On 10/08/2023 12:57, Bart wrote:
> On 10/08/2023 12:29, Paul Edwards wrote:
> > Hi Bart.
> >
> > I'm not sure what your goals are, but at one point you were
> > interested in the executable loader code in PDOS, to the
> > point that you even compiled it with a 64-bit C compiler
> > and pointed out issues with it.
> >
> > I corrected those issues, but I don't think you used it again.
> >
> > And at one point you released a generated 64-bit compiler
> > suitable for a Win64 environment with msvcrt.dll.
> >
> > Note that at the time I was only doing 32-bit coding, so I
> > didn't test the fixes that I made, nor was I able to use
> > the C compiler for anything I was trying to do.
> >
> > As of a couple of hours ago, that has changed.
> >
> > There is a version of PDOS - basically I convert the UEFI
> > OS into a basic Win64 OS - there wasn't a lot of "glue"
> > required to do that.
> >
> > Since we did the original activity above, there have been
> > new developments such as a public domain 64-bit coff
> > linker.
> >
> > So now there is a purely public domain distribution
> > called "University Challenge x64" available at http://pdos.org
> >
> > The only thing missing in the toolchain is a viable C compiler.
> >
> > SubC is included, but it is only a subset of C90.
> >
> > Can you tell me what the current status is of your public
> > domain C compiler and whether you are interested in
> > getting it to run on this new Win64 environment?
>
> I think it's best to forget my compiler:
>
> * I've lost interest in C and in implementing it. (It seems mainly
>   a tool for raising my blood pressure.)
>
> * The compiler is non conforming in many ways, and buggy
>
> * It's not written in C, and the tool used to create a C rendering
>   is being deprecated
>
> * Its output is anyway not usually an object file; it is ASM in
>   a special format, directly assembled into EXE files
>
> * There was a .OBJ format that could be produced, but that is no
>   longer viable, as my compiler's code is intended for low memory
>   (within first 2GB) at a fixed address. But modern OSes and linkers
>   now like to load at arbitrary, high addresses, which only works with
>   position-independent code.
>
> I may anyway, at any time, decide to just delete the lot. I just don't
> want any external pressures or responsibilities.

It's pretty much gone now together with other tools associated with C
(converting to C or from C). It only exists now as an archived ZIP on a
pen drive. I'll give it a week to see what to do about that.

It's a relief!

In article <ub2kbv$c6ds$1@dont-email.me>, Bart <bc@freeuk.com> wrote:
>On 10/08/2023 13:07, Paul Edwards wrote:
>> On Thursday, August 10, 2023 at 7:57:52 PM UTC+8, Bart wrote:
>>
>>> * There was a .OBJ format that could be produced, but that is no
>>> longer viable, as my compiler's code is intended for low memory
>>> (within first 2GB) at a fixed address. But modern OSes and linkers
>>> now like to load at arbitrary, high addresses, which only works with
>>> position-independent code.
>>
>> We aren't using what you would presumably call "modern
>> OSes and linkers". Even though they are only hours/days
>> old, loading below 2 GB is not an issue - we control all
>> components except EFI, but you can request EFI to
>> provide memory below a certain point, which is all that is
>> needed.
>>
>>> I may anyway, at any time, decide to just delete the lot. I just don't
>>> want any external pressures or responsibilities.
>>
>> Ok, sorry to hear your interests have changed.
>>
>>> However there must be 100s of small, amateur C compilers out there, you
>>> just have to look for them. Tiny C is open source and one of the best.
>>
>> Almost all of them are copyrighted.
>
>In a post-apocalyptic scenario (I think that's partly what you want this
>stuff for), I don't think many will care about copyrights!

PLEASE don't feed this troll.

- Dan C.

Hi Bart.

I'm not sure what your goals are, but at one point you were
interested in the executable loader code in PDOS, to the
point that you even compiled it with a 64-bit C compiler
and pointed out issues with it.

