A blog about my rantings, including Games, Game Development, Gaming, Consoles, PC Gaming, Role Playing Games, People, Gaming tips & cheats, Game Programming and a plethora of other stuff.
When you're working with PHP, hosted by Apache, if you are using the calls "exec", "passthru" or "system" to call into other applications you can miss changes in those apps, unless you change the PHP script before accessing the page again.
So, I have my browser pointed to my PHP page as "http://127.0.0.1/example.php" and in that PHP page script I have a call to "exec /usr/bin/myapplication" which is a program I've written in C++.
Now, if I change the application in C++ and recompile it, then refresh the browser I can fail to see any change, the Apache Server/PHP engine keeps the last version run cached. Great for speeding up your server, but a pain in the bum when you're trying to debug that C++ application through the web-browser interface.
The best solution to making sure your changes to the C++ show up appears to be to simple put an edit into the PHP file and save it again.
Note; Not even copying the PHP file into /var/www again makes a difference, the file itself has to change, if only by adding a space and then removing it again.
This edit then makes the Apache/PHP engine drop the cached version of the C++ Application and run it from scratch. This same problem must happen if you do anything to change any application you are calling with the above PHP commands.
No wonder I can't understand them when they're in India, if this is the kind of voice coaching they're getting:
"B-U-T is pronounced as 'but', however P-U-T is pronounced 'poot', [like foot] not 'putt'," explains teacher Stephen Rosario, as he coaches the class in how to enunciate English words.
But and Put... are the same changing the constonant from B to P does not change the inflection of the ut.... But, Put, Hut all sound the same, even Foot sounds like Fut to me (a very native British English Speaker)...
Poot, a pooter is a womans privates for heaven sake!
Ignore this "Stephen Rosario" he clearly can't speak English himself, his name sounds Italian. Gah.
So, taking a patient look back at all the graphics technology I have documented and on my reading list, I figured I needed to pick something worth reading that was going to further my actual work too. So, I picked up looking at scrolling and 2D effect games.
You know what? I am amazed at how much knowledge seems to have been lost, I was looking at old games from ID (the makers of Doom & Quake), I also looked at a bunch of 2D scrolling games on my old Atari ST, and started to put together some simple testing code to do the same soft of effects in Java & C#.
But, I'm after hard information about how to do certain tasks, how to achieve certain effects, but the knowledge is lost.
So, I cry out to all those old programmers out there, all you 2D scrolling nuts, blog about the technology! I'm going to be, but I have to reinvent your wheels first.
I've stumbled into a habit of late, something I can't stop myself doing, which I've never really done before... I'm parallel processing... Since 1989 when I became more aware of the internal workings of my Atari St, and the way it did one thing at a time for me I've always worked the same way. I used to write softwatre that way, picking up a component, designing, writing, testing and documenting it in sequence.
But, I found that as my software projets have gotten bigger and bigger I needed to be able to put parts of things together, slide in other parts and then go back to sew things together. This started out small, but in a current personal project I'm working on its a sprawling documented seething group of about 110 classes (and counting).
Concurrently running through each module in turn, going back reading my notes and reviewing each, incrementing and incorporating innovation from one place through the body of the whole project.
Its turning into a bit of a quality kicker for me, my personal software projects are far more higher quality than the ones I'm writing for my employer, simply because my employer don't really seem to see the point of my spending four weeks to do a task which they think should take one.
I myself however am reaping the benefits. My current software project is, unfortunately, about four months behind schedule and had a bit hiatus between June and July as I re-wrote vast swathes. But things are quieter now, for example in one day (with the improved quality) I was able to replace a major components of the display systems and rendering threading in a day... it should have taken at least a week alone, just to get a coloured screen to display. Instead, because of the quality of the underlying code I was able to get this feat of engineering done in less than 8 hours and had it integrated on my three test boxes for soak testing after 10.
The question is, how do I sell this contept to my bosses?
