Amiga GNU CC FAQ Version 0.2 This FAQ is intended to answer most (all, if possible) common questions concerning the Amiga Port of GCC. It is divided into the following subsections: 0. General Introduction a. The FSF, the GNU Project, and the GPL b. What GCC has to offer c. Hardware and Software requirements 1. Obtaining and Installing the Beast a. Sources of GCC b. The breakdown of the distribution c. Installation d. Common Problems with the Installation 2. Initial problems using GCC a. Set your stack properly b. Amiga specific library options i) ixemul.library ii) libnix iii) libamiga.a and others c. The object file format of GCC i) Converting libraries to this format d. Getting the Commodore includes e. The frontend of the compiler f. Optimization may work where no optimization doesn't g. Using the OS functions (an example) 3. Amiga-specific extensions a. The CHIP keyword (not currently implemented) b. Command Line options i) -noixemul (maybe goes in 2. (b)) ii) -fbaserel iii) -resident iv) Any more? c. Startup Code options 4. Advanced Questions a. How do I build a cross-compiler? b. Working with AmiTCP c. Writing code for ixemul.library d. Writing code for libnix f. Saving some RAM g. Writing shared libraries and resident (pure) code h. I want to change over from SAS/C to GCC i. Inline Headers j. Writing Hooks 5. C++ and Objective C (don't know much about these, except that they're not as well developed yet as the C compiler) 6. Notes for Un*x Hackers a. fork() is not implemented 7. Support Utilities a. Frontends b. Debuggers c. What else? 8. Additional Support a. amiga-gcc-port mailing list b. snail mail (?) - for people getting it off of CD-ROM c. You never know, it could be a problem in base FSF code (is there a newsgroup to point to here?) 9. Known Bugs a. the C compiler i) -resident option ii) -fbaserel option b. the C++ compiler c. the Objective C compiler 10. Maintainers and Contributors 11. Future 12. History Appendix A: How does the GPL affect programs compiled by GCC? Appendix B: GNU ports and who's dealing with them Appendix C: Another free C compiler (for those with little memory) Section 0, General Introduction ------------------------------- a. The FSF, the GNU Project, and the GPL The Free Software Foundation (FSF) is an institution dedicated to the free flow of software (and information in general (?)). To this end it has start the GNU (GNU's Not Unix) project, a collection of widely portable software of all sorts (compilers and Un*x utilities for the most part, but some other stuff as well). But I'm sure Richard Stallman, director and founder of the FSF, can say it better than I: [Note that this is heavily clipped; You should get the original at prep.mit.edu] ---------------------------------------------------------------------- Copyright (C) 1985, 1993 Free Software Foundation, Inc. The GNU Manifesto ***************** Why I Must Write GNU ==================== I consider that the golden rule requires that if I like a program I must share it with other people who like it. Software sellers want to divide the users and conquer them, making each user agree not to share with others. I refuse to break solidarity with other users in this way. I cannot in good conscience sign a nondisclosure agreement or a software license agreement. For years I worked within the Artificial Intelligence Lab to resist such tendencies and other inhospitalities, but eventually they had gone too far: I could not remain in an institution where such things are done for me against my will. So that I can continue to use computers without dishonor, I have decided to put together a sufficient body of free software so that I will be able to get along without any software that is not free. I have resigned from the AI lab to deny MIT any legal excuse to prevent me from giving GNU away. Why GNU Will Be Compatible with Unix ==================================== Unix is not my ideal system, but it is not too bad. The essential features of Unix seem to be good ones, and I think I can fill in what Unix lacks without spoiling them. And a system compatible with Unix would be convenient for many other people to adopt. How GNU Will Be Available ========================= GNU is not in the public domain. Everyone will be permitted to modify and redistribute GNU, but no distributor will be allowed to restrict its further redistribution. That is to say, proprietary modifications will not be allowed. I want to make sure that all versions of GNU remain free. Why Many Other Programmers Want to Help ======================================= I have found many other programmers who are excited about GNU and want to help. Many programmers are unhappy about the commercialization of system software. It may enable them to make more money, but it requires them to feel in conflict with other programmers in general rather than feel as comrades. The fundamental act of friendship among programmers is the sharing of programs; marketing arrangements now typically used essentially forbid programmers to treat others as friends. The purchaser of software must choose between friendship and obeying the law. Naturally, many decide that friendship is more important. But those who believe in law often do not feel at ease with either choice. They become cynical and think that programming is just a way of making money. By working on and using GNU rather than proprietary programs, we can be hospitable to everyone and obey the law. In addition, GNU serves as an example to inspire and a banner to rally others to join us in sharing. This can give us a feeling of harmony which is impossible if we use software that is not free. For about half the programmers I talk to, this is an important happiness that money cannot replace. How You Can Contribute ====================== I am asking computer manufacturers for donations of machines and money. I'm asking individuals for donations of programs and work. One consequence you can expect if you donate machines is that GNU will run on them at an early date. The machines should be complete, ready to use systems, approved for use in a residential area, and not in need of sophisticated cooling or power. I have found very many programmers eager to contribute part-time work for GNU. For most projects, such part-time distributed work would be very hard to coordinate; the independently-written parts would not work together. But for the particular task of replacing Unix, this problem is absent. A complete Unix system contains hundreds of utility programs, each of which is documented separately. Most interface specifications are fixed by Unix compatibility. If each contributor can write a compatible replacement for a single Unix utility, and make it work properly in place of the original on a Unix system, then these utilities will work right when put together. Even allowing for Murphy to create a few unexpected problems, assembling these components will be a feasible task. (The kernel will require closer communication and will be worked on by a small, tight group.) If I get donations of money, I may be able to hire a few people full or part time. The salary won't be high by programmers' standards, but I'm looking for people for whom building community spirit is as important as making money. I view this as a way of enabling dedicated people to devote their full energies to working on GNU by sparing them the need to make a living in another way. Why All Computer Users Will Benefit =================================== Once GNU is written, everyone will be able to obtain good system software free, just like air.