I corrected those issues, but I don't think you used it again.

And at one point you released a generated 64-bit compiler
suitable for a Win64 environment with msvcrt.dll.

Note that at the time I was only doing 32-bit coding, so I
didn't test the fixes that I made, nor was I able to use
the C compiler for anything I was trying to do.

As of a couple of hours ago, that has changed.

There is a version of PDOS - basically I convert the UEFI
OS into a basic Win64 OS - there wasn't a lot of "glue"
required to do that.

Since we did the original activity above, there have been
new developments such as a public domain 64-bit coff
linker.

So now there is a purely public domain distribution
called "University Challenge x64" available at http://pdos.org

The only thing missing in the toolchain is a viable C compiler.

SubC is included, but it is only a subset of C90.

Can you tell me what the current status is of your public
domain C compiler and whether you are interested in
getting it to run on this new Win64 environment?

Thanks. Paul.

On 10/08/2023 12:29, Paul Edwards wrote:
> Hi Bart.
>
> I'm not sure what your goals are, but at one point you were
> interested in the executable loader code in PDOS, to the
> point that you even compiled it with a 64-bit C compiler
> and pointed out issues with it.
>
> I corrected those issues, but I don't think you used it again.
>
> And at one point you released a generated 64-bit compiler
> suitable for a Win64 environment with msvcrt.dll.
>
> Note that at the time I was only doing 32-bit coding, so I
> didn't test the fixes that I made, nor was I able to use
> the C compiler for anything I was trying to do.
>
> As of a couple of hours ago, that has changed.
>
> There is a version of PDOS - basically I convert the UEFI
> OS into a basic Win64 OS - there wasn't a lot of "glue"
> required to do that.
>
> Since we did the original activity above, there have been
> new developments such as a public domain 64-bit coff
> linker.
>
> So now there is a purely public domain distribution
> called "University Challenge x64" available at http://pdos.org
>
> The only thing missing in the toolchain is a viable C compiler.
>
> SubC is included, but it is only a subset of C90.
>
> Can you tell me what the current status is of your public
> domain C compiler and whether you are interested in
> getting it to run on this new Win64 environment?

I think it's best to forget my compiler:

* I've lost interest in C and in implementing it. (It seems mainly
a tool for raising my blood pressure.)

* The compiler is non conforming in many ways, and buggy

* It's not written in C, and the tool used to create a C rendering
is being deprecated

* Its output is anyway not usually an object file; it is ASM in
a special format, directly assembled into EXE files

* There was a .OBJ format that could be produced, but that is no
longer viable, as my compiler's code is intended for low memory
(within first 2GB) at a fixed address. But modern OSes and linkers
now like to load at arbitrary, high addresses, which only works with
position-independent code.

I may anyway, at any time, decide to just delete the lot. I just don't
want any external pressures or responsibilities.

However there must be 100s of small, amateur C compilers out there, you
just have to look for them. Tiny C is open source and one of the best.

On Thursday, August 10, 2023 at 7:29:52 PM UTC+8, Paul Edwards wrote:

> So now there is a purely public domain distribution
> called "University Challenge x64" available at http://pdos.org

Oh - and the source code is distributed as part of the
"University Challenge 386" distribution.

I'll probably duplicate that later, but it's still mainly
"proof of concept".

BFN. Paul.

On Thursday, August 10, 2023 at 7:57:52 PM UTC+8, Bart wrote:

> * There was a .OBJ format that could be produced, but that is no
> longer viable, as my compiler's code is intended for low memory
> (within first 2GB) at a fixed address. But modern OSes and linkers
> now like to load at arbitrary, high addresses, which only works with
> position-independent code.

We aren't using what you would presumably call "modern
OSes and linkers". Even though they are only hours/days
old, loading below 2 GB is not an issue - we control all
components except EFI, but you can request EFI to
provide memory below a certain point, which is all that is
needed.

> I may anyway, at any time, decide to just delete the lot. I just don't
> want any external pressures or responsibilities.

