New version of gcc for the mac
Jonathan Kimmitt
A new version of gcc for the Macintosh is now available.
It is now on nic.switch.ch, software/mac/src/think_c/GCC_1.37_demo.sit.bin
Due to bandwidth limitations it will initially be available only in binary
form, however the source code is very similar to what is already available from
that site.
The main differences between this version and the previous one are that the
compiler engine is based on gcc-1.37 for MPW, and as such understands MPW-3.2
style include files directly, and secondly a significant effort has been made
to make the interface more mac friendly. The framework is based on THINK Class
Library, access to the compiler tools is via windows and menus, and a certain
amount of documentation is built into the program. Other than the toolbox
includes, only three files are needed to get going, but some example programs
are (hopefully) included.
Learning the lesson from the previous version, this program has been tested on
an LC-III and a Quadra 900 as well as a PowerBook 100. The compiler
application runs only under sys-7 and requires a machine with about 8M of
memory to compile non-trivial programs. The applications generated are
potentially sys-6 capable, depending on what toolbox facilities are called.
This version is described as a 'demo' since it lacks a standard C library, and
facilities for automated builds, scripts etc., and it generates applications
only. Only MacsBug debugging is supported. However it is in no way crippled
and these features could (I won't say easily) be added if there is interest.
Like the previous version, this program is Postcardware.
Jonathan Kimmitt
>A new version of gcc for the Macintosh is now available.
>It is now on nic.switch.ch, software/mac/src/think_c/GCC_1.37_demo.sit.bin
>Due to bandwidth limitations it will initially be available only in binary
>form, however the source code is very similar to what is already available
>from that site.
Before I get deluged with people telling me that I am in violation of the GNU
license, I just want to make it clear that this program is derived from and
uses GNU source code which is already available from the same site, and if
anybody wishes to read, reuse, improve this source code they are most welcome
to. It also incorporates class libraries which are available with a standard
development system for that machine. If anyone would rather work with the
hacked source code rather than the unadulaterated version, they can
send me a 90-meg Bernoulli/Iomega cartridge, with return postage paid, to get
the latest version.
Howard Berkey
>Before I get deluged with people telling me that I am in violation of the GNU
>license, I just want to make it clear that this program is derived from and
>uses GNU source code which is already available from the same site, and if
>anybody wishes to read, reuse, improve this source code they are most welcome
>to. It also incorporates class libraries which are available with a standard
>development system for that machine. If anyone would rather work with the
>hacked source code rather than the unadulaterated version, they can
>send me a 90-meg Bernoulli/Iomega cartridge, with return postage paid, to get
>the latest version.
And before he gets deluged (again :-) the last sentence of the above
fully satisfies the copyleft. He could even have charged for time
spent copying the sources, had he wished.
-H-
--
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
Howard Berkey how...@netcom.com
Plato, they say, could put it away... Half a pint of whiskey every day!
...........................................................................
Johnathon Suker
>>development system for that machine. If anyone would rather work with the
>>hacked source code rather than the unadulaterated version, they can
>>send me a 90-meg Bernoulli/Iomega cartridge, with return postage paid, to get
>>the latest version.
>And before he gets deluged (again :-) the last sentence of the above
>fully satisfies the copyleft. He could even have charged for time
>spent copying the sources, had he wished.
This is not meant as a flame and maybe better of in another newsgroup, but
since it started here, I will ask.
What exactly consitutes "availability". This only satisfies people that own
Bernoulli drives. Which is in itself exclusive, therefore not fully available.
Does anyone know the "ins-and-outs" of the GPL.
THIS IS NOT MEANT AS CRITICISM, MERELY A QUESTION TO WHAT I FEEL IS A VERY
UNFORCABLE PUBLIC LICENSE, I.E. I DONT AGREE WITH GNU'S POLICY'S FORE SEVERAL
REASONS, THIS BEING ONE OF THEM. THEY SEEM TO BE VERY SELECTIVE IN THER
"OPENNESS".