(2) This means much more than just saving everyone the price of a Unix license. It means that much wasteful duplication of system programming effort will be avoided. This effort can go instead into advancing the state of the art. Complete system sources will be available to everyone. As a result, a user who needs changes in the system will always be free to make them himself, or hire any available programmer or company to make them for him. Users will no longer be at the mercy of one programmer or company which owns the sources and is in sole position to make changes. Schools will be able to provide a much more educational environment by encouraging all students to study and improve the system code. Harvard's computer lab used to have the policy that no program could be installed on the system if its sources were not on public display, and upheld it by actually refusing to install certain programs. I was very much inspired by this. Finally, the overhead of considering who owns the system software and what one is or is not entitled to do with it will be lifted. Arrangements to make people pay for using a program, including licensing of copies, always incur a tremendous cost to society through the cumbersome mechanisms necessary to figure out how much (that is, which programs) a person must pay for. And only a police state can force everyone to obey them. Consider a space station where air must be manufactured at great cost: charging each breather per liter of air may be fair, but wearing the metered gas mask all day and all night is intolerable even if everyone can afford to pay the air bill. And the TV cameras everywhere to see if you ever take the mask off are outrageous. It's better to support the air plant with a head tax and chuck the masks. Copying all or parts of a program is as natural to a programmer as breathing, and as productive. It ought to be as free. --------------------------------------------------------------------- The GNU Public License (GPL) was written in order to keep software published under it free. Here's the Preamble of the GPL: --------------------------------------------------------------------- GNU GENERAL PUBLIC LICENSE Version 2, June 1991 Copyright (C) 1989, 1991 Free Software Foundation, Inc. 675 Mass Ave, Cambridge, MA 02139, USA Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed. Preamble The licenses for most software are designed to take away your freedom to share and change it. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users. This General Public License applies to most of the Free Software Foundation's software and to any other program whose authors commit to using it. (Some other Free Software Foundation software is covered by the GNU Library General Public License instead.) You can apply it to your programs, too. When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for this service if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs; and that you know you can do these things. To protect your rights, we need to make restrictions that forbid anyone to deny you these rights or to ask you to surrender the rights. These restrictions translate to certain responsibilities for you if you distribute copies of the software, or if you modify it. For example, if you distribute copies of such a program, whether gratis or for a fee, you must give the recipients all the rights that you have. You must make sure that they, too, receive or can get the source code. And you must show them these terms so they know their rights. We protect your rights with two steps: (1) copyright the software, and (2) offer you this license which gives you legal permission to copy, distribute and/or modify the software. Also, for each author's protection and ours, we want to make certain that everyone understands that there is no warranty for this free software. If the software is modified by someone else and passed on, we want its recipients to know that what they have is not the original, so that any problems introduced by others will not reflect on the original authors' reputations. Finally, any free program is threatened constantly by software patents. We wish to avoid the danger that redistributors of a free program will individually obtain patent licenses, in effect making the program proprietary. To prevent this, we have made it clear that any patent must be licensed for everyone's free use or not licensed at all. The precise terms and conditions for copying, distribution and modification follow. ------------------------------------------------------------------------ More information regarding the FSF, GNU, and the GPL can be obtained at gnu.prep.mit.edu. Note that the source code of any GPLed software must be made available by the distributor (is this wording accurate?). The GPL may also affect how you distribute your programs, or maybe not. For more information, check out Appendix A of this FAQ. (Do this before believing all the bad hype about it. Then decide for yourself). b. What GCC has to offer. A FREE C compiler, for starters. I'm sure there's some other stuff. It's available for a wide range of machines (and thus it can't be beat for cross-compiling, at least on the Amiga). For the interested, a chance to change it as you like, make your own bug fixes, be an active part in the development process. c. Hardware and Software Requirements This depends upon what you're doing. Here's an excerpt from Phillipe Brand's readme for his distribution on the net: - Systems: Any Amiga (ranging from A1000 up to A4000/40, including CD-32 & SX-1) will run amigados-gnu utilities. - Memory: A minimum of 4MB free memory is needed in order to compile small/medium projects. More memory will be needed for large projects, such as recompiling gcc itself. Gigamem is known to work with GCC so *maybe* less memory will work. But in this case, you'll need an MMU equiped amiga (A3000,A4000/40, etc...). VMM40 (Public Domain Virtual memory manager) is also known to work with GCC. [Editor's note: using the g++ compiler can require even more, like at least 6 MB of free RAM] - OS: Starting from this release [2.6.3], 1.3 systems are not longer supported. Gcc runs fine on all other systems, starting from 2.04 up to 3.1 (40.68). - Disk Space; An installation of gcc requires the use of a hard disk. Approximately 10MB is required at present to install the compiler and utilities required to use it. In addition 3MB is required for the Commodore Developer Kit, which is required to be able to compile AmigaDOS specific programs. This kit is available direct from Commodore-Amiga. [Note, this is not true for the FreshFish distribution, which can be run straight off the CD-ROM if desired.] ----------------------------end quote--------------------------------- Yes, it can take a lot of resources, but that's a function of its portability. Be glad you can count on having the same compiler on your Amiga at home or the Unix system at work (or vice versa :-)). 1. Obtaining and Installing the Beast a. Sources of the Amiga port of GCC There are a couple of options for getting GCC. If you have access to the Internet, and ftp service, you can obtain the current version of GCC from Aminet, in dev/gcc. The main site of Aminet is ftp.wustl.edu in /pub/aminet, and has mirrors world-wide: USA (MO) ftp.wustl.edu USA (WI) ftp.netnet.net USA (TX) ftp.etsu.edu Scandinavia ftp.luth.se Switzerland ftp.eunet.ch Switzerland ftp.math.ethz.ch Germany kelly.uni-paderborn.de Germany ftp.uni-paderborn.de Germany ftp.uni-stuttgart.de Germany ftp.uni-erlangen.de Germany ftp.cs.tu-berlin.de Germany ftp.tu-chemnitz.de Germany ftp.fh-augsburg.de Germany ftp.uni-bremen.de Germany ftp.uni-kl.de Germany ftp.uni-trier.de Germany ftp.informatik.rwth-aach France ftp.cnam.fr Portugal ftp.ci.ua.pt UK ftp.doc.ic.ac.uk UK micros.hensa.ac.uk ftp.funet.fi also serves to act as a hub for porting GNU software to the Amiga, and serves to mirror Phillipe Brand's Amiga GCC tree directly, so that you may always find GCC there (particularly to guarantee the availability of source code). The distribution is broken up into several parts, both to make it easier to download and to respect the GPL. Thus each archive is broken into two parts, one for binaries and one for sources. Gcc-2.6.3 is split up into 15 archives: gcc263-readme.lha holds readmes files (include installation notes). gcc263-base.lha basic gcc distribution, hold necessary files. gcc263-inclib.lha headers and libraries. gcc263-c.lha C compiler. gcc263-c-020.lha 68020+68881 version of C compiler. gcc263-c++.lha C++ compiler, headers and libraries. gcc263-c++-020.lha 68020+68881 versions of C++ compiler. gcc263-objc.lha Objective-C compiler, header and libraries. gcc263-objc-020.lha 68020+68881 versions of Objective-C compiler. gcc263-doc.lha