Ok, sorry to hear your interests have changed.

> However there must be 100s of small, amateur C compilers out there, you
> just have to look for them. Tiny C is open source and one of the best.

Almost all of them are copyrighted.

I believe yours is the most advanced public domain one,
and then there is SubC.

As you noted - your version can't really be maintained, so the
challenge would be on to bring SubC up to your (C90) level.

BFN. Paul.

On 10/08/2023 13:07, Paul Edwards wrote:
> On Thursday, August 10, 2023 at 7:57:52 PM UTC+8, Bart wrote:
>
>> * There was a .OBJ format that could be produced, but that is no
>> longer viable, as my compiler's code is intended for low memory
>> (within first 2GB) at a fixed address. But modern OSes and linkers
>> now like to load at arbitrary, high addresses, which only works with
>> position-independent code.
>
> We aren't using what you would presumably call "modern
> OSes and linkers". Even though they are only hours/days
> old, loading below 2 GB is not an issue - we control all
> components except EFI, but you can request EFI to
> provide memory below a certain point, which is all that is
> needed.
>
>> I may anyway, at any time, decide to just delete the lot. I just don't
>> want any external pressures or responsibilities.
>
> Ok, sorry to hear your interests have changed.
>
>> However there must be 100s of small, amateur C compilers out there, you
>> just have to look for them. Tiny C is open source and one of the best.
>
> Almost all of them are copyrighted.

In a post-apocalyptic scenario (I think that's partly what you want this
stuff for), I don't think many will care about copyrights!

On Thursday, August 10, 2023 at 8:16:56 PM UTC+8, Bart wrote:

> > Almost all of them are copyrighted.

> In a post-apocalyptic scenario (I think that's partly what you want this
> stuff for), I don't think many will care about copyrights!

No, a potential apocalypse is a scenario where we MAY
need to go through the entire development of processor
again.

I am after software that has been released without strings
attached. There are not 100s of C90-compliant compilers
to choose between. There are zero.

The rest of the toolchain, and editor, and OS, are all public
domain (by various people).

Only the C compiler remains, and I have seen about 6
talented people willing to donate their work to the
public domain who have tried to write a C compiler,
but all have failed to reach C90 (some haven't
officially given up yet though).

Despite the numerous number of people who like to
claim how easy it is, without actually doing it.

BFN. Paul.

On Wednesday, August 16, 2023 at 9:31:54 AM UTC+8, Bart wrote:

> > I may anyway, at any time, decide to just delete the lot. I just don't
> > want any external pressures or responsibilities.

> It's pretty much gone now together with other tools associated with C
> (converting to C or from C). It only exists now as an archived ZIP on a
> pen drive. I'll give it a week to see what to do about that.
>
> It's a relief!

Ok, you don't intend to challenge the legal status of the code
already released though, do you?

I ask because someone asked me "are you sure it's public
domain?" and I wasn't sure how to answer that.

Below is what it says - no sign of an author, but that's not
really a problem - I have other code included in PDOS
where the author wishes to remain anonymous (even
though I know his name) for some reason (which I've
never asked).

I can ask him why he has a specific concern about this
code, but I think I already know the answer - he looked
where bcc came from and noted that it wasn't public
domain, and had a known author. So people may ask
more questions - well - one person already has.

Anything you can say to help me address his concerns
(before we start using this in full knowledge that it is
unsupported and not expected to be maintainable)?

Thanks. Paul.

Toy C compiler for Windows.

This code is placed in the Public Domain.

Generated C code (not original non-C sources); build with gcc or tcc.

--------------------------------------

'BCC' Compiler for 'C'-like language.

