Hi

In this forum, there was a thread that mentioned that GNU code is being
used by lcc-win32.

This should be a violation of the GNU license. We did not know this,
and thought that just distributing the code would
be enough.

Therefore:

1: gdbm distribution is dropped from the lcc-win32 distribution
2: The function "edit_distance" is dropped from the library
3: The linux version of lcc-win32 is discontinued. We are trying to
evaluate how much effort will be to replace the bfd library
and use our own object file writing... Until then, no linux
version will be available.
4: The parts of lcc-win32 that called the gnu program "indent"
(that itself is derived from earlier code) have been deleted.
indent will no longer be distributed.


--
jacob navia
jacob at jacob point remcomp point fr
logiciels/informatique
http://www.cs.virginia.edu/~lcc-win32

jacob navia said:

> Hi
>
> In this forum, there was a thread that mentioned that GNU code is being
> used by lcc-win32.

Before the anonytrolls get started, can I just point out that Mr Navia's
response to discovering that his product violates the GPL licence seems to
me to be a perfectly reasonable one. Everyone makes mistakes - what counts
is not that we make them, but how we react when we discover them.

The world (or at least this bit of it) knows that I am far from being Mr
Navia's #1 fan, so I hope that his (many!) other critics will think
carefully about his response, rather than just lay into him blindly, which
would be a most foolish reaction.

<snip>

--
Richard Heathfield <http://www.cpax.org.uk>
Email: -http://www. +rjh@
Google users: <http://www.cpax.org.uk/prg/writings/googly.php>
"Usenet is a strange place" - dmr 29 July 1999

On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
> Everyone makes mistakes - what counts
> is not that we make them, but how we react when we discover them.
Well, what options where available at the time?
Either change the project, or GPL it, or infringe GPL.

jacob navia:
Why not GPL your project? or at least make the source available?
It would save you from trouble, and it would certainly save heaps
of your time, that is if other people find it interesting and decide
to contribute.

vipvipvipvip.ru@gmail.com wrote:
> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>> Everyone makes mistakes - what counts
>> is not that we make them, but how we react when we discover them.
> Well, what options where available at the time?
> Either change the project, or GPL it, or infringe GPL.
>
> jacob navia:
> Why not GPL your project? or at least make the source available?
> It would save you from trouble, and it would certainly save heaps
> of your time, that is if other people find it interesting and decide
> to contribute.

If Jacob doesn't want to share his source he is entirely within his
rights to do so. If you want to try and persuade him, please do so in
comp.compilers.lcc.

Phil

--
Philip Potter pgp <at> doc.ic.ac.uk

vipvipvipvip.ru@gmail.com said:

> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>> Everyone makes mistakes - what counts
>> is not that we make them, but how we react when we discover them.
> Well, what options where available at the time?

If, at the time he made the mistake, he was aware that he was making the
mistake, presumably he wouldn't have made the mistake. Your followup is
absurd.

--
Richard Heathfield <http://www.cpax.org.uk>
Email: -http://www. +rjh@
Google users: <http://www.cpax.org.uk/prg/writings/googly.php>
"Usenet is a strange place" - dmr 29 July 1999

Richard Heathfield wrote:
> vipvipvipvip.ru@gmail.com said:
>
>> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>>> Everyone makes mistakes - what counts
>>> is not that we make them, but how we react when we discover them.
>> Well, what options where available at the time?
>
> If, at the time he made the mistake, he was aware that he was making the
> mistake, presumably he wouldn't have made the mistake. Your followup is
> absurd.
>

Well... yes

I thought that redistributing the source would cover my compliance with
the GPL. But the GPL people say that if I distribute (with source)
the GPL code, the lcc-win compiler must be GPLed too.

I am not a lawyer and prefer to avoid legal battles that I
can't afford anyway.

The same for the linux distribution that has costed me a lot
of work. I thought that just using a GPLed binaries would not
mean that my code is forced to change its license. Apparently
that is the case. I have to write now a module to write
elf format object files. This will be a lot of work.