Gcc AmigaGuide(tm) documents & manpages. gcc263-utils.lha Useful utilities needed for development. gcc263-utilsdoc.lha Utilities documentation (guide & manpages). gcc263-texi.lha All Texinfo documents. gcc263-diffs.lha Diff files for all binaries. gcc263-src.lha Source for gcc-2.6.3 plus diff file. The second way is to utilize your CD-ROM drive and get one of Fred Fish's great CD-ROM collections. Any of the FreshFish or GoldFish Vol. 2 will have it, though the later the CD-ROM the later the version. He also includes as many ports of GNU utilities as he can. His restriction is simply that the binaries he includes must be compilable from the sources he includes (this has been a problem in the past, and Fred is a stringent observer of the GPL). So, yes, you also get all the sources with the binaries. (If you've got a CD-ROM drive, you should get one of these anyway; it's a great deal). These CD-ROMs may be purchased from Amiga Library Services by: Snail Mail: Amiga Library Services 610 N. Alma School Road, Suite 18 Chandler, AZ 85224-3687 USA FAX or Voice: (602) 491-0048 Phone: (800) 804-0833 or Email: orders@amigalib.com (but sending credit card information over e-mail is not secure, so one of the other methods would be preferable. Fred is working on using some kind of encryption though. More details when they're available.) There may be other CD-ROMs containing GCC, I just know of Fred's since he is directly involved in the porting of it. The breakdown of this distribution is different. Check the specific CD-ROM for details. They also include the Commodore Native Developer Updates (no Autodocs) for 2.04, 3.0, and 3.1, which is a big bonus. c. Installation Installation procedures may vary depending on your source, but the final directory structure should be the same. Check the README files included in your distribution package for more details. The ftp distribution (Phillipe Brand's) contains an Installer program. It utilizes the following layout: Name What Where ---- ---- ----- COPYING GNU LICENSE, read!! All archives COPYING.LIB GNU LIBRARY LICENSE, read!! All archives README-2.6.3 this file All archives NEWS-2.6.3 What's new in gcc-2.6.3 gcc263-base Installer Commodore installer utility gcc263-base GCC-Install Installer script to configure gcc All archives envarc/ global environment variables you should have set when using this programming gcc263-base environment include/ non-amiga specific C/C++ headers gcc263-inclib os-include/proto amiga specific protos headers. gcc263-inclib os-include/inline amiga specific inline C headers. Add gcc263-inclib Commodore headers!! os-lib/ amiga specific libraries gcc263-base guide/ Docs in AmigaGuide(tm) format gcc263-doc man/ this is the root for tons of man pages gcc263-doc bin/ this is /bin, and contains all gcc263-c binaries of this distribution that gcc263-c++ are meant to be directly invoked by gcc263-utils the user (contrary to the executables in lib/gcc-lib/, that are meant to be invoked by a driver program like gcc) lib/ normal (not base relatives) libraries gcc263-inclib lib/libm020/ normal 68020 libraries gcc263-inclib lib/libb/ base relatives libraries gcc263-inclib lib/libb/libm020/ base relatives using 68020 libraries gcc263-inclib lib/libnix/ Non-ixemul libraries gcc263-inclib lib/libm020/libnix/ Non-ixemul 68020 libraries gcc263-inclib lib/libb/libnix/ Non-ixemul base relatives libraries gcc263-inclib lib/libb/libm020/libnix Non-ixemul base relatives 68020 libs gcc263-inclib lib/gcc-lib/ home of compilers called by gcc gcc263-c gcc263-c++ gcc263-objc ixpipe/ a pipe handler needed by the library gcc263-base libs/ ixemul.library gcc263-base rexx/ ARexx wrappers for gcc and g++ gcc263-base src-patches/ source patches gcc263-diffs geninline/ Perl scripts to generate inline headers gcc263-inclib and -lamy glue A standard GNU directory tree structure has been adopted by the porting team to make life easier for everyone. Starting with the assign of GNU: to somewhere, we then have: GNU:amigados GNU:man GNU:Info GNU:etc GNU:bin GNU:include GNU:lib GNU:os-include GNU:os-lib Since one of the aims of porting GNU software is to be able to provide a Un*x like environment under AmigaOS, the porters decided to make a GNU "root directory" (GNU:) in which a standard Un*x directory tree can live, like /etc, /bin, and so on. This decision makes for good Un*x compatibility. Also, in the past, you would have to add many assigns (>=10) to have your GNU utilities working, and now it's just 5 (GNU:, MAN:, INFO:, ETC:, and BIN: ). We no longer have a conflicting assign LIB: (which most other compilers use). d. Common Installation Problems Using the Installer program for the ftp distribution should get rid of these problems. Check out the section on the specs file if you are having problems customizing your installation (say you have another compiler, and don't want to have 2 separate copies of os-include). Also, for those of you who refuse to read the README, you need the Commodore includes. Section 2 (d) tells you where to get them. See 2 (c) about the GCC object format. You may need to convert some standard libraries. Also check 2(a) on setting your stack properly. 2. Initial Problems with GCC a. Set your stack properly. GCC needs a lot of stack to run. For a nice, small (<1000 lines) source, a stack of 50,000 should do just fine. For more complicated code, you may need more, but 250,000 should do in any case, for gcc alone. Note that using make may increase the demands on the stack size, so the above suggestions may not hold. A word from Phillipe Brand's README: You need to have a 50.000 stack size in order to compile with GCC. This should be enough for most projects. Note than while recompiling gcc with itself it has taken more than 300KB stack. Stack can grow due to source complexity. Don't be afraid of it. To set the stack size, see the AmigaDOS Command 'stack'. To use ar and/or ranlib, 50KB is the minimum acceptable. You should have a much larger stack, if you use larger libraries. Starting with 2.6.3 a new environnement variable, GCCSTACK, enables gcc to read this variable and set stack upon startup. Thus now no need to set stack to huge values, only gcc/ld/cpp/cc1#? will automatically set new stack, according to GCCSTACK variable. Simply commit a: setenv GCCSTACK value to set gcc stack to value. Benefits: huge memory savings. ----------------end quote Note that GCC 2.7.0 will include automatic stack growth code, so you won't need to worry about this anymore. b. Amiga specific Library options i) Ixemul.Library This library was developed by Markus Wild when originally started porting GCC (up to version 2.3.3). It is a shared library that emulates a lot of Un*x functions, making life a lot easier for folks porting GNU utilities and such. Unfortunately, it is fairly resource hungry by Amiga standards, and has caused not a little irritation among Amiga users. GCC opens ixemul.library by default, so if you want to avoid it, you'll have to use -noixemul on the command line (see below in Amiga specific extensions and/or coding with x.lib) and link to one of the libraries below. The general rule of thumb should be to use ixemul.library if you're writing something non-Amiga specific (e.g. porting something) or one of the below link libraries when writing something specifically for the Amiga. Using ixconfig to work: ---------------------------------- [Joerg-Cyril Hoehle] > Which ixconfig should we use for ixemul v40.4? If you're using R.Luebbert ixem404lib.lha, use ixconfig from there. It's actually called bin in that archive :-( [generic] 2925 5396 54.2% -lh5- b464 Sep 3 1994 ixem/bin The ixemul.library from the gcc263 archive is R.Luebbert's 40.4. If you're using another one: Fred Fish's: I don't know whether he removed the ArpBase field in the include/library/ixemul.h file. R.Luebbert removed it between his 40.2 and 40.4, which was a big mistake IMHO (as this mail and others show). So I don't know which ixconfig is good for him. Leonard's: He told me that he had never removed this field, so get any old ixconfig you'll find. The one from the gcc263 archive will do. ii) libnix This is a standard link library to replace the functionality