If the answer isn't here, ask and it will be added...
See readme.txt.
From any SimTel mirror, in the directory gnu/djgpp/v2tk/, or from the Allegro homepage, http://alleg.sourceforge.net/.
See allegro.txt.
Make is part of the standard djgpp distribution, in the file v2gnu/mak*b.zip (whatever the latest version number is). You can get this from http://www.delorie.com/djgpp/.
Are you sure that you have installed djgpp correctly? Check the instructions in the djgpp readme.1st file, and in particular confirm that you have set the DJGPP environment variable and added the djgpp\bin directory to your path.
Are you sure that you are using the GNU make program, and not some other version like the Borland one? If in doubt, run "make -v" and check that the GNU copyright message is displayed.
You shouldn't be running make from the Win95 explorer! Bring up a DOS box, change to your Allegro directory, and type "make" at the command prompt.
Are you sure that your system time and date are set correctly? Alternatively you may have downloaded a fresh file made by someone in a more eastern timezone. In the latter case you can either wait a few hours, temporarily set your clock forward, or use a utility such as "touch" to change the timestamp of the files.
As the error message suggests, you need to provide more memory for the compiler to use. The go32-v2 program will tell you how much is currently available. If you are running under DOS, try to free up more disk space for use as a swapfile. When using win95, increase the DPMI memory limit in the properties for your DOS session to 65535 (you'll have to type this in by hand, because the pulldown list doesn't go above 16384).
As the error message suggests, there is a conflict between newer versions of gcc and older versions of the libc. You should read the readme.DJGPP file contained in the gcc*b.zip archive (a fix is appended to the end of the file).
You haven't read the docs, have you? :-) You need to link your program with the library file, liballeg.a. First, make sure you have installed everything properly (running make should do this for you). Second, if you are compiling from the command prompt or with a makefile, add -lalleg to the end of your gcc command line, or if you are using Rhide, go to the Options/Libraries menu, type 'alleg' into the first empty field, and make sure the box next to it is checked.
No, sorry. For starters, liballeg.a is about 450k, but you'd probably also want various utilities like the grabber, sound setup, etc. And what about all the example programs? If I included compiled versions of everything, a binary distribution would be over 7 megs: way too big to be practical! More seriously though, there just isn't any reason why you can't build it yourself. A compiler is a deterministic process: given a particular input file and command line, it will always produce the same output. If this isn't working, you either have the wrong input files (ie. your copy of Allegro is broken in some way), or the wrong command line (it is hard to see how that could happen, since all you have to do is type "make"...) or your compiler is broken, ie. you didn't install djgpp properly. You need to find and fix the problem, not just sweep it under the carpet by getting someone else to compile Allegro for you...
GNU tools write their error messages to the error stream, stderr. Unfortunately command.com is too stupid to know how to redirect this, but fortunately DJ was smart enough to work around that, so you can use his redir program to capture the output messages, for example "redir -eo make > logfile.txt".
Your mileage may vary. Some people have reported problems, while others say that it works fine. I don't have access to NT myself, so there is nothing I can do about it either way, and to be honest I don't care very much. Use the Windows version of Allegro if you want to make Windows programs. If you want to run DOS programs, use DOS!
Yes, but with some caveats. If you are using the OpenDOS version of EMM386, you must disable their DPMI implementation (specify DPMI=OFF on the EMM386.EXE device line in your config.sys). You should also make sure the PIC=OFF flag is set, but this is the default so it won't be a problem unless you have specifically enabled it.
You can't. The limit is imposed by the VGA hardware, not by Allegro. To access more than 256k of video memory you need to use an SVGA mode, which means either switching to a higher resolution or getting a copy of the SciTech Display Doctor, which provides several low resolution SVGA modes.
It does for some people, but not for others. The problem is that Creative Labs refuse to release any specs, so we don't know how to write a driver for it. Complain to them, or buy a different card from a more reasonable manufacturer.