On 28/08/2023 11:30, Paul Edwards wrote:
> On Wednesday, August 16, 2023 at 9:31:54 AM UTC+8, Bart wrote:
>
>>> I may anyway, at any time, decide to just delete the lot. I just don't
>>> want any external pressures or responsibilities.
>
>> It's pretty much gone now together with other tools associated with C
>> (converting to C or from C). It only exists now as an archived ZIP on a
>> pen drive. I'll give it a week to see what to do about that.
>>
>> It's a relief!
>
> Ok, you don't intend to challenge the legal status of the code
> already released though, do you?
>
> I ask because someone asked me "are you sure it's public
> domain?" and I wasn't sure how to answer that.
>
> Below is what it says - no sign of an author, but that's not
> really a problem - I have other code included in PDOS
> where the author wishes to remain anonymous (even
> though I know his name) for some reason (which I've
> never asked).
>
> I can ask him why he has a specific concern about this
> code, but I think I already know the answer - he looked
> where bcc came from and noted that it wasn't public
> domain, and had a known author.

Was that the Borland C Compiler?

If so that's nothing to do with me.

The file you have remains as public domain.

You can add the name 'Bart' or 'Bart C', but I don't want to divulge my
surname as it will likely be unique worldwide.

> So people may ask
> more questions - well - one person already has.
>
> Anything you can say to help me address his concerns
> (before we start using this in full knowledge that it is
> unsupported and not expected to be maintainable)?
>
> Thanks. Paul.
>
>
>
> Toy C compiler for Windows.
>
> This code is placed in the Public Domain.
>
> Generated C code (not original non-C sources); build with gcc or tcc.
>
> --------------------------------------
>
> 'BCC' Compiler for 'C'-like language.
>

On Monday, August 28, 2023 at 6:44:03 PM UTC+8, Bart wrote:

> > I can ask him why he has a specific concern about this
> > code, but I think I already know the answer - he looked
> > where bcc came from and noted that it wasn't public
> > domain, and had a known author.

> Was that the Borland C Compiler?
>
> If so that's nothing to do with me.
>
> The file you have remains as public domain.
>
> You can add the name 'Bart' or 'Bart C', but I don't want to divulge my
> surname as it will likely be unique worldwide.

Thankyou sir!

BFN. Paul.

Paul Edwards:

> The only thing missing in the toolchain is a viable C
> compile

Have you checked `lice':
https://github.com/public-domain/lice

--
() ascii ribbon campaign -- against html e-mail
/\ www.asciiribbon.org -- against proprietary attachments

On Monday, August 28, 2023 at 10:45:17 PM UTC+8, Anton Shepelev wrote:
> Paul Edwards:
> > The only thing missing in the toolchain is a viable C
> > compile

> Have you checked `lice':
> https://github.com/public-domain/lice

Jean-Marc emailed me in 2020 and said:

I found LICE ( https://github.com/public-domain/lice ) but I've not
worked on it (I can't find the name of the original author, this is a
issue).

If he's leery, I'm leery too.

BTW, this is the same guy who was leery about cc64 from Bart,
so I went back to him with this thread and asked him if that was
good enough, and he said he was happy with that, so I have now
incorporated cc64 into the PDOS source tree:

https://sourceforge.net/p/pdos/gitcode/ci/master/tree/cc64/

And the results have been fantastic. It accepted most of
PDPCLIB, and I was happy to change some minor things
that it needed.

I encountered one presumed bug, and worked around that.

I have since encountered another presumed bug when
building PDOS-generic, and worked around that too.

And I am able to build both Windows and EFI executables,
so the stack must be aligned on a 16-byte boundary, which
was a major concern (the most major concern was whether
it would accept the C90 syntax I actually use).

I am expecting to prioritize work on UCX64 at http://pdos.org
with this new tool, which gives close enough to full closure.

Thanks Bart!