All flames to /dev/null
Johnathon
--
Johnathon Suker | Thats SULTAN vile betrayer to you!
University of California, Irvine |
University Library | Iago
JLS...@UCI.EDU | Aladdin
Bill Johnston
Jonathon Kimmit writes, with regard to his offer to provide gcc sources:
>>> If anyone would rather work with the
>>>hacked source code rather than the unadulaterated version, they can
>>>send me a 90-meg Bernoulli/Iomega cartridge, with return postage paid,
>>>to get the latest version.
>how...@netcom.com (Howard Berkey) writes:
>>And before he gets deluged (again :-) the last sentence of the above
>>fully satisfies the copyleft. He could even have charged for time
>>spent copying the sources, had he wished.
>This is not meant as a flame and maybe better of in another newsgroup, but
>since it started here, I will ask.
>
>What exactly consitutes "availability". This only satisfies people
>that own Bernoulli drives. Which is in itself exclusive, therefore
>not fully available.
>Does anyone know the "ins-and-outs" of the GPL.
Of course. The FSF does. Questions regarding GPL compliance for
Mr. Kimmit's latest distribution of gcc for the Macintosh have been
referred to Len Tower (g...@prep.ai.mit.edu).
In this case, source code availability is a fairly minor problem.
According to Mr. Tower, Mr. Kimmitt's original announcement was a clear
violation of the GPL. Quoting from Mr. Tower's mail to Mr. Kimmitt:
[ Announcement ]
>> A new version of gcc for the Macintosh is now available.
>> It is now on nic.switch.ch, software/mac/src/think_c/GCC_1.37_demo.sit.bin
>> Due to bandwidth limitations it will initially be available only in binary
>> form, however the source code is very similar to what is already available
>> from that site.
[ Tower's reply ]
>Under the GPL, you MUST make the source code available at the SAME
>time you make the binaries available. If you have bandwidth
>limitations, you make the source code available first.
With regard to the second, amended announcement, the relevant part
of the GPL is as follows:
[ quote from text of GPL, copyright FSF, etc ]
3. You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the terms of
Sections 1 and 2 above provided that you also do one of the following:
a) Accompany it with the complete corresponding machine-readable
source code, which must be distributed under the terms of Sections
1 and 2 above on a medium customarily used for software interchange; or,
b) Accompany it with a written offer, valid for at least three
years, to give any third party, for a charge no more than your
cost of physically performing source distribution, a complete
machine-readable copy of the corresponding source code, to be
distributed under the terms of Sections 1 and 2 above on a medium
customarily used for software interchange; or,
c) Accompany it with the information you received as to the offer
to distribute corresponding source code. (This alternative is
allowed only for noncommercial distribution and only if you
received the program in object code or executable form with such
an offer, in accord with Subsection b above.)
Mr. Kimmitt's offer falls into category (b), above. Mr. Suker's
objection is reasonable, in my opinion. The GPL does not dictate
what constitutes a "medium customarily used for software interchange",
because this changes rapidly with changing technology. However,
I think that it is clear that Bernoulli cartridges do NOT meet
this criterion. [ The situation might be different if Mr. Kimmitt
had chosen to make his binary distribution only on Bernoulli cart,
which would guarantee that any recipient of the binaries would have
reasonable access to source distributed on the same media, but that
is not the case here. ]
So much for the _letter_ of the GPL. In what has evolved as
commonly accepted practice, binary distributions of GPL'd software
are typically accompanied by full sources, released at the same
time as required by the GPL, and available at the same ftp site.
John Coolidge's distributions of GNU software for Apple's A/UX
on wuarchive.wustl.edu:/systems/aux/gnu provide an excellent
example of the RIGHT WAY to redistribute GPL'd software.
For many reasons, not the least of which is the three-year
obligation imposed in section 3b of the GPL, cited above,
the simplest and best way to distribute GPL'd software is
to include binaries and full source in the same package.