I can't put my code under the GPL because I live from my
work. In a perfect world, I would give my work for free
and I would go to the supermarket and get food for free,
my home would have been given to me for free, etc. My kids
would get everything they need for free, etc.

This is apparently not the case. My supermarket is not GNU
and I have to pay for their goods.

--
jacob navia
jacob at jacob point remcomp point fr
logiciels/informatique
http://www.cs.virginia.edu/~lcc-win32

jacob navia <jacob@nospam.org> writes:

> Richard Heathfield wrote:
>> vipvipvipvip.ru@gmail.com said:
>>
>>> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>>>> Everyone makes mistakes - what counts
>>>> is not that we make them, but how we react when we discover them.
>>> Well, what options where available at the time?
>>
>> If, at the time he made the mistake, he was aware that he was making
>> the mistake, presumably he wouldn't have made the mistake. Your
>> followup is absurd.
>>
>
> Well... yes
>
> I thought that redistributing the source would cover my compliance with
> the GPL. But the GPL people say that if I distribute (with source)
> the GPL code, the lcc-win compiler must be GPLed too.
>
> I am not a lawyer and prefer to avoid legal battles that I
> can't afford anyway.
>
> The same for the linux distribution that has costed me a lot
> of work. I thought that just using a GPLed binaries would not
> mean that my code is forced to change its license. Apparently
> that is the case. I have to write now a module to write
> elf format object files. This will be a lot of work.
>
> I can't put my code under the GPL because I live from my
> work. In a perfect world, I would give my work for free
> and I would go to the supermarket and get food for free,
> my home would have been given to me for free, etc. My kids
> would get everything they need for free, etc.

Theoretically you can still earn money from the GPLd code as people are
"honour bound" to pay a license to use it if you so state.

Unfortunately in the real world it doesn't happen that often unless the
user really needs support.

> This is apparently not the case. My supermarket is not GNU
> and I have to pay for their goods.

In article <1192709803.337855.324270@q3g2000prf.googlegroups. com>,
<vipvipvipvip.ru@gmail.com> wrote:
>On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>> Everyone makes mistakes - what counts
>> is not that we make them, but how we react when we discover them.
>Well, what options where available at the time?
>Either change the project, or GPL it, or infringe GPL.
>
>jacob navia:
>Why not GPL your project? or at least make the source available?
>It would save you from trouble, and it would certainly save heaps
>of your time, that is if other people find it interesting and decide
>to contribute.

I, along with a lot of other people, think that Microsoft should GPL
their code. Then hackers around the world could post line-by-line
critiques of their code (which surely deserves it far more than any
newbie posting to CLC) and we'd have real fun.

In article <47175e38$0$25914$ba4acef3@news.orange.fr>,
jacob navia <jacob@nospam.org> wrote:
>Richard Heathfield wrote:
>> vipvipvipvip.ru@gmail.com said:
>>
>>> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>>>> Everyone makes mistakes - what counts
>>>> is not that we make them, but how we react when we discover them.
>>> Well, what options where available at the time?
>>
>> If, at the time he made the mistake, he was aware that he was making the
>> mistake, presumably he wouldn't have made the mistake. Your followup is
>> absurd.
>>
>
>Well... yes
>
>I thought that redistributing the source would cover my compliance with
>the GPL. But the GPL people say that if I distribute (with source)
>the GPL code, the lcc-win compiler must be GPLed too.

Isn't the usual work-around to this problem to not distribute the stuff
you can't distribute, but to provide pointers so that the user can d/l
it themselves? This:
a) Gets you off the hook - and you can continue to distribute
your source however you want to.
and b) Puts whatever legal onus there might be onto the user, who
presumably will just do whatever he wants anyway.

jacob navia <jacob@nospam.org> writes:

> Richard Heathfield wrote:
>> vipvipvipvip.ru@gmail.com said:
>>
>>> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>>>> Everyone makes mistakes - what counts
>>>> is not that we make them, but how we react when we discover them.
>>> Well, what options where available at the time?
>>
>> If, at the time he made the mistake, he was aware that he was making
>> the mistake, presumably he wouldn't have made the mistake. Your
>> followup is absurd.
>>
>
> Well... yes
<snip>
> The same for the linux distribution that has costed me a lot
> of work. I thought that just using a GPLed binaries would not
> mean that my code is forced to change its license. Apparently
> that is the case. I have to write now a module to write
> elf format object files.