I love when I get really really shit attempts at phishing scams. Here's the one that just showed up... I love it, this Chineese accented Engrish really needs to be docrumented somewhere...
Not to the degree of moving house, but when mobile's first came out I could not stand to sleep in the same room with one, I used to get a headache every morning (if it were on). And I had a Motorola phone which was just as bad, if I had it beside my bed when I slept I'd wake with an almost hung over feeling.
I now have a Nokia phone and I intentionally put it way over to the side of me, over two meters away, or I leave it to charge over 4 meters away, over night.
So there maybe something in this...
I'm not bothered by wireless networks, I don't think... but mobiles certainly have this running recurrent experience to me... I might do a blind test, get a sock or something, and ask the wife to put it under the pillow her side of the bed with, or without, the mobile in it... and see if I marry up a bad head with the phone being present...
Over the last few months there has been plenty of coverage of the recent civil unrest and rioting here in England, plenty of pundits, speculators, commentators and nanny-say-do-weller's have been on their podiums spouting off.
Some have blamed the parents, some the lack of morals, some video games, some rap & hip-hop culture, some the simple bad economic situation and lack of jobs.
I myself could speculate plenty about the problems, I was born in St Anns and raised in Top Valley, two of the roughest, dirtiest, dangerous areas of Nottingham city. When people think of Nottingham they think of guns and shootings, so much so that the city has recieved the monica "Shottingham" in the tabloid press. What seens to be reported however is always the shock and outrage at the time, but little ever gets done to address the situations.
So, why some three generations on is the place such a shit hole? And I have two view points as to why, firstly the lack of jobs. When I was a kid you could hear of peoples relatives going and getting casual work. In factories, in food processing, in road or building works. Work was, if not plentiful, certainly undersubscribed.
They went on about the unemployment figures in the 80's, but I strongly believe in our area a lot of the unemployment figures were from people simply not declaring they had a few hours here and there to make ends meet. They took their benefits and added extra's.
Now-a-days though, those self same people are simply taking the benefits, there is no way for them to even get those part time ad hoc jobs they once had, and certainly no hope of decent employment. So they're constantly struggling, the pressure builds and you get what we had here the other week.
That's all part of it, and being economics its all very convoluted and confused, so many inputs and so many expectactions that this is as far as I'm going to go with it and I'm going to move onto my other point.
Policing. The police. In the 80's and 90's you never saw them walking the beat on Top Valley, just as you never really see them today, but you never got the amount of violent disorder and crime you have today. The reason for this was two fold, firstly the police if they were called came down on whomever was causing the problems like a tonne of bricks. They took the work of two or more witnesses and acted, either telling the supposed-guilty party to shut up and buck up, or they gave them a night in the cells. It was an ever present threat which ever teenager in the area pretty much tried to avoid.
Yes there was still high spirits, yes some crime, and certainly crime from many passing through the area. But the actual people who made up that community did for the large part conform for a better standard of living all round.
Today however, you call the police and they come to mediate. You formally inform them of an infraction of the law, you spell out the actual laws someone has transgressed and you actually formally request that they take action... and they come to mediate... "oh that might inflame them, lets talk about this"...
What the fuck?
When I call the police, three times in a week, and have documented, diaried, recorded evidence of the situation at hand, when I am willing to testify, give statement and act to solve the problem, I want that active, reactive, policing approach from the 80's and 90's. I want the copper to listen to me, to understand I'm not a fool and to act.
Certainly not as my puppet, but to definately listen to my input and respond, hell I'd like to see them say "you're fucking moaning about nothing", then at least I would get off my high horse.
But oh no, modern policing... its mamby-pamby ball-less soul-less crack pop mediatory bollocks, and I'm saying this from experience.
Officers are either very confrontational to you the affronted party and very placating to the guilty party, or they simply do not engage with you. I've seen two different officers, both male, stand toe to toe staring down affronted parties (who were also male) and then being concilliatory and pacifying to guilty parties (who were female), its like the stab vests are holding in too much testosterone and they're trained to just pacify the women and to piss the men off.
