This was one of our first tries to solve the problem. Since session_start() is called in the index.php containing the frameset declaration, php.exe accesses the session file from both frames BEFORE processing any code out of it. So a delay contruct will only be processed when session file access is done :( Greetings Stephan Weiher ""Michael Southworth"" <mikesouthworth@xxxxxxxxx> schrieb im Newsbeitrag news:cbda40880802120650w3c87f1edy116647496d7b438a@xxxxxxxxxxxxxxxxx >I should preface this by saying it has been a while since I have used php > extensively, but my suggestion would be to try putting in a delay at the > beginning of your PHPfile, before the session_start() command (or > whichever > method you choose to begin your session). As far as I can recall, the > session needs to be initialized before any output is sent to the browser, > but this does not mean you cannot execute a php code delay before you > start > the session. This may give you enough time for the other PHP request to > complete, freeing up the required file. > > A potentially more robust, but equally untested option, would be to check > the file itself before initializing the session. The session_id will be > available in the super variables or via session_id(). Since the file names > for a given session are based predictably on this id, you could > potentially > try to do a read access on the file. If this fails, poll and try again or > timeout and fail gracefully. I am not very familiar with how the > permissions > on this would play out though. > > Those are two suggestions I could think of, hope that helps, > -Michael Southworth > > On Feb 13, 2008 8:15 AM, Stephan Weiher <stephan@xxxxxxxxxxxx> wrote: > >> Thanks for the answer but this won't help us anyway. We cannot enable >> javascript on any computer in the company's network to make this issue >> going >> away. I'll need a way to delay the requests for one of the two frames >> (defined in the frameset) for a tiny amount of time to prevent two >> php.exe >> instances from accessing the session file at the same time. But I don't >> know >> how to do so. Javascript is no option for us and as far as I know HTML >> has >> no such mechanism. So I need another way. Perhaps sth. that delays the >> call >> of php.exe - but I/O must work on this way :) >> >> Greetings >> Stephan Weiher >> >> >> >> ""Jacob Kruger"" <jacobk@xxxxxxxxxxx> schrieb im Newsbeitrag >> news:004501c86ce6$9da6a080$6401a8c0@xxxxxxxxx >> >A 'corny' workaround might be to get something like the nav frame in >> >fact >> >to >> > then load the main frame using something like JavaScript - just a >> > thought - >> > since it would in fact first have rendered before the main frame was >> > loaded. >> > >> > You could then also put something like a 'waiting...' notice/string in >> the >> > main frame in case one or two people's machines in fact rendered/loaded >> > the >> > two frames in the wrong order. >> > >> > Stay well >> > >> > Jacob Kruger >> > Blind Biker >> > Skype: BlindZA >> > '...Fate had broken his body, but not his spirit...' >> > >> > ----- Original Message ----- >> > From: "Stephan Weiher" < >> > stephan@xxxxxxxxxxxx> >> > To: < >> > php-windows@xxxxxxxxxxxxx> >> > Sent: Tuesday, February 12, 2008 12:06 PM >> > Subject: Concurrency issue on session file on harddisk >> >> -- >> PHP Windows Mailing List (http://www.php.net/) >> To unsubscribe, visit: http://www.php.net/unsub.php >> >> > -- PHP Windows Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