The key distinction here is between the GPL and LGPL code. Linking
against a library creates a derived work and if the library is GPL,
the derived work must also comply. The LGLP is specifically designed
to remove this condition (but there are strong arguments to prefer the
GPL[1]).

In the case of libbfd, the mistake is an easy one to make since I
believe some early versions were indeed LGPL.

This is in no way intended as a criticism -- I just wanted to clarify
a frequently misunderstood point.

[1] http://www.gnu.org/licenses/why-not-lgpl.html

--
Ben.

Kenny McCormack wrote:
> In article <47175e38$0$25914$ba4acef3@news.orange.fr>,
> jacob navia <jacob@nospam.org> wrote:
>> Richard Heathfield wrote:
>>> vipvipvipvip.ru@gmail.com said:
>>>
>>>> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>>>>> Everyone makes mistakes - what counts
>>>>> is not that we make them, but how we react when we discover them.
>>>> Well, what options where available at the time?
>>> If, at the time he made the mistake, he was aware that he was making the
>>> mistake, presumably he wouldn't have made the mistake. Your followup is
>>> absurd.
>>>
>> Well... yes
>>
>> I thought that redistributing the source would cover my compliance with
>> the GPL. But the GPL people say that if I distribute (with source)
>> the GPL code, the lcc-win compiler must be GPLed too.
>
> Isn't the usual work-around to this problem to not distribute the stuff
> you can't distribute, but to provide pointers so that the user can d/l
> it themselves? This:
> a) Gets you off the hook - and you can continue to distribute
> your source however you want to.
> and b) Puts whatever legal onus there might be onto the user, who
> presumably will just do whatever he wants anyway.
>

In the case of bfd this is not possible since the binary of lcc-win
will not start without that "bfd" library being present.

--
jacob navia
jacob at jacob point remcomp point fr
logiciels/informatique
http://www.cs.virginia.edu/~lcc-win32

In article <47173997$0$27390$ba4acef3@news.orange.fr>,
jacob navia <jacob@nospam.org> wrote:
>Hi
>
>3: The linux version of lcc-win32 is discontinued. We are trying to
> evaluate how much effort will be to replace the bfd library
> and use our own object file writing... Until then, no linux
> version will be available.

But compilers don't write object files. They write text files containing
assembly code. Assemblers and linkers write object files...

--
Alan Curry
pacman@world.std.com

In article <ff87nv$n9a$1@pcls6.std.com>,
Alan Curry <pacman@TheWorld.com> wrote:
>In article <47173997$0$27390$ba4acef3@news.orange.fr>,
>jacob navia <jacob@nospam.org> wrote:

>>3: The linux version of lcc-win32 is discontinued. We are trying to
>> evaluate how much effort will be to replace the bfd library
>> and use our own object file writing... Until then, no linux
>> version will be available.

>But compilers don't write object files. They write text files containing
>assembly code. Assemblers and linkers write object files...

There is no difference in principle between A) writing text that
an assembler parses into binary, or B) writing in a binary
"assembly language" (aka machine language) and having a "null"
assembler that leaves everything unchanged. Going through
a text assembly step can increase portability, but direct
generation can increase speed.
--
"History is a pile of debris" -- Laurie Anderson

Walter Roberson wrote:
> Going through
> a text assembly step can increase portability, but direct
> generation can increase speed.

Exactly. And it is QUITE an increase of speed; since
the assembler files tend to be HUGE and provoke
a lot of I/O to disk. Lcc-win does NOT
touch the disk from reading the original C source to
writing the object file.

No intermediate preprocessed
file, and no intermediate assembler file. This allows it to
have a 10-15 fold compilation speed compared to other
compilers: The problem is now to write the elf writing
routines, the relocations, and similar things.