This might be because you have the volume set very low: try changing this in the setup program. Also, Allegro is mixing several sounds into a single output buffer, unlike the Windows sound player that only plays one sample at a time, so each individual sound can only get a smaller percentage of the total output volume. This is just the price you pay for multiple output channels. If you don't like it, use the setup program to tweak the number of channels: this can be any power of two less than or equal to 64, and the smaller you make it, the louder the sound will be. Alternatively, use set_volume_per_voice(), described in the docs. This will enable you to adjust the overall volume of Allegro's digital sound output.
Try using a FreeBE/AF driver (http://www.talula.demon.co.uk/freebe/), or the commercial SciTech Display Doctor (http://www.scitechsoft.com). If it still doesn't work, post a description of your problem to the Allegro mailing list, along with a copy of the output from running the afinfo and vesainfo programs.
The VBE/AF interface already provides this for the video drivers: see the FreeBE/AF project on http://www.talula.demon.co.uk/freebe/. For more general things like the sound, VESA, and mode-X routines, this would be very difficult to do because the drivers depend on a lot of helper functions from the core lib. The djgpp DXE format is nowhere near flexible enough to support this, and I don't want to make Allegro dependent on any other dynamic linking packages.
Well duh, you need to increase the size of your environment then :-) You can do this by changing the settings for your DOS box (click the system menu and select "properties"), or at startup in your config.sys file (eg. "shell=c:\command.com /e:8192").
You need to tell your compiler how to find the DirectX include files and libraries: put the DirectX SDK /include and /lib directories in the compiler/linker path. Alternatively, if you don't want to modify any configuration stuff, you can simply copy the files from the DirectX SDK /include and /lib directories to the corresponding ones for your compiler.
You need to upgrade to a more recent version of the DirectX SDK, at least version 5, which you can get from the Microsoft Developer site.
You need to write END_OF_MAIN() just after your main() function. Allegro uses this, along with some preprocessor magic, to turn a normal main() function into a Windows-style WinMain() entrypoint.
Make sure you're building a Windows GUI Application, and not a Windows Console Application. This is a setting when you create the project in MSVC. This is specified by the '-tW' flag for the Borland C++ command line compiler. Either that, or #define USE_CONSOLE at the top if you really need the console for your program.
Well duh, you need to increase the size of your environment then :-) You can do this by changing the settings for your DOS box (click the system menu and select "properties"), or at startup in your config.sys file (eg. "shell=c:\command.com /e:8192").
Make sure that you don't have a semi-colon appended to your MSVCDIR variable if you are using MSVC, to your MINGDIR variable if you are using Mingw32 or to your BCC32DIR variable if you are using BCC.
You need to write END_OF_MAIN() just after your main() function. Allegro uses this, along with some preprocessor magic, to get a copy of your argv[] parameters (it needs those for various internal things).
You need to redirect stderr into a file, so you can view them later. The method for doing this depends on your shell: if you are using a Bourne-style shell like bash, try "make 2> logfile.txt".
There are two possible reasons: 1) DGA2 may support different resolutions/color depths than X, run the gfxinfo program to know what combinations you can use, 2) You may have a buggy DGA2 implementation, see next question.
You are probably using a XFree86 server with a very buggy DGA2 implementation, such as 4.0.3 (shipped with Red Hat 7.1 for example). Upgrading to 4.1.0 will probably solve your problem. You can obtain it from ftp://ftp.xfree86.org/pub/XFree86/4.1.0/binaries/ selecting the directory suited to your platform and following the instructions you find in the Install file.
The grabber needs to be linked with the code in datedit.c. But you shouldn't have to worry about doing this by hand: just run "make" and that will build everything for you.
You need to update your version of GNU binutils. See readme.txt to find out what the minimum required version you need is.
Not unless <foobar> is mentioned in readme.txt as one of the supported platforms. You could port it, but that is usually a lot of work. If <foobar> is a 16 bit DOS compiler like Borland C, you might as well just forget the idea :-)
WIP stands for "work in progress", and refers to any changes that are more recent than the last official release. WIP versions of the library can be obtained as patches from the Allegro website (http://alleg.sourceforge.net/), and are usually quite stable, although obviously not so well tested as a final release version.
