EVE Technology Lab

dang you poor php people, I found what I need (I think)

http://php-java-bridge.sourceforge.net/pjb/how_it_works.php is needed to get it up on TomCat.

For anyone following what I'm up, briefly reading this, it sounds Like java can directly call php and php can directly call java ....

cool beans. php might not have to mess with rabbitmq, however that can be available as well for apps that want to be async verses pulling/polling driven.

love it ... and with that advice ... waaa laaaa I have PHP running on Tomcat ... woot


Had to drop tomcat back down a version however, the master demanded it.

I worked at Novell during its hayday years, In 95 I came kicking and screaming from years of procedural C, at the time working on the Novel Directory Services (NDS). I got pulled from the C side and worked closely with Sun as they developed JNDI. I wrote java JNDI wrappers around NDS and LDAP at that time. I've been writing java ever since.

But Steve some people like slow overly engineered solutions to their problems

People do tend to only play in their own sandbox ... look at my naivety in the web-tier sandbox as exposed in this forum time and again since I showed up. I've never been in this neighborhood before!

For the last eight years of my life, I've worked for Panasonic in a very small R&D group of about 15 engineers. The following link might give a feel for the sandbox I live in. There is little to no web-tier. Last year we gave this core to the world, which I'm very proud of.

https://opendof.org/

One thing about the java silo/sandbox is I can bridge over to my neighbors PHP sandbox and **** in it. They can not do the same.

Serious question, as I try and understand your world. The tomcat/apache2 discussion highlights the question. tomcat is a simple application server which provides a servlet container. I think in posts here I've tried to liken this to a linux daemon, i.e. a background processes/application that is always running.

Perhaps mistakenly I'm viewing web-tier scripting languages to be limited to a streamed pipe environment where they get an inputstream (port 80) that they can apply complex filtering against to produce an output stream. I assume they are limited to a single thread, which they do not own and must do everything serial as they pipe from filter to filter. They are limited, static as it were to responding only to the stream.

With this single threaded filtered pipe stream, verses an open container, I'm wondering how PHP might be able to utilize RabbitMQ.

Side note, if you look way back at the history of Queue's (mailboxes), IBM led the way many moons ago with a queue I cannot remember the name of, which became the core Remote Procedure Call (RPC) mechanism between the WebSphere containers. JBOSS used the java open source java equivalent of Message Beans. Note containers can all live on the same box, or be split over several boxes (i.e. tiers).

Anyway, Queues have long been used for not only "mailboxing" purposes but also used to implement RPC request/reply type api's but also event/async push api's. It is these events that I'm wondering if we can leverage with PHP was well as request/reply api's?

For example, I can determine the crest endpoint is down (I use the Time endpoint as a heartbeat). So within 10 seconds of it going down, I can fire an event saying so. Same for the xml-api endpoint and for the DB (which could be on a different tier).

Can a PHP page, monitor for events and update with no other stream activity tickling the input stream pipe? I assume this might be were you utilize the likes for Cron jobs?

Yep PHP is just a filter Understand I'm not really into front-end user web design myself so this will probably be missing some things but should cover most of the things it does while being 'just a filter'

• Browser connect to URL and web server hands over serialize string of headers, data, and other misc stuff as a request to PHP.
• PHP tries to add a little order and break things up into actual useful information before starts up a front end controller(class) to continue on with processing the request.
• The front end controller creates a request object(class) that it uses to exam URL to determine which of possible hundreds, but usually only a few dozen other back end controllers(also classes) that have been registered with it should be used to handle the request.
• It pass off the request object to the back end where the request is usually farther processed to determine what other resources are need and do 'simple' things like connect to a database instance (local or remotely) to get data domain related objects need to 'filter' the request.
• The back end controller than start assembling a response object(class) which may require on the fly generating things like catcha graphic if the request is for a page with a web form, collect other resources as need to fulfill the request, pass over part or all of the domain data to a templating system(class) to plug the data into a template. After managing the one to possible dozens of part need for the request it pass back the response object to the front end
• the front end controller may do some additional processing before sending the simple 'filtered' stuff back to the web server having serialize all the headers into a nuch simpler form that the web server can understand.
• the web server now has the easy job of bit banging back to the browser.

Of course it does all those things and sometimes more usually in the time span of a couple hundred milliseconds or less without the benefit of having been pre-processed before hand for several seconds (or minutes) into immediate code so it try to keep up

BTW that does look like a cool project you were working on and I hope it takes off for you.

Here's a bit dated post I found comparing java vs something else for web apps you might find interesting. The answer by guy that went from Java to Python is really good I think. IMHO he'd been even better served by switching to PHP but it was still an improvement

So I'll say I did kind of miss what you were asking but I think I still indirectly answered it in part anyway but I'll try doing a better job here.

