_reporting though
<0> yeah
<1> Bungopolis: yeah, i'd put in a feature request if you get it working
<0> cool, thanks a lot I'll play around with that
<1> Bungopolis: np
<2> Nevermind, magic quotes was on
<1> man!! where is aidan when you want to rub his nose in it??
<1> duanear: there's your first mistake :-P
<2> I've never had it on... always used php.ini-recommended
<2> but i installed through debian this time
<2> and its default php.ini had it on
<1> duanear: there's your second mistake!!
<2> ?
<1> i'm a freebsd ports diehard, but i always install php manually
<2> when i need turckmmcache i do it manually
<2> but debian is so nice
<2> :P
<6> i guess my problem is actually with apache, i'll try there
<0> dools another solution might be to try something like header('') because header will also fail if content has been sent
<1> Bungopolis: ahh, yeah, there you go
<0> but I dont know what the implication of sending an empty header is
<1> Bungopolis: if it fails, surely nothing has been sent
<0> right
<7> what's wrong with headers_sent() ?
<7> oh, didn't scroll up
<1> GarethAdams: he's checking for post headers content
<0> garethadams becuase I believe headers_sent() only tells you if a single header has been sent, and in my case I have nothing against a user sending headers, only to them sending content after the headers are sent
<7> Bungopolis: no headers are sent until content starts. When any content is sent then all queued headers are sent together
<0> garethadams are you positive? (i'm not questioning, just confirming)
<7> not 100%
<0> ok, well I'll test it
<7> that's what I always ***umed
<0> I think you might be right
<0> I cant really remember why I thought I couldnt use headers_sent before
<7> headers_sent in the manual says "You can't add any more header lines using the header() function once the header block has already been sent."
<0> and the only way to send headers is to send some content, right?
<0> or to exit()
<0> presumably
<7> and the header() function in the manual says "The optional replace parameter indicates whether the header should replace a previous similar header, or add a second header of the same type."
<7> indicating that they aren't send immediately
<7> also "Note: The HTTP status header line will always be the first sent to the client, regardless of the actual header() call being the first or not."
<0> right
<0> it seems like headers_sent isnt very reliable, i'm running on windows right now and header('Content-type: text/plain'); echo 'test'; if (headers_sent()) echo 'YES'; else echo 'NO'; outputs testNO (indicating it thinks headers have not been sent)
<1> Bungopolis: does a call to header() fail?
<8> (isset($task) ? @$task : null) <-- what the sign '@' means in this case?
<1> !+@
<8> ?
<1> !tell xmlaxis about @
<3> !+@
<9> [@] The PHP error control operator that suppresses errors (@foo()), see: http://php.net/operators.errorcontrol - Don't use it. Think I'm wrong? Ask me about "GO".
<1> ahh, i see
<1> you'll have to make me an op then Jymmm :-)
<10> does anyone know if there are any cl***es available for PHP5 that will allow you to do remoting similar to .NET remoting?
<8> thanks
<1> Psykick: what type of remoting?
<0> dools header() is actually a void function so it's always false, I just realized
<0> so that wouldnt have been a useful test anywya
<1> Bungopolis: d'oh!! what about session_cache_limiter()?
<0> havent tried that yet, still trying to figure out what the hell is going on with headers_sent()
<11> $get_something = $_GET['p_time']; can I initiate that at the beginning of the cl***?
<1> Frosh: do it in the constructor
<11> ok
<1> Frosh: you don't have to, but it's good practice
<11> put it before the constructor?
<11> or put it in the constructor?
<1> in
<1> Frosh: i always initialise all variables in the constructor
<1> Frosh: member variables i should say
<0> dools: why?
<1> Bungopolis: makes more sense when you think about extending the cl***. only member vars i would initialise outside of the constructor are static members
<11> question, maybe I'm doing it wrong
<11> $_GET[];
<1> Frosh: what is that?
<11> if I call that and there is no $_GET
<11> I get error right
<11> no $_get variable
<1> Frosh: the only time there is no $_GET is when the script is called from the command line
<0> dools but those members vars could still be replaced in a cl*** extension
<1> Frosh: in which case count($argv) will return true
<1> Bungopolis: that's the idea
<11> index.php?id=3 and index.php
<11> in the second one $_get would echo error about missing $get
<1> Bungopolis: hold on... no, that's not the idea!!
Return to
#php
or
Go to some related
logs:
posh-lia movie
modify routing table gentoo
proftpd ubuntu connection refused
gaim plugin autoreply
y2mpsh
mlingojones proxy
#css
alter all tables mysql
#debian
quicktables for slackware
|
|