And as such I'm not surprised that there were riots, I'm not surprised police stations were attacked. If I operated the prime civil law enforcement infrastructure and I witnessed my organization being so easy to trample at the bread and butter front line policing interface with the public, like the police here are, then I'd be very scared to be running that organization even ashamed.
None of this is meant to mean that we need draconian policing, but some balls at least, some common sense from practical living, rather than namby-pamby self reassurance course invested pseudo science clap trap approaches.
This blog post is a following of the excellent tutorial at http://mikeos.berlios.de/write-your-own-os.html by Mike Saunders. My thanks to him for his permission to post this information and quote his original work.
Something I've been mooching over recently is writing my own assembly code, as part of my degree I did a tiny bit of 68000 assembly and before that I'd included assembly based interrupt driven Mouse controls into my Turbo C and Turbo Pascal programs but I had never actually sat down and understood Assembly on the x86 class processor.
So I have set about learning a little bit of assembly, using a text editor and nasm. After getting the basics I asked the question 'What will I make?'. Assembly is a really alien world, I'd seen listings of coding in it many times, all full of strange seemingly massively repetive commands like mov, add, dx, si and other scary looking nonsense; even as an experienced Software Engineer I had a certain amount of trepidation going through the topic at hand.
But, I soon realised I knew very little about how the BIOS in the machine picks itself up and starts the operating system. Sure I knew how to install GRUB, how to format a drive to make it bootable and add a new master boot record, but I'd never actually looked at how the software picks itself up, how it "bootstraps"... A PC is often said to try to pick itself up by its own boot straps when you turn it on.
The first bit of software run sits in some hardware chips (ROMS), this primevil code's job is to call launch an application loaded from the first sector of a disk (the first sector is the master boot record - or MBR). This application is the boot-strapper or boot-loader application, and you can write your own...
So I hit the inter-webs for some help and came across the MikeOS project, specifically the excellent introductory boot loader tutorial, and this is where I'm going to pick up the rest of my blog post, it struck me that as brilliant as that tutorial is it doesn't really hold ones hand through the process step by step. So I'm going to try and help you write your own MikeOS inspired boot-loader with Ubuntu (or Kubuntu or Edbuntu or Debian) you can do this on your Windows machine too.
So, lets get started, first of all you'll need to download a whole bunch of software from the internet, you'll need the following:
Development System
Ubuntu 32bit Desktop Edition - I used v10.04 LTS - so anything newer will do:
If you want to run this tutorial as I have, get yourself a spare PC from somewhere, burn the Ubuntu ISO to a CD and install your system. If however you've not got a spare machine at hand, then you can read on anyway...
Next, you'll need VMWare Player. Many tutorials on writing your own boot-loader use qemu, but I'm going with VMware player because you can get it for windows and linux... so perversely if you want to follow this tutorial on a windows machine you can run VMWare player inside your Windows installation, and let it run the Ubuntu 32bit system you downloaded in the first step!... neat.
So, go now and set yourself up, if you're really really stuck setting yourself up then drop me a line, we'll have a skype about what's going on. But I'm going to assume you're tooled up, you've install your machine (or host a new virtual machine with VMware Player under windows) with Ubuntu and you're sat looking at its desktop.
Configure Ubuntu Tools (good for Debian, Kubuntu, Lubuntu etc).
Once your Ubuntu Install is up and running set it up to look however you wish, and open a terminal window (with either the CTRL+ALT+T key stroke, or by pressing ALT+F2 and typing xterm). Inside your console window you will now need to install some software, namely the essential items to build software with and the assembler we'll be using; which is called nasm.
So, inside your Ubuntu terminal you will need to run the command:
sudo apt-get install build-essential nasm
This will install both the build essentials and the nasm assembler in one fell swoop.