of ixemul.library. Make sure you link to it if you use the -noixemul command line option for gcc. Here's the readme file from the distribution: Short: A library for amiga specific development on gcc Type: dev/gcc Uploader: fleischr@izfm.uni-stuttgart.de Author: fleischr@izfm.uni-stuttgart.de, gnikl@informatik.uni-rostock.de This is libnix, a static (i.e. link) library for gcc 2.3.3 or above. It's not a replacement for ixemul.library (though it's possible to recompile most of the gcc environment with libnix) but a good thing for amiga specific development on gcc: * It's mostly compatible to SAS's way of handling things, i.e. you get even an automatic shared library opening feature and some other things you may miss in ixemul.library. This also means it's ANSI compliant. * It doesn't need any shared libraries than normal Amiga OS ones. * It is not copyrighted by the FSF. Therefore you neither need to include sources nor objects together with your executable. (read the GLGPL _before_ flaming on this statement) * And it's short! I was able to compile a 492 byte 'hello, world' using normal main. * It uses OS20 features whenever necessary. To cut it short: Use ixemul.library for porting Un*x programs, libnix for compiling amiga-only programs and gcc becomes one of the best Amiga compilers. There is no need to download this archive if you use gcc2.6.0 or above since the libraries itself are included with the normal gcc distribution. But if you use an older gcc version or if you want to get the sources you can take this package. But be warned: The ld that comes with earlier versions of gcc has some serious trouble with set elements. You cannot use libnix without the fixed linker that comes with gcc2.6.0. iii) Gerlib Obsolete iv) the PDC library Obsolete c. The object file format of GCC Yes, GCC uses its own object file format. This means you aren't currently able to link to AmigaDOS hunk format object files (standard). But hunk2gcc (by Markus Wild) will do the conversion for you as described below. Why does GCC use its own object file format? It's really just a design decision that allows other gnu utilities to deal with these objects (for example, gdb) without having to convert them to deal with AmigaDOS hunks. >From Brand's README-2.6.3 Starting from this release an AmigaDOS compliant library is provided, thanks to libnix authors (Matthias Fleischer and Gunther Nikl). Anyway if you want to rebuild one, there are two methods: 1) Using hunk2gcc; the AmigaDOS object converter made by Markus Wild. To achieve this, simply grab a copy of latest amiga.lib (from Commodore Development Kit) and make a new directory where you want your converted object files to go, cd into it, and enter hunk2gcc amiga.lib [..further libs if you like..] This generates an a.out object file for every program unit present in the hunk file (in this case, from amiga.lib). As the final step convert all those files into an a.out style library by issuing: ar qc libamiga.a obj.* ranlib libamiga.a The ranlib run builds a symbol table in the archive, and makes accesses to the library much faster. 2) Creating a libamiga.a library with libnix is fairly easy, but takes some time. Just uncompress sources.lha from libnix distribution and run a 'make libamiga.a'. NOTE: As long as you make no AmigaDOs specific calls, you can create a dummy library using: cat "int dummy;" >dummy.c gcc -c dummy.c ar crv libamiga.a dummy.o mv libamiga.a gcc:lib A small libamiga.a (dummy) is also provided with libnix. d. Getting the Commodore includes [Jochen Wiedmann] You can obtain the includes and some developer tools from any Fish CD and some other CD's as well. However, you will be missing the autodocs, which are definitely the best information source for the OS functions. You can get them either by buying the The Amiga ROM Kernel Manual: Includes and Autodocs, ISBN 0-201-56773-3 (about 50-60$, as far as I know), or by buying the so-called NDU (Native developers update kit, 5 disks, about 40$) from Fa. Hirsch & Wolf Mittelstr. 33 56564 Neuwied Germany Phone: (0049) +2631 83990 Email: hhhirsch@carla.adsp.sub.org Sorry, but this is still the *only* source for the NDU, at least as long as CATS (Commodore Amiga Technical Support) isn't alive. I'd prefer buying the NDU, because it contains the newer information than the book and I prefer it online. e. The frontend of the compiler amigados-gcc and gcc are the same thing. AREXX scripts: The provided ARexx scripts have been contributed by Loren J. Rittle. If you like ARexx, they're an alternate way of calling gcc. They automatically make sure you're using a large enough stack setting, and enable you to compile C++ programs with less obscure options. This approach is furthermore useful if you're not able to use the g++ /bin/sh script. f. Optimization may work where no optimization doesn't The folks who write GCC almost always use -O when compiling GCC and other stuff. Hence, the -O routines are better tested than those without optimization. So you may want to try it. g.) Using the OS functions with gcc. [Jochen Wiedmann, with a small suggestion by Christian Stieber] Let's write a simple "HelloWorld.c": /* Compile me with gcc -noixemul -o HelloWorld HelloWorld.c -lauto */ #include #include #include int main(int argc, char *argv[]) { struct EasyRequest er; er.es_StructSize = sizeof(er); er.es_Flags = 0; er.es_Title = "Message"; er.es_TextFormat = "Hello, world!\nintuition.library is at 0x8l."; er.es_GadgetFormat = "Ok"; EasyRequest(NULL, &er, NULL, IntuitionBase); exit(0); } Some notes: - We are using the function EasyRequestArgs() from intuition.library. Thus we have to include the appropriate headers: intuition/intuition.h for the structures and constants, proto/intuition.h for the function prototypes. Do not use headers from other compilers (for example pragmas/intuition.h), gcc headers (inline/intuition.h, included by proto/intuition.h) or even OS headers (clib/intuition_protos.h). The only exception are clib/alib_protos.h and clib/alib_stdio_protos.h: These represent link libraries and not shared libraries. - We did *not* open intuition.library. gcc does this for you by including proto/intuition.h and linking against libauto.a. However, this works only for OS libraries. Consult the GNU:libauto directory, if you want to know how to get the same possibilities for other shared libraries. Using proto/intuition.h you are safe and even source compatible to SAS/C and Dice. - If you *need* to open a library manually, do it as follows: #include #include #include #include struct IntuitionBase *IntuitionBase = NULL; /* Explicit initialization with NULL is a *must*! */ void Cleanup(void) { if (IntuitionBase) CloseLibrary(IntuitionBase); } int main(int argc, char *argv[]) { if (atexit(Cleanup)) { perror("atexit"); exit(20); } if (!(IntuitionBase = (struct IntuitionBase *) OpenLibrary("intuition.library", 37))) { fprintf(stderr, "Cannot open intuition.library, V37"); exit(20); } /* Same program as above */ } Note the use of atexit(), which makes exit() calling Cleanup(). Note this possible alteration: > struct EasyRequest er; > > er.es_StructSize = sizeof(er); This seems to be correct, but I suggest hardcoding 5*sizeof(ULONG) here. That field serves as version information; using sizeof() and compiling with newer headers would tell the OS that you're using the new structure. But the additional fields will contain garbage... 