Slashdot Mirror
PHP 5 Beta 1
Sterling Hughes writes "The PHP development community is proud to announce the release of PHP 5 Beta 1. Downloads are available in both source and binary form (for Windows users). A full list of changes is available in the ChangeLog. Some of the new features include much improved OO support, completely revamped XML support, and the default inclusion of SQLite."
Top Most Bizarre/Disturbing Error Messages
Take a look at the OO changes page. The syntax seems to be converging with Java. I find this amusing in some ways.
And what about apache 2? Up until now, they recommend that users stick with apache 1 if they want to use PHP as a module.
Does PHP5 cache compiled scripts yet? That's where ASP gets it's speed from.
It's still lumped into a great big inconsistent namespace
... pool database connections maybe? no, mysql_pconnect() doesn't count. Oh, and what's with this SQLite thing? had a bit of a fallout with the MySQL team?) ... because that kinda defeats the point of PHP anyway.
..... (x47) really want to like Python, but it's not re-entrant and has a big interpreter which makes its threading capabilities into nothing more than a silly joke (and last I checked, efforts to rectify the situation died back in, what, 1997?), so yeah I admit I use PHP for quite a few web devel jobs. But just because everything else sucks more, doesn't mean PHP doesn't suck any less.
It's still made by a for-profit company who hobble the product in order to not cut into their profit margins too much (Hello? Zend Cache? Optimiser? Compiler? Everything's free in PHP-Land, for a small fee in PHP-Land...)
I don't mind so much the fact that you can't have servlet-like objects which handle entire sections of your URLspace (as opposed to one URL -- how very un-spider-friendly. Most choke on a ? in a URL and rightly so) and remain persistent (allowing you to do funky stuff like
But come on. Fellas. PHP: Hypertext Preprocessor is the name. Not PHP: Application Server. If those first two issues were fixed it might actually make a seriously powerful hypertext preprocessor. That's something it's reasonably good at. But at the moment it's some sort of bastard preprocessing language run amok that people use to write whole web applications with and other stuff Nature never intended. Perl's got an awful syntax and a total lack of convention (and mod_perl is really byzantine), and I really really really really
I'm not even sure what my point is anymore. But, I think what I was trying to say was this isn't much. Same stuff is true of PHP as has always been true of it... wake me up when they get round to PHP6. An earlier rant I made comparing Perl to PHP (I think I preferred Perl back then) is here. The extended comment history is pretty much the only reason I got a subscription and to be fair I think it's worth the money.
God damn!
I'm half-way through re-writing our website to remove our dependance on the depreciated register_global_variables stuff which I learned to live with and then was told was a stupid idea.
I'm hoping they don't break too many other things 'cause being a php beginner, the whole process is a bit of a nightmare.
MySQL isn't bundled with it, but you can easily add it yourself when compiling.
Compiling?
Compiling PHP for Windows requires the Microsoft Visual C++ compiler version 6.0 or later.
The Microsoft Visual C++ optimizing compiler version 6.0 or later is available only from Microsoft as part of Microsoft Visual Studio .NET, which costs $1,079 for one non-academic seat. (Microsoft no longer sells a Visual C++ optimizing compiler separately.)
Some people are bound to bring up the $109 Microsoft Visual C++ Learning Edition, but 1. the EULA attached to its library probably does not permit distribution of generated binaries nor public performance (i.e. use on a public web site) of generated binaries, and 2. because it does not have an optimizer, the speed of generated binaries is closer to that of an interpreted program than to that of a compiled program.
If I had any spare time, I'd fix this by porting the build to MinGW.
Will I retire or break 10K?
I'm a PHP programmer and I've begun to recognize a pattern in software: when will we see the end?
Truly, what more can you add to PHP? Why do most software packages continue to add features without actualy providing a subjective goal to strive for?
For example, in the world of Microsoft(R) Windows we see the same operating system have plastered above it "Where do you want to go tomorrow" and above all "n% faster than previous, more stable, etc." When will the goal of a products feature-list finally be met?
I know Perl5 accomplished its goals, and then they had an {ap-if-in-ee} to add the RegEx in yet another release of Perl titled Perl6. When will they ever make a product that has a goal? I don't call this competition...I call it beating a dead horse from its grave, like how Intel puts rocket boosters on its crumy brick CPU architecture (Pentium Pro) and adds some more features.
Look at netBSD; it isn't dying, it's still working on its number-one goal: security.
Linux is the same way; it stated from an original design and now is being extended. Am I sounding like I expect a new feature to be a new product? I don't think so... GNU/HURD, of which I know many people are skeptics unto, is builind upon its goals of being a Micro Kernel and add some. What if Linux all-of-a-sudden wanted to become a Micro Kernel? What if Microsoft(R) Windows(TM) all-of-a-sudden wanted to become a Micro Kernel?
The software names, after huge changes to extend their capabilites, are becoming misleading! If you take out a RedHat Linux 5.2 system and compare it with RedHat Linux 8.x, you can't say they are both similar Linux; they are completly different! Sure, it's like comparing apples with oranges on the old and new RedHat, but Kernel-wise the latest Linux Kernel may look completly different and have completly different goals and features than what was hoped for in the verry early Linux kernels. Shouldn't to cause less confusion and more inspiration, they leave the Linux name behind with the old design and all the new stuff that would completly change Linux's software (mechanical?) appearance become known as the new project Herring(TM). I anticipate most of you will reply with "That's why we have VERSION NUMBERS", but hey? You're missing the point.