I hope that calling the linker using

system("gcc output.o") will NOT make my compiler
break the GPL...


--
jacob navia
jacob at jacob point remcomp point fr
logiciels/informatique
http://www.cs.virginia.edu/~lcc-win32

In article <47177987$0$25952$ba4acef3@news.orange.fr>,
jacob navia <jacob@nospam.org> wrote:
>Kenny McCormack wrote:
>> In article <47175e38$0$25914$ba4acef3@news.orange.fr>,
>> jacob navia <jacob@nospam.org> wrote:
>>> Richard Heathfield wrote:
>>>> vipvipvipvip.ru@gmail.com said:
>>>>
>>>>> On Oct 18, 2:27 pm, Richard Heathfield <r...@see.sig.invalid> wrote:
>>>>>> Everyone makes mistakes - what counts
>>>>>> is not that we make them, but how we react when we discover them.
>>>>> Well, what options where available at the time?
>>>> If, at the time he made the mistake, he was aware that he was making the
>>>> mistake, presumably he wouldn't have made the mistake. Your followup is
>>>> absurd.
>>>>
>>> Well... yes
>>>
>>> I thought that redistributing the source would cover my compliance with
>>> the GPL. But the GPL people say that if I distribute (with source)
>>> the GPL code, the lcc-win compiler must be GPLed too.
>>
>> Isn't the usual work-around to this problem to not distribute the stuff
>> you can't distribute, but to provide pointers so that the user can d/l
>> it themselves? This:
>> a) Gets you off the hook - and you can continue to distribute
>> your source however you want to.
>> and b) Puts whatever legal onus there might be onto the user, who
>> presumably will just do whatever he wants anyway.
>>
>
>In the case of bfd this is not possible since the binary of lcc-win
>will not start without that "bfd" library being present.

So? You just tell them they have to download/install bfd before they
can run your compiler.

On Oct 18, 11:45 pm, jacob navia <ja...@nospam.org> wrote:
>
> This should be a violation of the GNU license. Therefore:
>
> 1: gdbm distribution is dropped from the lcc-win32 distribution
> 2: The function "edit_distance" is dropped from the library
> 3: The linux version of lcc-win32 is discontinued. We are trying to
> evaluate how much effort will be to replace the bfd library
> and use our own object file writing... Until then, no linux
> version will be available.
> 4: The parts of lcc-win32 that called the gnu program "indent"
> (that itself is derived from earlier code) have been deleted.
> indent will no longer be distributed.

Good to see that Linux has DRM too!

jacob navia <jacob@nospam.org> writes:
[...]
> I hope that calling the linker using
>
> system("gcc output.o") will NOT make my compiler
> break the GPL...

Since this thread is already marked "[OT]" ...

Why do you invoke the linker via gcc? Why not just
system("ld output.o")
(which is what gcc is going to do anyway)?

I'm not sure that the behavior of "gcc output.o" is exactly the same
as "ld output.o", but it's easy enough to figure out just what gcc is
going to do and reproduce the logic. It's possible that a Linux
system might have ld installed, but not gcc. (On Linux, both gcc and
ld are typically GNU programs, so this has no bearing on your question
about GPL. I *think* that invoking either gcc or ld is ok, but my
opinion on this matter is worth no more than what you're paying for
it.)

--
Keith Thompson (The_Other_Keith) kst-u@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

jacob navia <jacob@nospam.org> writes:
> In this forum, there was a thread that mentioned that GNU code is
> being used by lcc-win32.
>
> This should be a violation of the GNU license. We did not know this,
> and thought that just distributing the code would
> be enough.
>
> Therefore:
>
> 1: gdbm distribution is dropped from the lcc-win32 distribution
> 2: The function "edit_distance" is dropped from the library
> 3: The linux version of lcc-win32 is discontinued. We are trying to
> evaluate how much effort will be to replace the bfd library
> and use our own object file writing... Until then, no linux
> version will be available.
> 4: The parts of lcc-win32 that called the gnu program "indent"
> (that itself is derived from earlier code) have been deleted.
> indent will no longer be distributed.