We could now create some software but we'd have nowhere safe to run it...
VMware Player
This is where VMware player comes it, it'll let us create a virtual driev image (a floppy image) into which we'll put our boot laoder, without risking our real hard drives or wssting burning CD's all the time.
You will now need to install VMWare Player into your Ubuntu machine (if you are already hosting Ubuntu from Windows don't worry about this step, use the player from windows and copy the floppy image; mentioned in a moment; to windows to run the software we create).
To install the player simply open a terminal (or use the one you already have open) navigate to wherever you have downloaded the player bundle to and use the command:
sudo sh VMware-Player-3.1.4-855356-i386.bundle
(The last item there is the file name I've downloaded, by the time you come to read this tutorial you may have downloaded a different build of the product, so use whatever file name you have exactly in place of the one above: sudo sh <file name>).
Once this starts up to install you'll be back into the land of a GUI, so you can praise whatever god you believe in and use the mouse to set up the player software.
Once the player has installed on Ubuntu you'll be able to launch it from the terminal with the command "vmplayer" or you can launch it with a mouse click from the menu via "Applications->System Tools->VMware Player".
Go a head and start the player up, we need to use it before we can get coding.
Floppy Image
The first task we have is to find somewhere to put our boot-loader when we've built it. We don't want to use our hard drive because we still wan to boot our machine as it is... and we don't want to keep burning and wasting CD-R's... we don't want to use a USB stick to boot because they work slightly differently to normal drives... so the option left to us is a floppy.
Now, not many of us have a real floppy drive any more, infact I'm sure some of you reading this will have never held a floppy disk in your life, but don't worry, because we're going all virtual. We're going to get VMware Player to create us a blank floppy image.
With VMware Player open, click to create a new virtual machine, select to install the operating system later, and opt for a type of "Other" and select "MS-DOS" from the drop down... we're not going to actually let this virtual machine come into being, just advance through the screens until you get to the hardware selection screen, when you see the options to allow you to select "Customize Hardware" click that button and you'll be presented with the list of devices.
Inside the customization page you should see in the list of devices "Floppy", if you left click this item you will see several options, one of which is "Use a floppy image:". Check that option and you'll now be able to get at a button called "Create"... you guessed it, hit create and save yourself a floppy image (FLP) file somewhere (confusingly the option is "open" on the dialog, so 'open' the image to save it).
Once you have the image just cancel the creation of the virtual machine, we don't want it at this time.
So, I now have a "blankfloppy.flp" file, which I've saved out... that image can be very useful so take a copy of it somewhere as a back up...
Code
Now, I'm going to give you the code of a whole simple, well commented, boot loader (derived from the boot-loader in MikeOS). All this assembly does is specify the stack, define a constant string and output that string to the screen. It then sits in an infinite loop.
; This is the assembly code for our Bootloader
BITS 16; Instruct the system this is 16 bit code
jmp start; Jump to the start routine, short being never to return
nop; No operations from here
; This is the entry point put nothing else above this, not even
; include statements
start:
mov ax, 07C0h; Setup 4k stack space after this bootloader
add ax, 288; (4096+515) / 16 bytes per paragraph
cli; Disable Interrupts
mov ss, ax
mov sp, 4096
sti; Enable Interrupts
mov ax, 07C0h; Set data segment to the load point of
mov ds, ax; our application
call PrintWelcome; Call the print welcome message routine
jmp .loopHere; Jump to the loop here label
.loopHere:
jmp .loopHere; Jump to the loop here label
; this is now an infinite loop
; so all our system did was boot
; welcome us and sit spinning
; on this label
;------------------------------
; Constants
;------------------------------
c_WelcomeMessagedb"HELLO AND Welcome to our OS", 0x0d, 0x0a, 0x00
;------------------------------
;----------------------------------------
; Routine to print the welcome message
; this is our only procedure in the
; application, so enjoy!