Microsoft(R)'s Windows(TM) name isn't describing the project's code name, yet the product's retail names are somthing like 95, 98, NT, 2000, XP. Well, it is honest to say that the NT and 2000 products are similar, 95 and 98 products are similar, and the XP product isn't quite similar to 95 and 98 and 2000, but then there was the fiasco that all the 9x users went out to buy product Windows(TM) 2000 and found they had been tricked by Microsoft that 2000 would be like a 95 or 98. Anyone see my point?
As for PHP, and perhaps Perl... Does anyone think they should continue calling those products by their initial names AFTER the programming syntax and methodology becomes completly different or non-compatible than they were first designed?
I'm looking forward to some intelligent answers. Thanks and I'll be a PHP and Perl programmer for a long time to come. d:-)
To the best of my recollection, there isn't much else that's not backwards-compatible. Even where functions have been renamed (e.g. socket_get_status), the old function names still work, and while deprecated, they don't seem to be going anywhere soon. I have no trouble digging up stuff I wrote back in '99 or 2000 and getting it to work under PHP 4.3; though I do have to enable register_globals in those cases.
Only problem I ever ran into with PHP where stuff quit working after an upgrade was on a test Apache2 server. It turned out to be a bug related to posted form data. I wouldn't use Apache2/PHP on a production server yet anyway, though; and 1.3.27 still gets the job done. I haven't had time to play with PHP5 yet, so I'm not sure what the differences are in that version.
I agree at the surprising number of hosts who simply haven't updated, though. There are a lot of hosts still running 4.1x, and even (yikes) 4.06, who just won't upgrade for whatever reason. I do most of my coding these days on 4.2 or 4.3, and have run into plenty of belligerent hosts who refuse to upgrade from a two-year-old release. Typically I just have my clients move to a better host; the providers who don't stay reasonably with the times will eventually figure out that it's hurting their bottom line.
"BSD: Free as in speech. Linux: Free as in beer. Windows 10: Free as in herpes." --Man On Pink Corner in #52607549.
The will be an interesting battle. JSP and PHP are now broadly identical in syntax and OO implementation. Who'll win? PHP is OS, but JSP has a huge amount of support from corporations.
I'm betting on JSP
((lambda x ((x))) (lambda x ((x))))
I have always wanted PHP to have an XSLT based transformation pipeline, similar to what Apache Cocoon has. I would like to be able to write source documents in XML, arbitrarily transform XML into HTML and PHP, interpret the PHP, possibly transform again and finally output the results. I know I could build this top-down from within a single PHP document, but I'd like this pipeline to be a part of the mod_php distribution, with the .XML file extension mapped to the pipeline processor.
Anyone wanna help me build it?
std::disclaimer<std::legalese> sig=new std::disclaimer; sig->dump(); delete sig;
Considering that I've been a Linux/Apache/MySQL/PHP developer since they each were available, it's just disgusting that they're making it harder to support MySQL at a time when LAMP is a recognized contender against Microsoft. I am not going to fscking rebuild a bunch of sites' dbs and recode MySQL-specific PHP code just because the GPL gives these guys a rash. Certainly when PHP was under Rasmus's authorship and control, he was never this sort of jerk.
I'm not saying I can't take the trouble to link in MySQL libraries, just that there's no good excuse for the PHP folks to make me - and thousands of others - go to this trouble. They could, if nothing else, distribute their nonGPL PHP, plus a GPL kit that adds MySQL support, if they're too scared that the GPL will give them cooties.
"with their freedom lost all virtue lose" - Milton
Hey HeadDown. You're right to take the guy to task, s/he made some crazy comments. But I can at least partially substantiate speed issues. Back in 2000, I worked with Sabeer Bhatia (the Hotmail guy) on a startup called Arzoo. Our product (a Web site similar to epinions) was almost 100% Java, except for a bit of Perl for screen-scraping and searches. But anyway, it was slow -- first with Tomcat, then with JRun. At one point, we gave a private preview to 1,000 journalists. They didn't even visit the site all at once, they trickled in over the course of 3 days or so. Just that was enough to hammer the site. We ended up running cron jobs that would reboot the farm, round-robin, just to solve memory issues and instability.
Now, you can say, well, that was 2000! Try it now! OK. At SST, we have a team that is using Tomcat now. Although the instability is gone, the speed is still an issue -- they have wait screens as you click through the app. My team is working with PHP, and has no wait screens, and no need for them (with 1 exception). Our pages are actually more computationally stressful than the Java stuff, yet PHP is delivering the result to the browser faster.
As a final point, you might suggest that the teams I've worked with do not understand Java or how to run it well. It's no skin off my back if you make that argument -- it's not me doing this stuff, so no blow to my ego. But I think working with 2 different teams over the course of 3 years says something. Perhaps, at the very least, if Java really can handle a bigger load, it is so difficult to tune that mere mortals would do better with PHP.
My Greasemonkey scripts for Digg &
That's kind of funny, considering the Linux users I know can all get things done. The Windows users I know quite often have to ask the Linux users how to get things done.
Feel free to mod me "-1 - Angry Jerk".
Oracle has has ODBC support for as long as I can remember. Also, if you're using Windows, it's better to use OLE DB since you then remove an entire layer for better efficiency.
And, SQL isn't anything close to all that a real database (ie: Oracle, DB2, etc.) has to offer. Amateur programmers often completely overlook an entire layer: stored procedures. Mature databases (ie: Oracle as opposed to MySQL) has a very robust native programming language in which the code is compiled and cached in the database, making it often the best way to add business logic. (Note: whitepapers say to use some bullshit middle tier like CORBA or COM which just doesn't fucking scale, no matter what the "experts" say). The best thing that an amateur developer can understand is that something like Oracle or DB2 is a very very mature product which screams compared to any middle tier bullshit. Learn PL/SQL and REALLY build your resume.
Note: This post was written by a former Oracle Certified Developer who got paid MAD cash in the day to write web apps.