I applaud your quick and sensible response.

--
Keith Thompson (The_Other_Keith) kst-u@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"

Richard Heathfield wrote:
> jacob navia said:
>
> > Hi
> >
> > In this forum, there was a thread that mentioned that GNU code is being
> > used by lcc-win32.
>
> Before the anonytrolls get started, can I just point out that Mr Navia's
> response to discovering that his product violates the GPL licence seems to
> me to be a perfectly reasonable one. Everyone makes mistakes - what counts
> is not that we make them, but how we react when we discover them.
>
> The world (or at least this bit of it) knows that I am far from being Mr
> Navia's #1 fan, so I hope that his (many!) other critics will think
> carefully about his response, rather than just lay into him blindly, which
> would be a most foolish reaction.

Depressing to find you've become such a craven apologist for Navia's
selfishness.

The situation is this: Navia has invested a bit of time in making a
Linux version of his notoriously poor compiler. Under the terms of the
GPL, he has two choices: he can distribute the resulting program under
the GPL, or send it to /dev/null.

There is *no way* for Navia to make money now, but he'd rather write
off the time he spent porting to Linux rather than give people access
to his compiler with the same spirit of freedom and generosity with
which he benefited from the GPL'd components he wanted to include in
it. Taking under the GPL is fine for him - but a different story when
it's his turn to give.

Mind you, I don't think many Linux users will be crying /too/ hard if
they don't get the chance to use Navia's bad compiler for non-standard
C.

rosewater@mailinator.com wrote:
> Richard Heathfield wrote:
>> jacob navia said:
>>
>> > Hi
>> >
>> > In this forum, there was a thread that mentioned that GNU code is
>> > being used by lcc-win32.
>>
>> Before the anonytrolls get started, can I just point out that Mr
>> Navia's response to discovering that his product violates the GPL
>> licence seems to me to be a perfectly reasonable one.

<snip>

> The situation is this: Navia has invested a bit of time in making a
> Linux version of his notoriously poor compiler. Under the terms of the
> GPL, he has two choices: he can distribute the resulting program under
> the GPL, or send it to /dev/null.

Or use a GPL compliant license.

> There is *no way* for Navia to make money now,

The GPL doesn't prohibit selling software covered by it.

<snip>

rosewater@mailinator.com said:

>
> Richard Heathfield wrote:

<snip>

>> The world (or at least this bit of it) knows that I am far from being Mr
>> Navia's #1 fan, so I hope that his (many!) other critics will think
>> carefully about his response, rather than just lay into him blindly,
>> which would be a most foolish reaction.
>
> Depressing to find you've become such a craven apologist for Navia's
> selfishness.

Depressing to find that you can't read. I haven't made a case either for or
against what you call his "selfishness". All I have said is that we should
not criticise someone simply for taking appropriate corrective action
after realising that a mistake has been made. Okay, so you'd rather see
him release the source. Well, it's clear that he isn't going to. That is
his decision, not yours, mine, or anyone else's.

> The situation is this: Navia has invested a bit of time in making a
> Linux version of his notoriously poor compiler. Under the terms of the
> GPL, he has two choices: he can distribute the resulting program under
> the GPL, or send it to /dev/null.

Indeed. And he has chosen the latter. That is *his* choice to make, not
yours or mine.

> There is *no way* for Navia to make money now, but he'd rather write
> off the time he spent porting to Linux rather than give people access
> to his compiler with the same spirit of freedom and generosity with
> which he benefited from the GPL'd components he wanted to include in
> it.

Perhaps. But I think it more likely that he doesn't reveal his source code
because he knows full well that it will not withstand expert scrutiny.