Do you have a copy of patch.exe? If not, go and get it from the same place that you downloaded the rest of djgpp: this tool is a standard part of the compiler distribution. If you do have the patch program but it isn't working properly, make sure that you are installing the patch over the top of an unmodified copy of whatever Allegro version it is intended to update (this will usually be the most recent official release from before the WIP was made, but check the text file that comes with the WIP to be sure).
That file is distributed separately in the WIP versions. It is here: http://alleg.sourceforge.net/wip.html. (alldata.zip)
This is just the way that the video hardware works: there can only be one palette in use at any given moment. Either convert your images so they all use the same palette, or switch to a truecolor graphics mode.
See the Allegro.cc homepage (http://www.allegro.cc/) for some utilities, for example FixPal and Smacker.
The VGA hardware only uses 6 bits for each color component, which means the red, green, and blue values in the palette structure range from 0 to 63, not all the way up to 255. That gives you a potential 2^18 = 262144 different colors, or 64 shades of grey. If you need more than this you could try using VESA function 0x4F08 to select an 8 bit wide DAC, but Allegro doesn't support this directly and I don't know how reliable it will be across different hardware.
When you are in a 256 color mode, the VGA card displays color #0 around the border of the display area (in truecolor modes it displays black). Your funny color will go away if you change the palette so that color #0 is black.
With great difficulty :-) There is no such easy trick as just altering the palette, so you will have to repeatedly redraw the image in a lighter or darker form. You could draw black translucent rectangles over the screen to darken it down, or use draw_lit_sprite() to tint a bitmap while copying it to the screen, but be warned that these operations are expensive and will require a fast PC!
fade_in() and fade_out() only work in 8-bit paletted modes. One of the ways to fade a bitmap in is to loop through all the pixels and change them individually. An alternative is to draw lots of translucent black boxes over the screen. However, these methods can be slow. The example exxfade.c demonstrates how to fade between two bitmaps. You could modify this so it fades in or out from/to a black screen. See also the previous question.
See the Allegro.cc homepage (http://www.allegro.cc/) for some utilities, for example DTA and Smacker.
In your favourite paint program, get hold of the RGB sliders and drag the red and blue ones up as far as they go (usually to 255, but this will depend on what units your software uses), and the green one right down to zero. The result is a special shade of Magic Pink, or as some people prefer to call it, magenta.
Remember that the vertex positions are stored in fixed point format, so you must use the itofix() macro or shift your coordinates 16 bits to the left.
Remember that the angle of rotation is stored in fixed point format, so you must use the itofix() macro or shift your coordinates 16 bits to the left. For example, rotate_sprite(bmp, spr, x, y, itofix(32)) will rotate the graphic by 45 degrees.
You are probably trying to initialise the dialog structure with a pointer to your bitmap, right? That won't work because the dialog is created at compile time, but the bitmap is only loaded at runtime, so the compiler doesn't yet know where it will be located. You need to fill in the dialog structure with a NULL pointer, and then copy the real bitmap pointer into the dp field as part of your program init code, after you've loaded the bitmap into memory.
It depends on exactly what you are doing. If your images are totally opaque, there is no advantage to using an RLE sprite, and it will probably be faster to use a regular bitmap with the blit() function. If your graphics contain masked areas, an RLE sprite will almost always be both smaller and faster than the draw_sprite() function. Compiled sprites are in general quite a bit faster than both the others for masked images, and slightly faster for opaque graphics, but this is far more variable. They are at their best with small sprites, on older machines and in mode-X, and may actually be slower than blit() when using SVGA modes on a pentium (the large size of a compiled sprite is very bad for the cache performance).
You need to make sure the game logic gets updated at a regular rate, but skip the screen refresh every now and then if the computer is too slow to keep up. This can be done by installing a timer handler that will increment a global variable at your game logic speed, eg:
volatile int speed_counter = 0;
void increment_speed_counter()
{
speed_counter++;
}
END_OF_FUNCTION(increment_speed_counter);
void play_the_game()
{
LOCK_VARIABLE(speed_counter);
LOCK_FUNCTION(increment_speed_counter);
install_int_ex(increment_speed_counter, BPS_TO_TIMER(60));
while (!game_over) {
while (speed_counter > 0) {
update_game_logic();
speed_counter--;
}
update_display();
}
}
Add a call to save_bitmap() somewhere in your code. See the save_bitmap() documentation for a discussion of one common pitfall when doing this, and some example code.