3. Amiga-specific extensions a. The CHIP keyword Most C compilers for the Amiga have a CHIP keyword or equivalent to signify that an object should stored CHIP memory when declared. I.e., CHIP struct Image *my_graphics = { ... } will cause the memory for the data to which my_graphics points to be allocated in CHIP memory (pretty handy, huh?) This is obviously Amiga-specific, so of course it's not implemented in the FSF code. BUT, the linker (ld) currently recognizes -chip and -fast options. However, these haven't been put in gcc itself quite yet, though I have been assured they will be here soon (there have been more important things to work on, like getting -resident to work again, see below). So if you're using 2.6.3 or below, here are some basic work-arounds: (I gotta find these) b. Command Line options i) -noixemul (maybe just reference 2. (b)) This option prevents gcc from making your executable open ixemul.library (of course, your program can still do so). Make sure you link to libnix or something similar to access those functions. See also 2. (b) (ii) ii) -fbaserel (currently broken until 2.7.0) [from Christian Stieber] -fbaserel turns on base relative adressing; which means that global/static variables are referenced with a 16 bit offset relative to an address register (a4 is generally used for that). The startup-code loads a4 with a pointer into the data segment. Result: every access to a global/shared variable is only 16 bit instead of the usual 32 bit address -> shorter & faster. It also means that you're limited to at most 64K of global/static variables. iii) -resident [from Christian Stieber] A "resident" program (the correct term is "pure") is a program that can be loaded into memory just once, but executed by serveral processes at the same time. Therefore, "resident" is actually "code sharing". -resident turns on -fbaserel and links with a special startup code. That first thing that special startup code is to allocate some memory, copy the global/static variables into it and load a4 with a pointer to that memory. Then, a normal startup-code and your normal program is run. When the program exists, the memory block is freed. Result: Every process has its very own, private data segment for global/static variables. Since there all global/static variables are accessed via a4+16-bit-offset (because of -fbaserel), the original data segment is untouched. Result: the program can be executed independently by several processes, with every process getting its own data segment. c. Startup Code options (may be redundant) 4. Advanced Questions a. How do I build a cross-compiler? [from Phillipe Brand] How to generate a cross-compiler, AmigaDOS side: ------------------------------------------------ - Get gcc-2.6.3.tar.gz from ftp.gnu.ai.mit.edu or mirror site - Get ftp.telesys-innov.fr:/pub/amigados-gnu/gcc-2.6.3-amiga.diffs file >From CLI: CLI> sh # zcat gcc-2.6.3.tar.gz | tar xvf - # cd gcc-2.6.3 # zcat ../gcc-2.6.3-amiga.diffs | patch -p1 # ./configure --host=amigados --target=CPU-COMPANY-SYSTEM # make When compilers are built, all you have to do is installing it using make install, and to grab other architecture's libraries (libc.a, etc...), and headers. More infos: See GCC AmigaGuide documentation, look for "Cross-Compiling". How to generate a cross-compiler, Other architecture side: ---------------------------------------------------------- - Get gcc-2.6.3.tar.gz from ftp.gnu.ai.mit.edu or mirror site - Get ftp.telesys-innov.fr:/pub/amigados-gnu/gcc-2.6.3-amiga.diffs file >From CLI: CLI> sh # zcat gcc-2.6.3.tar.gz | tar xvf - # cd gcc-2.6.3 # zcat ../gcc-2.6.3-amiga.diffs | patch -p1 # ./configure --target=amigados (host should be determined by configure itself) # make When compilers are built, all you have to do is installing it using make install, and to grab AmigaDOS GCC libraries (libc.a, etc...), and headers. More infos: See GCC AmigaGuide documentation, look for "Cross-Compiling". A working example of a cross-compiler running on sunos4.1.3 can be found in ftp.telesys-innov.fr:/pub/amigados-gnu/gcc-cross/.... b. Working with AmiTCP > Could somebody please explain where exactly gcc and the libraries > that come with it fall down when it comes to using AmiTCP? > > What's missing etc. > > Then, Why can't (I assume you can't since veryone says you can't) > use the C= Developers kit with gcc without fiddling. The problems are: a) gcc uses it's own object module and library format, so lots of the GNU tools don't have to change too. b) gcc uses ixemul by default, which is not compatible directly with the SAS libraries supplied with AmiTCP. a) is just a design decision, and indeed there are tools to convert. (b) is harder to see; basically the SAS AmiTCP libraries assume the environment of SAS/C, complete with references to the SAS FILE structure internals and the functions which aren't in SAS or which need to be patched to work with AmiTCP. In addition, the netincludes supplied with the AmiTCP port include many items and structure definitions which are also defined in gcc, albeit differently, resulting in a very confised compiler if you mix them. Unfortunately, a simplistic port (recompile with the right -I directives) fails for two reasons; the includes are wrong, and even if they were matched, the SAS-directed library patches the wrong things. Resulting in an executable which will fail to operate as expected (it might work OK, or it might not - depends what you call & when). What is needed is a new library - take out the duplicate definitions, patch the right functions and remake the library for gcc. Does this help? Peter Ivimey-Cook. c. Writing code for ixemul.library After looking at Markus Wild's README, I couldn't see anything to include in particular, and the thing is way too long to put here. Any specific suggestions of stuff to clip out? Actually, considering the update activity, the current maintainers/ developers should be providing a new README. Particularly the functions available in the different versions (regular fxns, the ones in the network version, and the conflicts. Somebody recently answered a question about this last part, saying he was trying to integrate the 2). i) Finding bugs with trace() [Joerg-Cyril Hoehle] more than one year ago, I corrected a huge bug in the program trace (a buffer overflow). Together with ixemul.trace, this program allows you to see every ixemul call (aka SnoopDOS for ixemul.library programs) and may be useful for debugging. Before, because of that overflow, trace hung the whole system very often (and very soon) when tracing every ixemul call. Now, I've been able to trace long sessions of gcc compilation to a (K)CON: window without trouble. d. Writing code for libnix Maybe a referral to the libnix docs in the source file, or should I stick them in here (I believe they're only in the source archive, so people may tend not to grab it)? e. Problems with asm() [Insert your answer here,please] f. How do I save RAM? [Jochen Wiedmann] A simple try is to do a setenv TMP MyHardDrive:t before compiling. This will create temporary files on the harddisk and not in RAM. Depending on what you compile this can save some 100K of RAM while compiling and it doesn't slow down the compiler very much. Another try is to turn off optimization. This prevents the "inline" headers from being included, which need *very* much RAM. If you like optimization (for example because of the additional checks!) you can use the options -O -D__NOINLINES__ which turn on optimization, but prevent the inline headers from being included. (The negative effect is, that you need an additional function call any time, you use an OS function, but one can live with it.) g. Writing shared libraries and resident (pure) code Here is some discussion from the amiga-gcc-port mailing list [not cooked up yet] [Peter Ivimey-Cook] > Obviously there must be other solutions. Being able to build a library > is not far away from being able to build a pure (=residentable) > program, and gcc-2.3.3 was both largely above 64KB in size and > residentable. So what's the trick? > Don't use global data referenced absolutely! residentable implies reentrant, which means that global data is read-only apart from the first instance (i.e. it's OK to initialise it, once, but once initialised it can't be modified or reinitialised). This effectively means the only data used in the must be on the stack or referenced from it by pointers, as the stack is separate for different invocations of a resident program. Of course this does pose a major headache for programs which use lots of global data. And watch out for library use too! Failing this, one must use a global data access mechanism which allows: (a) the base address to be set up - e.g. by calling AllocMem() and storing the result in a register (NOT a global!) (b) *all* accesses to "global" data are relative to this register. [Peter Ivimey-Cook] > > BUT: If we build a shared lib from the static link lib, code exists only once > and is SHARED between callers of the lib (callers means different executables, using the lib) > Here the problem comes up that the DATA MUST not be shared between different callers. > This is because it is PRIVATE to the individual caller. > IMHO the only way to realize multiple callers is the use of -fbaserel, since now access > is relative to the A4 register and thus, every program can have IT'S OWN data, without > interferring with each other. It depends on the way the library is constructed. Most Amiga libraries are written to be re-entrant - i.e. they don't use global data items which would require loading - they use the stack. For items which can't be done this way, there are two solutions - 1. Use the library data space. It's global, unless you make provision for it to be otherwise - and availabl easily off a6. 