--
Richard Heathfield <http://www.cpax.org.uk>
Email: -http://www. +rjh@
Google users: <http://www.cpax.org.uk/prg/writings/googly.php>
"Usenet is a strange place" - dmr 29 July 1999

In article <1193004036.074981.186250@e34g2000pro.googlegroups .com>,
<rosewater@mailinator.com> walked into a trap and wrote:
>
>Richard Heathfield wrote:
>> jacob navia said:
>>
>> > Hi
>> >
>> > In this forum, there was a thread that mentioned that GNU code is being
>> > used by lcc-win32.
>>
>> Before the anonytrolls get started, can I just point out that Mr Navia's
>> response to discovering that his product violates the GPL licence seems to
>> me to be a perfectly reasonable one. Everyone makes mistakes - what counts
>> is not that we make them, but how we react when we discover them.
>>
>> The world (or at least this bit of it) knows that I am far from being Mr
>> Navia's #1 fan, so I hope that his (many!) other critics will think
>> carefully about his response, rather than just lay into him blindly, which
>> would be a most foolish reaction.
>
>Depressing to find you've become such a craven apologist for Navia's
>selfishness.

hah hah hah heee heee hoe hoe hah hah hee hah hah hah


OK, OK, I've recovered (oops, I'm still ROTFL...)

I believe the operative phrase here is: YHBT.

(Good show, Heathfield, BTW)

rosewater@mailinator.com writes:

> Richard Heathfield wrote:
>> jacob navia said:
>>
>> > Hi
>> >
>> > In this forum, there was a thread that mentioned that GNU code is being
>> > used by lcc-win32.
>>
>> Before the anonytrolls get started, can I just point out that Mr Navia's
>> response to discovering that his product violates the GPL licence seems to
>> me to be a perfectly reasonable one. Everyone makes mistakes - what counts
>> is not that we make them, but how we react when we discover them.
>>
>> The world (or at least this bit of it) knows that I am far from being Mr
>> Navia's #1 fan, so I hope that his (many!) other critics will think
>> carefully about his response, rather than just lay into him blindly, which
>> would be a most foolish reaction.
>
> Depressing to find you've become such a craven apologist for Navia's
> selfishness.
>
> The situation is this: Navia has invested a bit of time in making a
> Linux version of his notoriously poor compiler. Under the terms of the
> GPL, he has two choices: he can distribute the resulting program under
> the GPL, or send it to /dev/null.
>
> There is *no way* for Navia to make money now, but he'd rather write
> off the time he spent porting to Linux rather than give people access
> to his compiler with the same spirit of freedom and generosity with
> which he benefited from the GPL'd components he wanted to include in
> it. Taking under the GPL is fine for him - but a different story when
> it's his turn to give.
>
> Mind you, I don't think many Linux users will be crying /too/ hard if
> they don't get the chance to use Navia's bad compiler for non-standard
> C.

Here's an idea : why don't you publish everything YOU contribute to the
OS community or simply fuck off?

In article <ff87nv$n9a$1@pcls6.std.com> pacman@TheWorld.com (Alan Curry) writes:
....
> But compilers don't write object files. They write text files containing
> assembly code. Assemblers and linkers write object files...

Wrong. The first Pascal compiler I ever did use (the original compiler
from ETH Zuerich for CDC systems) did write out the binary directly, in
a single pass. And I have known many compilers that *never* did use an
assembler in whatever form.
--
dik t. winter, cwi, kruislaan 413, 1098 sj amsterdam, nederland, +31205924131
home: bovenover 215, 1025 jn amsterdam, nederland; http://www.cwi.nl/~dik/

In article <JqAKs7.876@cwi.nl>, Dik T. Winter <Dik.Winter@cwi.nl> wrote:
>In article <ff87nv$n9a$1@pcls6.std.com> pacman@TheWorld.com (Alan Curry) writes:
>...
> > But compilers don't write object files. They write text files containing
> > assembly code. Assemblers and linkers write object files...
>
>Wrong. The first Pascal compiler I ever did use (the original compiler
>from ETH Zuerich for CDC systems) did write out the binary directly, in
>a single pass. And I have known many compilers that *never* did use an
>assembler in whatever form.

Would you feel better if Alan has explicitly used the implied word "most"
in his statement?