BFN. Paul.

nothing

Description

Purchase DMT powder online

Buy DMT powder online is an effective hallucinogen, as well as a kind of tryptamine alkaloid. It is a normally happening material, discovered in various plants and also pets, and also in small quantities in the human mind, where its function is unidentified. DMT is famous for its power, though the psychedelic trip it creates only lasts 5 to half an hour when smoked. The impact is extensive and also remarkable, with the sensation that the user is transferred to a completely various location, submersed in kaleidoscopic noises as well as pictures. In its pure kind, the medication is a white to yellow crystalline strong.

https://methamphetaminebox.com/product/buy-dmt-crystal-online/
https://methamphetaminebox.com/product/buy-dmt-crystal-online/

Acquire DMT powder online has been taken in throughout history and into prehistory by aboriginal individuals, specifically in South America, where it is taken in throughout shamanic routines and called ayhuasca. This is done by combining plant product that contains it with a monoamine oxide inhibitor, an unique chemical that permits the medication to prevent digestion by the tummy and also get to the bloodstream. Proof of DMT consumption by aboriginal peoples in South America stretches back to a minimum of 2130 BC. A pipeline made from puma bone of that age was checked favorable for the material. Smoking it would give the individuals visions and feelings that they connected with magical sources, placing them into contact with “spirits” they can seek advice from on issues of plants, illness, and so on.

A few of the most unusual psychedelic trip reports originate from customers of DMT, who report “rotating quadrate vortices,” discussions with intelligent alien-type animals, and so forth. These reports are uncommon because of their strength and also the sensation of meeting smart beings, which is evocative what happens to lots of people each night in desires. Though clinical examination of the results of the medication has actually been restricted, cognitive scientific research may be able to find out more concerning the human mind by seeing just how it transforms its procedure in feedback to tryptamines. Spiritualists might be inclined to believe that the beings that individuals “fulfill” intoxicated might in fact feed on identical aircrafts, which has actually introduced alternating religion systems or worldviews based on the experience.

DMT is a powerful hallucinogen, suggested to be thoroughly administered in a tranquil setting to somebody who has prior experience with various other psychedelic drugs. The medicine is relatively uncommon because of the lack of industrial demand as well as the shortage of individuals with the understanding and also motivation to separate it from plants. Still, as a molecule, it appears like a terrain ripe for exploration. Untried conjectures have actually suggested that the DMT located normally in the mind may be implicated in specific neurological states, and if it is synthetically provided, it may pull these “switches as well as bars” in manner ins which can be extra exactly defined and studied. As the human brain is the most complicated well-known things in the universe, identifying the specific way in which it engages with complicated particles like this may be just one of the biggest clinical obstacles of all time.

Get DMT powder online is a white crystalline powder that is stemmed from specific plants discovered in Mexico, South America as well as parts of Asia.

It is usually evaporated or smoked in a pipeline, or eaten by mouth in brews like ayahuasca. Sometimes, DMT is grunted or injected.

Acquire DMT powder online’s chemical root structure is similar to the anti-migraine medicine sumatriptan, and it functions as a non-selective agonist at most or all of the serotonin receptors. Serotonin is a neurotransmitter that widely affects most of our body’s mind cells.

When smoked, the ordinary dose of DMT is 30-150 mg, and also the start of action can be really felt virtually promptly. The results peak and also plateau for 3-5 minutes, and slowly leave with the duration of impact totaling 30-45 mins.

When eaten as a brew, the dose is in between 35-75 mg. Effects begin after 30-45 mins, optimal at 2-3 hours and are resolved in 4-6 hours.
Road names of Buy DMT powder online

Dmitri
Businessman’s journey
Businessman’s unique
Fantasia
Forty-five-minute psychosis.

Degree of DMT use

Purchase DMT powder online has no approved medical use in the US however can be made use of by researchers under a Schedule I study registration that requires authorization from both the DEA as well as the Fda (FDA).

DMT is utilized illicitly for its psychoactive, hallucinogenic impacts. Customers report “spiritual understanding” as one of the most commonly reported favorable impacts of the drug.

The vast bulk of new DMT individuals are already experienced with making use of hallucinogens, and also as holds true with other unlawful hallucinogens, customers acquire the drug through the Internet.