Call srand(time(NULL)) at the beginning of your program, and then use rand()%limit to obtain a pseudo-random number between 0 and limit-1.
There is no need. The linker will only include the parts of the library that you actually use, so if you don't call any of, say, the texture mapping or FLIC playing functions, they will be left out of your executable. This doesn't work perfectly because a lot of the Allegro code uses tables of function pointers that cause some unnecessary routines to be linked in, so the majority of the graphics functions will be included in every executable, but I have tried to keep this to a minimum. See allegro.txt for information about more precise ways to remove some of the graphics and sound drivers.
No. I think this sort of hardware support would be most useful as part of a proper 3D API, which Allegro is not, and will never be. If you want to do some work on this, the MESA library (a free implementation of OpenGL) is IMHO the place to start.
Several very good ones already exist, for instance the JGMOD package. See the audio library extensions section on the Allegro.cc website (http://www.allegro.cc/). You are not allowed to suggest that one of these libraries be merged into Allegro, because this topic has already been done to death on the mailing list and I am tired of it.
That would be a lot of work, and I don't have time to do it myself. There are several networking packages currently in development or floating around on the net, though, and in my opinion this sort of code is more useful as an external library than it would be as part of Allegro.
Unisys has a patent on the LZW compression algorithm that is used by the GIF format. I want everything in Allegro to be freely usable without any restrictions whatsoever, which means I can't include any code that is subject to licensing or the payment of royalties.
Perhaps. Try to isolate the smallest fragment of code that is able to reproduce the problem, and I'll have a look at it. If you can send me a 10 line program, I will fix it. 100 lines, and I can probably fix it. 1000 lines, and I don't have a chance :-)
Sure. See the giftware terms in readme.txt. I don't mind what you do with it, and there are no problems with commercial usage.
Whenever it is done! A little encouragement is always welcome, but I don't have any completion deadlines and I'm not going to make one up for you :-) As soon as it is finished, it will be released.
The grabber can import directly from GRX or BIOS format .fnt files, or you can draw them onto a .pcx image using any paint program. See http://www.talula.demon.co.uk/ttf2pcx/ for a utility that will convert Windows TrueType fonts into this .pcx format.
See the Allegro homepage (http://alleg.sourceforge.net/) for some links. You can use Gravis patches (.pat format), or SoundFont 2.0 (.sf2) files, but the latter must be converted into a patches.dat file with the pat2dat utility.
You need to download the makertf conversion utility (ftp://ftp.coast.net/Coast/win3/winhelp/mkrtf104.zip), and the Windows Help compiler (ftp://ftp.microsoft.com/Softlib/MSLFILES/HC505.EXE). Make a temporary directory, copy the allegro.txi file from the allegro/docs dir, and run the commands "makertf --no-warn allegro.txi -o allegro.rtf -J allegro.hpj" followed by "hcp allegro.hpj". The second command will give a lot of warnings, but they can safely be ignored.
The allegro.rtf file can be read directly into Microsoft Word and printed from there, but you should right-click and update the table of contents and index fields to fill them with the correct data first. Alternatively you can install the TeX package and use the tex and dvips programs to convert allegro.txi into Postscript format.
Check the Allegro.cc homepage, http://www.allegro.cc/. If you have anything to add to this, please post the URL!
It simplifies the maintenance of your program, in case the value of PI ever needs to be changed. Also it will make your program more portable to other compilers that use different values of PI.
A number of graphics cards have buggy or incomplete VESA implementations, and often the vsync() function is not implemented. For examples on flicker-free drawing, look at the code for the demo game, which uses a variety of methods to draw itself.
If the code works without optimisations, then it could be the compiler's fault. You can try beating the compiler into submission, for example:
while (!key[KEY_ENTER])
yield_timeslice();
For this case, however, it would be better to use readkey() instead. Or consider upgrading or downgrading your compiler.