Help Center Live Community

Development => Core System Development => Topic started by: HCL Admin on June 15, 2007, 12:22:36 AM

Title: Progress on Core.
Post by: HCL Admin on June 15, 2007, 12:22:36 AM
Without surprise, the current 2.1.2 core is functioning well on PHP 5.2.3, and MySQL 5.0.41  However the eTicket module that is to be included in the next version unfortunately is having problems while interfaced (but not stand alone).  At this point it appears that something in PHP 5 is breaking the core system during the interface.  Once this issue is resolved, we will be ready for either a beta or a RC (depending on what all needs to be done to get the issue resolved.)

Now the easy question is why worry about it.  Well, two things, a problem of this nature could (and probably will) affect other modules, and also, if there is an issue, we want track the sucker down and kill it ASAP.

It appears that for some reason PHP 4 is allowing headers to be redefined and PHP 5 is not.  Strictly speaking, that's probably the best way to handle things, however in our situation there really is no "fix" for it, other then to comment the header lines out of the integrated application. 

While I hate this idea, I'm thinking that it would help integration for eTicket for them to make a "switch" to turn off these header lines, as I was able to cause the same error integrating another app to eTicket and getting the same exact error under PHP 5.2.2

I have yet to try this with a PHP 5.1.x platform, as I don't have any PHP5 servers running less then 5.2.

Title: Re: Progress on Core.
Post by: HCL Admin on June 17, 2007, 03:54:22 AM
Just an update.  Core was ready to go when I was made aware of a couple of bugs, one is again preventing users from being seen, the second is much more irritating, and I think it's been there for at least a year.  It seems that the initiate chat is not working.  Because the initiate bug appeared to be the most daunting, I decided to tackle that first.

The initiate bug is odd, because everything seems to work, yet the window is not opened.  To boot we had a storm pass through and I had all my working computers off for most of the day (storm is an understatement, we have tornado watches going until 10pm tonight...)  So I haven't accomplished much.

The other bug I think may be related to the bot killer.  Just a hunch, but until the initiate bug is squashed, it's on the back burner.

The long and short of it is that unless these two bugs are squashed in our test environments, I've postponed any betas for the next few days.

Title: Re: Progress on Core.
Post by: HCL Admin on June 22, 2007, 06:58:28 AM
I should probably update this.  At this point the system is running smooth, it looks like both bugs are squashed.  One needed a timeout adjustment, the other appears to be an issue with IE 6 & 7, which can be worked around, but until MS decides to fix that browser, it's one of those "that's life" things...

We found a couple more issues, and resolved them, one being an install issue for Windows servers, another being an upgrade issue that has also been resolved.

While I'm going to put out another test in a few days, I would like to wait and see if any additional issues crop up.  I'd like to see 2.1.3 be the end of the 2.1 line.