According to a recent communication with Mr. Kimmitt, this
is exactly what he has decided to do; a copy of the full
source should have arrived at nic.switch.ch by now.
Unfortunately, a closer examination of the gcc binary that
was included in Mr. Kimmitt's distribution reveals other,
more serious GPL violations. The most obvious of these is
the integrated text editor, which includes a copyright and
license statement that clearly violates the GPL in several ways.
Mr. Kimmitt has further indicated that he is working to address
this issue. [ I might add that FSF has been remarkably prompt
and professional in explaining how the distribution can be
brought into compliance with the GPL, this despite their well-known
political objections to porting software to the Apple Macintosh. ]
For those who have read this far and wonder what this is all
about, copies of the GPL and other FSF position papers are
available on prep.ai.mit.edu in /pub/gnu/GNUinfo.
I do not speak for FSF, but having used and maintained GNU
software on a variety of platforms from Sun workstations to
my own Macintosh, I've come to appreciate the sensibilities
that underly the GNU project and the GNU General Public License
which has made its success possible. Proof of this can be
seen in freely redistributable Unix clones such as Linux,
and even in the MacBSD and MacMint projects, which have largely
been made possible by the existence of free software like gcc.
When a fully GPL-compliant distribution of Mr. Kimmitt's gcc
is finally made available, everyone will benefit.
--
-- Bill Johnston (john...@me.udel.edu)
-- 38 Chambers Street; Newark, DE 19711; (302)368-1949
Chris Wysocki
>Unfortunately, a closer examination of the gcc binary that
>was included in Mr. Kimmitt's distribution reveals other,
>more serious GPL violations. The most obvious of these is
>the integrated text editor, which includes a copyright and
>license statement that clearly violates the GPL in several ways.
As the author of said text editor, I find it rather ironic that you
would take objection to its inclusion in Mr. Kimmitt's distribution of
GCC. While you are correct in noting that the terms of the license
agreement for my text editor are not compliant with those of the GNU
General Public License, in my opinion there are far more serious
transgressions in Mr. Kimmitt's distribution of GCC. Most notably,
his shell application is based on and requires Symantec's THINK Class
Library, a collection of copyrighted commercial source code which,
unlike my text editor, cannot be freely distributed. Without the TCL
source code, it is virtually impossible to compile Mr. Kimmitt's
version of GCC, which to me violates the underlying spirit and
intention of the GNU GPL.
Chris.
--
------------------------------------------------------------------------------
Chris Wysocki Internet: wys...@netcom.com
Software Engineer America Online: AFA ChrisW
Global Village Communication Compuserve: 72010,1140
Matthias Neeracher
>>> A new version of gcc for the Macintosh is now available.
>>> It is now on nic.switch.ch, software/mac/src/think_c/GCC_1.37_demo.sit.bin
>>> Due to bandwidth limitations it will initially be available only in binary
>>> form, however the source code is very similar to what is already available
>>> from that site.
>[ Tower's reply ]
>>Under the GPL, you MUST make the source code available at the SAME
>>time you make the binaries available.
Actually, the source code *is* now available for ftp, as Jonathan sent it
to me as soon as it became clear that a controversy was erupting. It has been
on nic.switch.ch as software/mac/src/GCC_1.37_demo_src.cpt.bin since Friday,
but then I was too exhausted to remember informing Jonathan or the net of that
fact.
Our initial motivation for not immediately putting up the sources was that
the only means of data transfer between Jonathan and me is email and sending
Multi-Megabyte packages per email once a month is relatively unpleasant.
Therefore, I agreed to delay the source transfer for some time until a first
wave of bug reports would have been addressed. Part of my mistake was
underestimating the interest that a relatively old compiler would generate
(I certainly would have insited on immediate source availability for a
GCC 2.x port). I apologize for my mistakes in this issue.
>Mr. Kimmitt's offer falls into category (b), above. Mr. Suker's
>objection is reasonable, in my opinion. The GPL does not dictate
>what constitutes a "medium customarily used for software interchange",
>because this changes rapidly with changing technology. However,
>I think that it is clear that Bernoulli cartridges do NOT meet
>this criterion.