2. Manage your own data space, using tags or whatever to allocate space for callers. Note that the only indication you have as to who is the current caller is the result of FindTask(). So, in general libraries try to make required data passed through parameters, or available on the stack, or shareably global. In which case no baserel or anything else is required. [Jochen Wiedmann] - Of course it is possible to write shared libraries (ixemul.library, for example) with gcc, but *only* if the source is written with a shared library in mind. - It is possible (Matthias told how) to develop a solution that would allow to use existing source of link libraries for shared libraries (libbfd sources, for example) with at most very minor modifications - -fbaserel would be a must for the above solution. On the other hand it is not possible yet, at least not without going to the assembler level. (Really deep, not just using some headers with asm keywords.) [Christian Stieber] > Jochen Wiedmann writes: > > - Of course it is possible to write shared libraries (ixemul.library, > > for example) with gcc, but *only* if the source is written with a > > shared library in mind. > > Could you elaborate on that so that we know a little more about it, > please? - keep in mind that all callers use the same global and static variables; that means: either protect them with semaphores (e.g. if your library is using a memory pool), or allocate a block of memory inside your library and pass pointers around. - the "startup" code will contain the ROMTag structure and related data, as well as stubs to push parameters on the stack, calling the C function, cleaning up the stack. - I generally write the libOpen(), libClose(), libInit() etc. functions as normal C functions, but some people might prefer to put them into the "startup code". That's about all. It also means that many link-lib functions (e.g. from libnix) can't be used; this has to be decided on a per-function basis. It's pretty safe to say that you can't use stdio and malloc(). strcpy(), isdigit() are safe, etc. Just be careful. Think about what the function does; that should give you an idea whether you can use it or not. [Jochen Wiedmann] > Jochen Wiedmann writes: > > - Of course it is possible to write shared libraries (ixemul.library, > > for example) with gcc, but *only* if the source is written with a > > shared library in mind. > > Could you elaborate on that so that we know a little more about it, > please? > > Does it mean that you can't use any global variable? Or only one but > declare it with something like__asm__("a4") and use it as a base > pointer? Does it imply that you must use the inline/*.h? Does it imply > that you must use a special startup code (of course :-), what must it > do? In short: You may well use something like struct ExecBase *SysBase; and read this variable from anywhere within your program, because you can a) initialize it from the library startup code and b) be sure that it never changes but you must not use global or static data otherwise. h. I want to change from SAS/C to GCC Great! [Kriton Kyrimis] > I am also wondering if it is possible to link object files that were > created by different conpilers, specifically GCC and SAS. I need to do Try compiling SAS programs using DATA=FAR CODE=FAR, then use the hunk2gcc program to convert the object files from amiga format to sun format. One problem with this approach would be with floating point code and with integer multiplication/division. If you can compile with CPU=68020 (or higher) and MATH=68881, then SAS/C will produce inline code, and you will have no problem. Otherwise, you'll have to grab the necessary modules from the SAS libraries and convert them to sun format as well. (Compiling with the option that produces code that uses utility.library, and linking with -lauto might alleviate the problem.) ---------------------------------- And here's a little AREXX program to automagically convert sas libs to gcc libs: This arexx program is based on help from Philippe BRAND. /********************************HISTORY********************************** * 18-DEC-94 : DJR : Created * ************************************************************************** * * * Name : saslib2libgcc * * Author : David J. Ruscak ruscakd@polaroid.com * * Created : 18-DEC-94 * * Purpose : to convert SAS libraries to GCC libraries * * * *************************************************************************/ /* To convert a SAS library to a GCC format library several steps are involved. 1) cd to RAM: Either copy the library to RAM: or give the full path name. EX: gnu:saslib2libgcc X11:lib/Xtnb.lib 2) You need the SAS Librarian 'oml' with a version => Library otherwise you'll get an error such as 'unknown hunk'. 3) You need hunk2gcc, mv, ar and ranlib. If your using gcc you have the fastram, VMM2.1 swap works fine too. */ /* TRACE R */ SIGNAL ON ERROR OPTIONS RESULTS IF ~SHOW('l','rexxsupport.library') THEN IF (~ADDLIB("rexxsupport.library",0,-30,0)) THEN DO ECHO "Can't load rexxsupport.library" EXIT 10 END fullname = "" PARSE ARG fullname IF fullname == "" THEN DO ECHO " No Library to convert" ECHO "" ECHO " EX: gnu:saslib2libgcc X11:lib/Xtnb.lib" EXIT 10 END IF (~EXISTS(fullname)) THEN DO ECHO "Can't find" fullname EXIT 10 END ADDRESS COMMAND /* make 2 directories for object files */ IF EXISTS('object') THEN DO ECHO "Directory object exists Please delete or rename it." EXIT END 'makedir object' IF EXISTS('object2') THEN DO ECHO "Directory object2 exists Please delete or rename it. " EXIT END 'makedir object2' 'oml -oobject 'fullname 'X *' /* SAS Object Librarian */ IF EXISTS('object') THEN DO objfiles = SHOWDIR('object','F') /* list of all object files */ nfiles = WORDS(SHOWDIR('object','F')) /* count of object files */ DO UNTIL nfiles == 0 /* must be a more efficient way... */ 'hunk2gcc >NIL: object/'WORD(objfiles,nfiles) /* convert to GCC object */ 'mv >NIL: obj.#? object2/'WORD(objfiles,nfiles) /* restore original names*/ nfiles = nfiles - 1 END END 'cd object2' /* directory of converted objects */ path = PRAGMA('D','object2') IF (INDEX(fullname,':') > 0) THEN /* strip out assign */ fullname = SUBSTR(fullname,LASTPOS(':',fullname) + 1) IF (INDEX(fullname,'/') > 0) THEN /* strip off directories */ libname = SUBSTR(fullname,LASTPOS('/',fullname) + 1) ELSE libname = fullname dot = POS('.',libname) gcclibname = 'lib'SUBSTR(libname,1,dot)'a' 'ar qc 'gcclibname' #?.o' /* archive new library */ 'ranlib 'gcclibname IF EXISTS(gcclibname) THEN DO needslash = LASTPOS(':',path) IF ~(needslash == LENGTH(path)) THEN path = path'/' ECHO " The following indicate a hunk2gcc error:" ECHO " Short reloc into N_DATA, what should I do? ECHO " Short reloc into N_BSS, what should I do? ECHO "" ECHO "" ECHO " Remember to move/rename "path"object2/"gcclibname" appropriately ." END ELSE ECHO "No library created, what could have gone wrong ????" EXIT ERROR: ECHO " error = "RC ECHO " Hope its not the dreaded *Invalid Hunk type*" EXIT h. Inline Headers [Christian Stieber] (maybe should clear out the specific bgui reference) > compiling bgui:demos/font.c: > gcc -O3 -noixemul bgui:demos/font.c > > got errors: > font.c:100: unterminated macro call This is because the inlined varargs aren't useful at all. Do #define NO_INLINE_STDARG (or whatever it is called), and create