Use DMT has actually enhanced in recent years, suggesting that the medicine might be obtaining in popularity. The National Study on Drug Use and Health and wellness disclosed that the number of individuals in the United States that reported making use of some kind of DMT had actually increased from 688,000 individuals in 2006 to 1,475,000 in 2012.
Adverse effects of Buy DMT powder online

A person is having a surreal hallucination with clocks.

The primary result of DMT is the experience of intense hallucinations that change the individual’s assumption of the globe around them.

The main effect of DMT is mental, with extreme visual as well as acoustic hallucinations, ecstasy and also a modified feeling of space, body as well as time. Several individuals explain profound, life-changing experiences such as going to other worlds, chatting with alien entities as well as full changes in the assumption of identification and also fact.

In contrast to various other psychedelic drugs (LSD, ketamine, magic mushrooms), recreational users of DMT consider it to have the most affordable adverse effects account.
Possible side effects of Buy DMT powder online may include:

Enhanced heart price
Raised high blood pressure
Chest pain or rigidity
Agitation
Dilated students
Quick rhythmic movements of the eye
Wooziness

When taken orally, DMT can cause queasiness, throwing up as well as looseness of the bowels.

Depending upon the specific customer, the DMT experience can be either extremely amazing or overwhelmingly frightening. The experience can be so effective that users might have problem handling as well as integrating the “trip” right into their reality. Mental adverse effects may remain for numerous days or weeks after ingestion of the medicine.
Health threats of DMT

Due to the fact that DMT is structurally related to the natural chemical serotonin, a problem called serotonin syndrome is a possibly dangerous health and wellness threat that can be connected with its use. People taking antidepressants go to greatest risk for this problem.

Serotonin syndrome happens when the body builds up a too much amount of serotonin. The condition is often brought on by taking a mix of different medicines. Way too much serotonin in the body can cause signs and symptoms such as:

Anxiety
Confusion
Hypertension
Loss of muscle coordination
Frustration

At higher dosages, Purchase DMT powder online can create seizures, breathing arrest and coma.

DMT can have significant unfavorable consequences for individuals with pre-existing mental problems or a mental disorder such as schizophrenia.

As a result of limited research data, DMT is not known to create physical reliance or addiction, although frequent recreational customers may create emotional food cravings for the medicine. The National Institute on Drug Abuse (NIDA) recommend that, unlike various other hallucinogens, DMT usage does not appear to cause tolerance of the medication.

5-MeO-DMT, additionally called 5-Methoxy-N, N-dimethyltryptamine is a normally occurring psychedelic of the tryptamine class, 5-MeO-DMT exposes structural similarities to N, N-Dimethyltryptamine (DMT), nonetheless studies have actually shown chemical characteristics that may resemble a greater effectiveness than the better recognized substance. It is distributed in a wide range of plant types and also in the venom of a few psychedelic pathogens such as Bufo Alvaris.

PURCHASE 5-MeO-DMT

5-MeO-DMT, additionally referred to as 5-Methoxy-N, N-dimethyltryptamine is a normally occurring psychedelic of the tryptamine class, 5-MeO-DMT bares architectural resemblances to N, N-Dimethyltryptamine (DMT), nevertheless researches have actually revealed chemical features that might appear like a greater strength than the better recognized substance. It is distributed in a wide array of plant types and in the poison of a couple of psychedelic virus such as Bufo Alvaris.

5-MeO-DMT can be located in the milky white venom of the Colorado River Toad. South American medicine men have actually used toad poisonous substance for hundreds of years Today, both the extracts of the toad poisonous substance and the synthetic powder form are examined. Its synthesis as well as outcomes were first documented in Alexander Shulgin’s 1997 book TiHKAL (” Tryptamines I Have Actually Known and Liked”). Recent researches have revealed that it can supply magical end results during research study and has actually expanded in popularity over the past couple of years.

Just how to utilize 5-MeO-DMT?