Given that the FSF used media such as tapes, and that all concerned consider
ftp, which in large parts of the world is much more exotic than Bernoullis,
a perfectly acceptable medium, I have to disagree.
> [ The situation might be different if Mr. Kimmitt
>had chosen to make his binary distribution only on Bernoulli cart,
>which would guarantee that any recipient of the binaries would have
>reasonable access to source distributed on the same media, but that
>is not the case here. ]
You have a point here, and this is probably the best criterium for what
makes reasonable distribution.
>When a fully GPL-compliant distribution of Mr. Kimmitt's gcc
>is finally made available, everyone will benefit.
Definitely. Over all these distribution controversies, it should not be
forgotten that Jonathan's GCC is to my knowledge the closest that anybody
has ever come to a standalone version of GCC for the Mac so far.
Matthias
-----
Matthias Neeracher ne...@iis.ee.ethz.ch
"One fine day in my odd past..." -- Pixies, _Planet of Sound_
Matthias Neeracher
>In article <CF0E7...@news.udel.edu> john...@me.udel.edu (Bill Johnston) writes:
>>Unfortunately, a closer examination of the gcc binary that
>>was included in Mr. Kimmitt's distribution reveals other,
>>more serious GPL violations. The most obvious of these is
>>the integrated text editor, which includes a copyright and
>>license statement that clearly violates the GPL in several ways.
>As the author of said text editor, I find it rather ironic that you
>would take objection to its inclusion in Mr. Kimmitt's distribution of
>GCC. While you are correct in noting that the terms of the license
>agreement for my text editor are not compliant with those of the GNU
>General Public License, in my opinion there are far more serious
>transgressions in Mr. Kimmitt's distribution of GCC. Most notably,
>his shell application is based on and requires Symantec's THINK Class
>Library, a collection of copyrighted commercial source code which,
>unlike my text editor, cannot be freely distributed. Without the TCL
>source code, it is virtually impossible to compile Mr. Kimmitt's
>version of GCC, which to me violates the underlying spirit and
>intention of the GNU GPL.
The GPL says:
& For an executable file, complete source code means
& all the source code for all modules it contains; but, as a special
& exception, it need not include source code for modules which are standard
& libraries that accompany the operating system on which the executable
& file runs, or for standard header files or definitions files that
& accompany that operating system.
And it would appear to me that since the Think Class Library is shipped with
every copy of Think C, it might just barely fall under this exception. There
are no standard libraries or headers as such on the Mac, and if the TCL
does not qualify, all MPW C GNU ports which call SpinCursor() (And I have
done a few of these in my time :-) and probably also most DOS, Atari, etc.
GNU ports would be illegal.
The TCL is more problematic as it contains more code than a typical C library,
but, after all, the spirit of the GPL is to further the distribution of free
software without erecting artificial barriers, and Jonathans use of the TCL
certainly doesn't qualify as that.
Matthias
-----
Matthias Neeracher ne...@iis.ethz.ch
"The fundamental act of friendship among programmers is the sharing
of programs." -- Richard Stallman, _The Gnu Manifesto_
Jon Wätte
>>what constitutes a "medium customarily used for software interchange",
>>because this changes rapidly with changing technology. However,
>>I think that it is clear that Bernoulli cartridges do NOT meet
>Given that the FSF used media such as tapes, and that all concerned consider
>ftp, which in large parts of the world is much more exotic than Bernoullis,
>a perfectly acceptable medium, I have to disagree.
Just settle for punch cards, will you? It's definately the most
mentioned input/output media in all written computer litterature
up 'til today.
You can buy GCC for the Mac for $500! You can also get the
FULL SOURCE on industry-standard Punch Cards for only $110,000!
Yes, that's right, we GIVE away the source for only the cost of
the paper! :-)
--
-- Jon W{tte, h...@nada.kth.se, Mac Hacker Deluxe --
This article printed on 100% recycled electrons.