;----------------------------------------
PrintWelcome:
mov si, c_WelcomeMessage; Move the string into si
call PrintString; Call our routine to output til 0x00
ret; Return to the call point
;----------------------------------------
;----------------------------------------
; Print string Routine
; PrintString (SI)
;----------------------------------------
PrintString:; Routine to output string in SI to the Screen
push ax; Push the AX register to the stack
mov ah, 0Eh; int 10h 'print char' function code
.repeat:
lodsb; Get character from string
cmp al, 0x00; Compare the character with zero (0)
je .done; if the character is zero then end of string
int 10h; Otherwise call BIOS interupt 10 to print
jmp .repeat; And now jump back up to the repeat label
.done:
pop ax; Pop the copy of the AX register off the stack
ret; Return the call stack back to the call point
;----------------------------------------
; Now we need to pad the remainder of the boot sector with 0s
times 510-($-$$) db 0; Times is a macro for nasm to repeat an action during
; compilation of the the application binary, in this
; case to add 510 lest the amount of data and code in
;the program to the output file
; Why 510, why not 512? A boot sector being 512 bytes?... well...
; to tell the BIOS this application, when written to the MBR, is
; a boot sector, we need to add two bytes to the end of the
; application, so we call define word with 0xaa and 0x55 now
Read the code yourself to understand what its up to... that's the key to this tutorial, you have to apply yourself to understanding the assembly, everything else I hope you're having thought about, so you're free to look at that code in detail.
With this file in hand on disk open a terminal window once again and move to that folder, so for me its under my home directory in a folder called 'asm' in a file called '1.asm'. To get to it and to build it with nasm I'm going to perform these commands:
CTRL+ALT+TTo open the terminal window
cd /home/jonb/asmTo move to my home folder's asm folder
nasm -f bin -o OS.bin 1.asm
To build as binary into a file called 'OS.bin' the 1.asm source
Everything should be fine, and you should now be able to list the directory (ls) and see 'OS.bin'.
Copying Boot Sector to Virtual Floppy
So, we now need to move that OS.bin into the blank floppy image (or a copy of the blank floppy image)... so first things first, I copy my blankfloppy.flp into the /home/jonb/asm folder.
This does a none-truncating move of the OS.bin into 'blankfloppy.flp' (so it's not blank anymore, its now got a boot loader installed), but the flp file stays at a size of 1.44MB and the first 512 bytes of the file is overwritten with the bin file.
Running our Code
Now we want to run our boot-loader, so lets jump back to VMware Player; if we now create a virtual machine of type "MS-DOS" with a tiny disk (0.1gb), we can select the floppy (under customize hardware), ensure it is connected on power up and browse to and select the 'blankfloppy.flp' file.
Running the virtual machine now, we should see the VMWare BIOS post, it attempt to start over a network boot, fail and then it use our floppy disk image as the boot point.
And you should see out "Welcome" message from the assembler as the only thing that comes out... that is our bootloader... we have literally control at the earliest moment possible of that virtual machine.
(You could run the floppy image in qemu too if you want - see the MikeOS tutorial - link at bottom).
(If you are hosting a Ubuntu Virtual Machine, simply copy the floppy image out of Ubuntu onto the harddrive of yoru Windows machine, launch a second copy of VMware Player and create your MS-DOS type virtual machine to boot with the floppy there, you just introduce a copy back to Windows each time you change the floppy image with the 'dd' command).
Video
They say that a picture speaks a thousand words, so for you lucky people I've create a series of videos documenting the highlights of this process.
Its best if you expand the videos, and view them in HD full screen.
Credits
The boot loader code, with the exception of a little tidying up and extra commenting is copyright Mike Saunders: http://mikeos.berlios.de/write-your-own-os.html and as per his website I publish this code here completely open source (under a BSD-like license) with the caveat that any usage of this code, either directly, or as part of your own work should always carry a reference to this blog post within.