Buy 5-MeO-DMT in the best quality freebase as well as hydrochloride type. Buy 5-MeO-DMT research chemicals in the quantities 0,25 gr, 0,5 gr, 1gr, 2gr, 5gr, and 10gr. Make certain to keep the 5-MeO-DMT in a completely dry as well as great place for maximum shelf-life. When handling study chemicals ensure to constantly take the appropriate precautions busy like cleaning down surface areas and wearing handwear covers, a mask & protective garments. We advise additional caution with 5-MeO-DMT as a result of its high potency..

On 29/08/2023 03:19, Paul Edwards wrote:
> On Monday, August 28, 2023 at 10:45:17 PM UTC+8, Anton Shepelev wrote:
>> Paul Edwards:
>>> The only thing missing in the toolchain is a viable C
>>> compile
>
>> Have you checked `lice':
>> https://github.com/public-domain/lice
>
> Jean-Marc emailed me in 2020 and said:
>
> I found LICE ( https://github.com/public-domain/lice ) but I've not
> worked on it (I can't find the name of the original author, this is a
> issue).
>
>
> If he's leery, I'm leery too.
>
> BTW, this is the same guy who was leery about cc64 from Bart,
> so I went back to him with this thread and asked him if that was
> good enough, and he said he was happy with that, so I have now
> incorporated cc64 into the PDOS source tree:
>
> https://sourceforge.net/p/pdos/gitcode/ci/master/tree/cc64/
>
> And the results have been fantastic. It accepted most of
> PDPCLIB, and I was happy to change some minor things
> that it needed.
>
> I encountered one presumed bug, and worked around that.
>
> I have since encountered another presumed bug when
> building PDOS-generic, and worked around that too.

You can try raising issues here:

https://github.com/sal55/langs/tree/master

I just need to notice that something has been posted, as I don't check
it often.

On Wednesday, August 30, 2023 at 5:01:32 AM UTC+8, Bart wrote:
> > I encountered one presumed bug, and worked around that.
> >
> > I have since encountered another presumed bug when
> > building PDOS-generic, and worked around that too.

> You can try raising issues here:
>
> https://github.com/sal55/langs/tree/master
>
> I just need to notice that something has been posted, as I don't check
> it often.

Sorry - I thought you were very clear that this is unsupported
software and you don't want the headache? I checked the
link and it clearly says that bcc is archived.

Here are the workarounds I needed:

__PDPCLIB_HEADFUNC FILE **__gtin(void);

#define __stdin (*(__gtin()))

#ifdef __CC64__
*__stdin->fbuf = '\0';
__stdin->fbuf++;
*__stdin->fbuf = '\0';
__stdin->fbuf++;
#else
*__stdin->fbuf++ = '\0';
*__stdin->fbuf++ = '\0';
#endif

static int (*genstart)(OS *bios);

#ifdef __CC64__
rc = (*genstart)(&bios);
#else
rc = genstart(&bios);
#endif

Regardless, things are looking REALLY good and this is
likely to change my direction in life (as a Win64 vendor
on modern machines instead of scouring landfill for
Compaq BIOSes).

BFN. Paul.

One more issue has just come up:

From the author of pdld analyzing a problem I sent him:

The kernel32.obj produced by cc64 is lacking .drectve section which tells the linker which symbols should be exported.

We're still seeing if we can work around this problem.

BFN. Paul.

On Wednesday, August 30, 2023 at 3:25:27 PM UTC+8, Paul Edwards wrote:

> We're still seeing if we can work around this problem.

An "export-all-symbols" option was added to pdld and
that seems to be sufficient for my immediate needs.

BFN. Paul.