Chris Wysocki
>And it would appear to me that since the Think Class Library is shipped with
>every copy of Think C, it might just barely fall under this exception. There
>are no standard libraries or headers as such on the Mac, and if the TCL
>does not qualify, all MPW C GNU ports which call SpinCursor() (And I have
>done a few of these in my time :-) and probably also most DOS, Atari, etc.
>GNU ports would be illegal.
Every C compiler that I've used on the Macintosh has included, in
addition to the standard K&R/ANSI headers and libraries, a licensed
set of Apple's headers and libraries for the Macintosh Toolbox, so I
think that these would qualify as the GPL's "standard libraries" and
"header files...that accompany the operating system." Beyond that, I
feel any other header files or libraries (such as the TCL or MPW's
CursorCtl library) are problematic, unless the port is done in a
manner independent of the availability of these libraries (e.g. "#if
applec/SpinCursor(1)/#endif"). On the other hand, if the TCL is
considered a standard library with regard to the GNU GPL, then I
wouldn't think that the inclusion of my freely-distributable text
editor with Mr. Kimmitt's distribution of GCC would be a problem.
Then again, one could argue that Mr. Kimmitt's TCL-based shell
environment is separate and distinct from the various GNU components
(compiler/assembler/linker) and hence is not subject to the terms of
the GPL. After all, the Apple ATG port of GCC requires the MPW Shell
to run, and clearly the MPW Shell does not comply with the GPL. I
think this is an issue which the FSF needs to address more directly.
>The TCL is more problematic as it contains more code than a typical C library,
>but, after all, the spirit of the GPL is to further the distribution of free
>software without erecting artificial barriers, and Jonathans use of the TCL
>certainly doesn't qualify as that.
It's a matter of degree. If I'm a user of MPW C, the use of the TCL
in Mr. Kimmitt's port is far more of an artificial barrier than the
use of SpinCursor is to a user of THINK C. In general, I feel it's a
bad idea to make code covered by the GPL dependent on a particular
commercial compiler, since it unnecessarily restricts one's ability to
modify the code.
Don't get me wrong; I applaud Mr. Kimmitt's efforts to make a
standalone version of GCC available on the Macintosh, for it is
something that I have wanted to see for a long time. It is simply my
opinion that he should have written the entire shell from scratch, so
as to avoid making his implementation dependent upon a commercial
source code library that is not freely distributable.
Joseph Hall
>Don't get me wrong; I applaud Mr. Kimmitt's efforts to make a
>standalone version of GCC available on the Macintosh, for it is
>something that I have wanted to see for a long time. It is simply my
Should we start an RFD for comp.sys.mac.programmer.pedantic?
Maybe it's just my Sunday thinking cap, but this thread seems a
little harsh.
--
Joseph Nathan Hall | "Hi, there, little green BSD. Hi Donna. Login. Login.
Software Architect | Going to bite me today? Hey! Ooo, you are a nice
Gorca Systems Inc. | vicious little birdy, aren't you? Like more veggies?"
(on assignment) | (602) 732-2549 (work) Josep...@sat.mot.com
Bill Johnston
>>the integrated text editor, which includes a copyright and
>>license statement that clearly violates the GPL in several ways.
>As the author of said text editor, I find it rather ironic that you
>would take objection to its inclusion in Mr. Kimmitt's distribution of
>GCC.
Two things: first, in case it wasn't clear enough before, I think
that the work done by both Mr. Kimmitt and Mr. Wysocki is terrific.
The GPL violations were not deliberate or malicious, and the effort
required to bring the distribution into compliance should be trivial
compared to the effort made in getting this far--well worth making.
I also want to emphasize that I wasn't trying to single out the
editor. It just happens to be easy to explain the nature of the
problem in this case, because the violation is readily apparent
in the copyright and license statement.