Kind of. PHP does do much like a lot of Unix/Linux view the outside world as just a bunch of data stream the thing it doesn't just work with the ones it's give it actual spends more time making it own to DBs and other services then dealing with the original input stream. The input is just a request for something usually by way of a simple message of some form. That message take many forms from the URL being requested to form data sent in the POST or GET. Even the fact that it's a POST or a GET can be view as part of the message as well. In most cases this is the easy part figuring out what it is being asked for. How it fulfills the request is by far the most complex thing it does. Outputting the response is likewise usually a very minor part of what it does.

sometimes. Some web server do use thread per connection and have PHP running within a single thread they start it in, others use a multiple thread per request model or in other ways use it with them that's why PHP has both thread safe and non-thread save versions. There are even some web servers that can switch or use a hybrid model and PHP can be used with them as well. PHP also has CLI version for non-web related stuff and testing and newer versions of PHP can be it's own web server as well for light weight local testing when even a local web server is just to much overhead. PHP even has everything it needs to make a full desktop application for that matter though most of the extensions aren't that well maintained as most people choose to view PHP through the very narrow view as just something you use to do web pages. My own projects Yapeal and Yapeal-ng are very atypical uses of PHP as a server side back end service / whatever you want to call it. Using it like this is becoming more common as site developers are pushing up against the limits of the only able to do stuff while the user is connected to web server. They have to choose between writing something in another language like Python, Java, or something else that has usually been used here or as more and more are doing just use PHP instead and be able to share their code base between both parts of what they are doing.

Personally right from the start I've seen PHP as a great general language that could replace all the nasty Perl and Bash scripting you usually see running around on Linux machines. A lot of people have gone with Python for these things now which is an improvement I agree but I feel that PHP development by focusing so much on the web missing have as much impact as it could have had as a more general cross-platform scripting/ system language. One of the reason I really love PHP is that it's usually installed on any Linux machine by default and on Windows its one short download, unzip, and maybe add a couple things to your path and your done. Installing Bash or Python or even Java for write new code is a lot harder to get setup.

I'll kind try summing up the middle part as this is already getting to be a long post again. Does PHP under stall RPC and all the other stuff? yes Anything that has been done especially in any way remotely web related PHP has built in or an extension you can use. Just taking a quick look at the RabbitMQ web site it looks like there are more tools to work with it in PHP than several of the other listed languages.

Here I believe you are miss under stand the nature of all web -centric/mobile application development which is what most things are about any more not just for PHP but for most other languages used as well.Maybe you do know it as well but just aren't use to thinking about it since that wasn't part of what you have done. Most of the app is on the client side in their phones, pads or other mobile device as a actual app they download or through their browser where the web pages are designed to make it feel like they installed something. Which in many ways they might have since everything is being done through JS which makes the request back in forth to the server without any page loads. Notice that even though things are designed to make it seem like there's a constant connection between the client and site which has state in actually it's all over a stateless on and off connection where each connection might not even be with the same server each time but through a lot of really interesting 'magic' is made to seem like it is. Now like I said you probably know most of this but the part you are missing is the web server is not in control and really has no way to 'push' anything to the client and in most cases firewalls etc actively try to keep you from doing so. The client has to basically poll for updates and then the site can let it know about anything new but only if the client knows to ask for it. Just by the nature of how it all work there's no not anything really sitting around to receive push info from what you are doing. PHP or anything else can do the same thing the client does and poll your service but push is not something that is really useful since it's only active for a couple seconds at a time at most. Now something like Yapeal-ngcould but its easy just to use pull there as well.

Dragonaire,

Thank you for all of the information! I can see you are a good man. I don't quite know how to say this without being offensive.

I'm getting old, and I tire of my career but financially have to continue. Most people in other careers do not realize that software engineering is like continually working on a masters degree. The technology changes Soooo rapidly with many of them becoming obsolete about every 5 years. One must constantly be learning the new technologies and keeping up with the brilliant kids coming out of college every year.

As you web browse leading edge sites today, have you noticed that annoying popup ads are not really used anymore, you get streaming video etc constantly updating long after initial page loads instead? Often times some are ill behaved and lock up as they run your machine out of memory and everything starts paging to disk and brings your box to a screaming halt if you leave the page open for a long time.

Have you also noticed that these sites will be pushing out adds to something else you recently googled? It approaches a breach of personal space what google has done with crawlers and a technology called ajax.

check out the summary at the top here: http://www.w3schools.com/ajax/

The computer world has come full circle in several things. Years back the push was to keep the servers freed up and push the work to the client. In leading edge web technologies today everything is moving back to the server. Like the Wicket framework, its all about super dynamic webpages, of which google via ajax can index your dynamic data (images and all). The NodeJs framework, I assume Python frameworks and many others are using these DHTML, ajax etc. for totally server side instantiated content. Static pages are dead (well dying quickly).

The google search for PHP documentation is revealing to those that have done this for a long time, where you have 15 years of undated web hits following several generations of a technology, but the main download org site has basically zero documentation on the current release. To me, this is an indication of a legacy technology with a huge use base in existence. With everyone interested in keeping code releases coming to maintain their legacy implementations, but no interest in maintaining the documentation. Realize like mentioned above, many technologies become legacy every 5 years.

I do realize that PHP has a huge following and has for 20 years. But it's old school now, its not industry standard anymore and its far from leading edge, it's not where new commercial projects are going to go.

Of course this is a don't care for hobbyist programmers that have a real life and certainly is not a requirement for what we are doing here.

Now has somebody found a way to run PHP7 on Windows XP... ?

That would be interesting


Also I immediately tought about Yapeal when I read this: ;)