varargs stubs like this: Prototype: BOOL SomeFunction(int, char *, ...); and BOOL SomeFunctionA(int, char *, APTR); Stub function: #define NO_INLINE_STDARG #include BOOL SomeFunction(int A, char *B, ...) { return SomeFunctionA(A,B,(&B)+1); } That isn't really ANSI-C, but it works on all Amiga compilers I know of. Create a .c file for every function, compile, ar and link it to the program. Make sure the optimizer is on, else SomeFunctionA won't be inlined. How to create stubs for non-varargs functions: In order to be able to compiler without -O, you want to create stubs for the inline functions as well: #define NO_INLINE_STDARG #define SomeFunctionA InlinedSomeFunctionA #include #undef SomeFunctionA BOOL SomeFunctionA(int A, char *B, APTR C) { return InlinedSomeFunctionA(A,B,C); } Again, create such a file for every function and put the *.o files into the archive mentioned above. j. Writing Hooks [Tomi Ollila] > 1) How to I write a hook function - I.E. Specify which registers parameters > should be received in and that the result should be returned in register D0. One way to do this is to write the function entry the following way: ULONG hook(void) { register char * a0 __asm("a0"); char * buffer = a0; register ULONG d0 __asm("d0"); ULONG size = d0; ... return length; } We wrote some macros for AmiTCP/IP project to make it easier to write this kind of function entries... #define RAF2(funname, type1, arg1, reg1, type2, arg2, reg2) \ funname(VOID) \ { \ register type1 reg1 __asm(#reg1); \ type1 arg1 = reg1; \ register type2 reg2 __asm(#reg2); \ type2 arg2 = reg2; would make the above function look like: ULONG RAF2(hook, char *, buffer, a0, ULONG, size, d0) #if 0 { #endif The whole macro file `amiga_raf.h' is available at kampi.hut.fi, directory /AmiTCP. It defines RAF1 - RAF7 for both GNU C and SAS C compilers (someone could write these to DICE as well). > 2) Is there anyway to specify that a function should NOT be made inline (As > I wouldn't want the hook function inlined. Compile it in a separate module. This way gcc doesn't see the function code when compiling. If you give a pointer to the function, then it must also exist as a callable function. 5. C++ and Objective C a. C++ i) Use _complex.h instead of complex.h Because Amiga Dos is not case-sensitive, and there are both complex.h and Complex.h on the Unix distribution, one had to be changed. This choice was recommended by the libg++ maintainer. (don't know much about these, except that they're not as well developed yet as the C compiler) 6. Notes For Un*x Hackers [additions to this section are welcome] a. fork() is not implemented Here's a clip from Markus Wild's README in the ixemul source distribution: Some notes to vfork() and friends ================================= **IX/BSD process management is one of the most nasty design differences between AmigaDOS ans **IX/BSD. `fork()' for example is hardly possible to implement on AmigaDOS, as it requires to create an identical copy of the parent process. This is only feasible with virtual memory, where processes can be mapped at equal places in memory. Under AmigaDOS this would have to be simulated by copying of stack and malloc'd data whenever a process is activated, and copying them to a safe place before it is disactivated. This problem can be avoided, if the program to be run under AmigaDOS is only fork()ing, because it just wants to start another process. In that case, no such copying as described before is necessary, and BSD therefore invented the `vfork()' function, which works like `fork', but runs the child on the parents memory segments (stack and malloc'd data). While the child is using the parents resources, the parent is sleeping in a not interruptible state. That much for theory;-) I tried to implement an as compatible as possible vfork() function, that behaves like the BSD one. This should work under any OS version, for Kickstart 1.3 the arp.library is used, starting with OS 2.0 dos.library is powerful enough to do it itself. Since I won't try to implement `fork', I provided a possible alternative (you tell me;-)). As an extension, you get the `vfork_resume()' function, which causes the parent to resume, just like it would if you called `_exit()' or one of the `exec*()' functions. Since this function is quite dangerous (and an even bigger hack than vfork() itself..), here's what's happening in `vfork_resume()': o the child switches to its own stack. After vfork(), the child is using the stack of the parent process. Since no two processes can share the same stack in parallel, vfork_resume() causes a switch to the `real' stack of the child. o the parent is sent a wakeup message. o both processes run concurrently The first point is the most important one: Since vfork_resume() changes the stack pointer of the running process, you can't refer to any variables or parameters anymore after calling vfork_resume()! Only register variables survive such a call, and you have to explicitly store values in register variables that are subject to survive! There's another potential problem with vfork_resume(): ************************************************************************** Don't exit() from the parent before all vfork()'d children have died!! ************************************************************************** Since exiting from the parent causes the parents code and data segments to be deallocated, the child would find itself without code space to run on, and would probably cause a severe machine crash! So always call at least `wait (0)' before returning from the parent. exec*() ======= In most cases, you just use `vfork()' to later overlay the process with a new image, that is you want to start another program. The way AmigaDOS loads processes is not too well suited to do `exec' style program starting, yet it is possible, although with slight resource wasting.. First problem is, that all exec* functions pass an argument vector to the new program, whereas AmigaDOS programs expect to be passed an argument line (instead of the vector of arguments). Since in my opinion it would be a good thing if a program could get an argument vector directly (in that case the inherent problem of passing multi word arguments to a program would be finally solved, no more weird quoting needed!). That's why I provided a mechanism that allows this vector passing, and it works like this (look at crt0.c for a concrete implementation of this concept!): The program has to provide a magic header at the first executable location in its code. This magic header looks like this: o JMP instruction to common AmigaDOS startup o struct exec area. Use the OMAGIC a_magic code. o provide an alternate entry vector in a_entry. execve() jumps thru this vector to pass vectors to your program, instead of going thru the normal AmigaDOS startup part. As long as you use my crt0.o and libc.a, this whole thing is completely transparent to your program, you only have to care for it, if you want to support the mechanism in other languages as well. The second problem is how to start `old' AmigaDOS programs from execve(). If the program has the described magic header, starting is easy. Else another approach is taken, depending on the OS version. Common to both OS versions (1.3 and 2.0) is redirection of stdin and stdout. Since the new program can't refer to the real file descriptors (I can't pass the open library without my startup code), I have to setup DOS fields to use my filehandles. This may succeed or not, depending on whether the descriptors in question are realized by DOS files or not (in the future a not-compatible alternative would be descriptors that refer to sockets!). Actual starting of the program is done with RunCommand() under 2.0, and some own hack under 1.3. If someone is interested to get this working well under 1.3, I'd be happy to include a better starter function, my kludge doesn't particularly deal graceful with BCPL functions... 