> While you are correct in noting that the terms of the license
>agreement for my text editor are not compliant with those of the GNU
>General Public License, in my opinion there are far more serious
>transgressions in Mr. Kimmitt's distribution of GCC. Most notably,
>his shell application is based on and requires Symantec's THINK Class
>Library, a collection of copyrighted commercial source code which,
>unlike my text editor, cannot be freely distributed. Without the TCL
>source code, it is virtually impossible to compile Mr. Kimmitt's
>version of GCC, which to me violates the underlying spirit and
>intention of the GNU GPL.
Fortunately this is not the case. The GPL makes explicit allowance
for linking to proprietary C libraries, such as those which ship
with various proprietary unix implementations, or those that are
necessary parts of various proprietary compiler systems like
Symantec's Think C libraries, MacTraps, TCL, etc. Otherwise
porting GNU software to SunOS or DOS with Microsoft's compiler
would be impossible.
Bill Johnston
>john...@me.udel.edu (Bill Johnston) writes:
>> The GPL does not dictate
>>what constitutes a "medium customarily used for software interchange",
>>because this changes rapidly with changing technology. However,
>>I think that it is clear that Bernoulli cartridges do NOT meet
>>this criterion.
>
>Given that the FSF used media such as tapes, and that all concerned consider
>ftp, which in large parts of the world is much more exotic than Bernoullis,
>a perfectly acceptable medium, I have to disagree.
The key point here is that FSF does not distribute binaries
without source. In fact, I can't recall them ever distributing
binaries, period.
Source-only distributions can be done by stone tablet, if the
author so chooses. The obligation to make available machine-
readable source on 'customary' interchange media arises only
when a binary is distributed without full source.
Peter Lewis
>Maybe it's just my Sunday thinking cap, but this thread seems a
>little harsh.
I agree, if I was the fellow, I'd never bother making any future stuff
available (well, ok, that might be an excessive over reaction). Really,
as long as he makes all the code that he wrote and all the modifications
to gcc available under the GPL (ie, if I get a copy off him, I can give
the copy to someone else), what more can you ask? If it uses TCL and he
can't distribute that, thats not his fault, at least you've got all the
code he's written, and you can compile it yourself with easily available
files from THINK C. People should also remember that the "free" in
"free software" means that you are free to do what you like with it, not
necessarily that you can compile it for free (you generally can't
compile any software for free, since you need to buy a compiler first
anyway). The free really means that you have access to the source code,
so you can make and share your changes.
I'm quite amazed at the amount of flack this guy's taken for making his
code available, and taking the time to make a standalone gcc available.
If you people don't like its dependence on TCL and the editory, then get
the code, and write gTCL and gEditor and stick them in. I'm sure a GNU
class library would be exceptionally useful, and if a portable one could
be made (at least for msdos and mac initially) then we could all forget
about bedrock (those who havent already).
Peter.
--
_______________________________________________________________________
Peter N Lewis <peter...@info.curtin.edu.au> Ph: +61 9 368 2055
Chris Wysocki
>hal...@sat.mot.com (Joseph Hall) writes:
>
>>Maybe it's just my Sunday thinking cap, but this thread seems a
>>little harsh.
>
>I agree, if I was the fellow, I'd never bother making any future stuff
>available (well, ok, that might be an excessive over reaction). Really,
>as long as he makes all the code that he wrote and all the modifications
>to gcc available under the GPL (ie, if I get a copy off him, I can give
>the copy to someone else), what more can you ask?
Sorry if I came off as a bit heavy-handed in my previous messages. At
the time I simply thought it strange that Bill Johnston was
questioning the inclusion of a copyrighted-but-free piece of source
code (which I happen to have written) when the entire distribution was
dependent upon a commercial source code library that is not freely
distributable. Since then Bill has clarified that the dependency on
the TCL is not a problem, which presumably means that the inclusion of
my text editor is not either. As I mentioned in my previous message,
I think it's great that Jonathan Kimmitt has ported GCC to the Mac as
a standalone application. It's something that's long overdue and I'm
sure that many people, including myself, will appreciate his efforts.