7. Support Utilities a. Frontends I'm thinking of GUI's here. There is one (something like GCC_Frontend) on the April 1994 Aminet CD-ROM collection. If I remember correctly though, when I tried using it, it bugged out on me. Anybody know of anything else? b. Debuggers (Maybe just GDB?) GDB is the GNU debugger, and Fred Fish is the port master for it. Here is something from his readme. GDB 4.12 has been ported to the extent that you can build an AmigaDOS executable that knows how to load and examine executables from non-AmigaDOS systems. Much work remains. See the gnu:src/diffs/gdb-4.12-README file. [He's referring to the general gnu:source directory. If you unpack the gdb source archive to your GNU: directory, it should be in that same directory.] c. What else? 8. Additional Support a. amiga-gcc-port mailing list b. snail mail (?) - for people getting it off of CD-ROM c. You never know, it could be a problem in base FSF code gnu.gcc Newsgroup and daughter newsgroups ftp: prep.ai.mit.edu in /pub/gnu is the primary archive for the GNU Project. If you need the FSF baseline code for something, it'll be there. 9. Known Bugs a. C compiler i) -resident option This has been broken since 2.3.3. Hence, you'll need to get that version if you want to compile a program you can easily make resident. Of course, you can always write your code to be resident, it just takes a little more effort (see above). This is fixed in 2.7.0. ii) -fbaserel option Won't be fixed until the -resident option is (as it's the cause of the -resident problem). b. C++ compiler (g++) c. Objective C compiler 10. Maintainers and Contributors Gcc v2.2.2 port: Markus Wild Gcc v2.3.3 port: Markus Wild Gcc v2.4.5 port: Philippe Brand, Lars Hecking, Fred Fish Gcc v2.5.0 and up: Philippe Brand, Fred Fish, Leonard Norrgard Ixemul.library: Markus Wild, Leonard Norrgard, R. Luebbert Libnix: Matthias Fleischer, Gunther Nikl Also, much testing, suggestions and debate have been provided by Jorg Hoehle Peter Ivemey-Cook Christian Stieber Walter Harms Lars Hecking Kriton Kyrimis Niklas Hallqvist Jochen Wiedmann Thomas Walter not necessarily in that order. [ I just compiled this list from looking at the 1995 archive and some of the '94, whoever consistently either (a) made suggestions and wrote some code or (b) has been reporting bugs from attempts at compiling GNU software for a long time (i.e. not someone who's just having problems porting their favorite package). If I've left someone out, or if you have a problem with my criteria, or if you think I should just thank the mailing list as a whole, mail me]. The present FAQ maintainer is Lynn Winebarger (owinebar@indiana.edu). However, you should send corrections to amiga-gcc-port@nic.funet.fi for a look over, as that's what I'll do anyway. [or maybe they should just send them to me, and I can screen out the very common screw-up from people who think they know.] Flames directed to /dev/null. 11. Future 12. History Appendix A: How does the GPL affect programs compiled by GCC? [Niklas Hallqvist] > If I compile one of my sources with gcc, do I have to > distribute it under the GPL? I'm not using any GNU libraries or > startup or anything. My interpretation of the GPL is that i > don't have to, but one of the teacher (lecturers?) say that I > have to. Who's right? You are. The GPL (and LGPL) only covers distributed code, not code generated by tools which are GPL:ed. As there are code which get linked by GCC distributed with GCC, libgcc.a, one might think that the GPL would apply to generated code, but no, libgcc.a is written without GPL just to enable that use of the GNU compiler. Think about it, several object-only commercial applications including OSes are compiled by GCC for enhanced performance, would they do that if they had to give up their objectcode-only policy? The key behind the GPL is that no code based on work GPLed, should be locked to to that specific compilation, but the user should be able to customize the GPLed code as he wants to and recompile. You are still allowed to protect your own source as you see fit. For example, let's say you use som LGPLed library libfoo.a in your application app. When you distribute it you must also distribute a linkable objectfile app.o which, when linked with the distibuted libfoo.a (with source) generates app. If the end-user wants, he should be able to customize libfoo, and relink it with app.o to get a customized app. It's not a very severe limitation on distribution rights IMHO. Remember that the (L)GPL was written to give programmers more Freedom, not to limit their chance of protecting their own work. As long as their own work is clearly delimited from others GPLed work, it's perfectly OK to keep a separate copyright policy on it. In your case, your lecturer has misunderstood the intentions of the (L)GPL which is easy to do. These discussions come up ever so often because of the legaleze used to express the GPL. Too bad many get the GPL wrong, as rumours like "you cannot use GNU products for business work" severly harms the usage of GNU products. To name a few uses of GCC where the program did NOT fall under the GPL: Dell Unix SVR4, NextStep & some OSF/1 port. Tell your lecturer about these, and ask him why he thinks the GPL prevents them to be sold during other than GPL conditions? I ask you to actually convince him he was wrong as it is harmful for the programming society that such misconceptions exist. This is where I'd like to put that list of libraries that are GPLed. Appendix B: GNU Porting activities and who's dealing with them Not really GCC (Which is why it's an Appendix), but certainly one of GCC's main uses. It would be good to keep track of these things anyway, in case someone wants to start working on porting something (just as administratvia). Appendix C: Another Free C compiler (for those without much memory) Warning, apparently lcc is not yet fully ported (no backend) for easy use. [Fred Fish] > > One of the things I hope to eventually add to my FreshFish CD is a > > port of lcc, which is much smaller and faster than gcc, and apparently > > somewhat more ANSI compliant. I just haven't had time yet to look > > at doing so. > > What's lcc? I never heard of it. ============================ begin inclusion =========================== lcc is the retargetable compiler for ANSI C described in our book `A Retargetable C Compiler: Design and Implementation' (Benjamin Cummings, 1995, ISBN 0-8053-1670-1), which will be available in December 1994. lcc is in production use at Princeton University and AT&T Bell Laboratories. The public distribution directory contains the following files. README this file. install.{ps,txt} describes the distribution and gives installation instructions. install.ps is the PostScript generated from the HTML document, install.html, which is included in the the distribution. X.Y.tar.{Z,gz} a compressed tar files for the distribution of version X.Y, e.g., 3.0.tar.Z the tar file compressed with compress. This distribution includes user documentation, the front end, the driver program, code generators for the SPARC, MIPS R3000 and x86, and the code-generator generator that produced them. A .gz file is the tar file compressed with gzip instead of compress. The distribution is available via `anonymous' ftp from ftp.cs.princeton.edu (128.112.152.13) in the directory pub/lcc. Obtaining and extracting the distribution into its own directory is accomplished by the following commands. Replace `3.0' with the latest version, which is the only one typically available; versions like `3.0a' identify minor updates and versions like `3.1beta' identify pre-releases (which might be incomplete). As suggested, use your login as the password, and use ftp's binary transfer mode. mkdir lcc cd lcc ftp ftp.cs.princeton.edu anonymous yourlogin cd pub/lcc binary get 3.0.tar.Z dist.tar.Z quit zcat dist.tar | tar xpof - rm dist.tar.Z To be added to the lcc mailing list, send a message with the 1-line body subscribe lcc to majordomo@cs.princeton.edu. This line must appear in the message body; `Subject:' lines are ignored. To learn more about mailing lists served by majordomo, send a message with the body `help' to majordomo@cs.princeton.edu. Additional information about lcc and about our book is available on the WWW at URL http://www.cs.princeton.edu/software/lcc. Chris Fraser / cwf@research.att.com David Hanson / drh@cs.princeton.edu Thu Aug 18 13:36:15 EDT 1994