On 30/08/2023 04:00, Paul Edwards wrote:
> On Wednesday, August 30, 2023 at 5:01:32 AM UTC+8, Bart wrote:
>>> I encountered one presumed bug, and worked around that.
>>>
>>> I have since encountered another presumed bug when
>>> building PDOS-generic, and worked around that too.
>
>> You can try raising issues here:
>>
>> https://github.com/sal55/langs/tree/master
>>
>> I just need to notice that something has been posted, as I don't check
>> it often.
>
> Sorry - I thought you were very clear that this is unsupported
> software and you don't want the headache? I checked the
> link and it clearly says that bcc is archived.

I can't guarantee anything, but sometimes I will try and fix something I
find irksome, just for my own satisfaction.

(The bcc project is a major test program for my own language.)

> Here are the workarounds I needed:
>
> __PDPCLIB_HEADFUNC FILE **__gtin(void);
>
> #define __stdin (*(__gtin()))
>
> #ifdef __CC64__
> *__stdin->fbuf = '\0';
> __stdin->fbuf++;
> *__stdin->fbuf = '\0';
> __stdin->fbuf++;
> #else
> *__stdin->fbuf++ = '\0';
> *__stdin->fbuf++ = '\0';
> #endif
>
>
>
> static int (*genstart)(OS *bios);
>
> #ifdef __CC64__
> rc = (*genstart)(&bios);
> #else
> rc = genstart(&bios);
> #endif

The latter bug (needing to call a function pointer F as (*F)() rather
than F()), I know about. The increment one is new. However, I can't
reproduce it. This program:

-----------------------------
#include <stdio.h>

//typedef char T;
typedef short T;
//typedef int T;

typedef struct {
int dummy;
T* fbuf;
} rec;

int main(void) {
T a[] = {10,20,30,40,50};
rec x = {999, NULL};
rec* p = &x;

x.fbuf = &a[2];

printf("*p->fbuf = %d\n", *p->fbuf);

*p->fbuf++ = 77; // From your example

for (int i=0; i<5; ++i) {
printf("%d: %d\n",i, a[i]);
}
printf("*p->fbuf = %d\n", *p->fbuf);
} -----------------------------

gives the same results with both bcc and gcc, whatever T is chosen. What
is the actual type of fbuf? And at what offset is it within __stdin? I
tried with and without that dummy field.

In any case, the code you've had to use for CC64 should work with other
compilers too.

Hi Bart.

I have encountered the below problem.

mingw64 works, but cc64 crashes on the fprintf.

Any idea? I will answer your other post now.

Thanks. Paul.

C:\devel\pdos\pdpclib>temp2

C:\devel\pdos\pdpclib>x86_64-w64-mingw32-gcc -S -O2 -D__WIN32__ -D__64BIT__ -I . -I ../src -o pdptest.s pdptest.c

C:\devel\pdos\pdpclib>pdas --oformat coff --64 -o pdptest.obj pdptest.s

C:\devel\pdos\pdpclib>rm -f pdptest.s

C:\devel\pdos\pdpclib>pdld -s -nostdlib --no-insert-timestamp -o pdptest.exe w32start.obj pdptest.obj msvcrt.lib

C:\devel\pdos\pdpclib>pdptest
xhello 5
yhello, world

C:\devel\pdos\pdpclib>temp

C:\devel\pdos\pdpclib>cc64 -c -out:pdptest.obj pdptest.c
Compiling pdptest.c to pdptest.obj

C:\devel\pdos\pdpclib>pdld -s -nostdlib --no-insert-timestamp --image-base 0x400000 -o pdptest.exe w32start.obj pdptest.obj msvcrt.lib

C:\devel\pdos\pdpclib>pdptest
xhello 5

C:\devel\pdos\pdpclib>type pdptest.c

int fprintf(void *stream, const char *format, ...);
int printf(const char *format, ...);

extern char *__imp__iob;

int main(void)
{ printf("xhello %d\n", 5);
fprintf(__imp__iob + 48, "yhello, world\n");
return (0);
}

C:\devel\pdos\pdpclib>

On Wednesday, August 30, 2023 at 6:38:30 PM UTC+8, Bart wrote:

> I can't guarantee anything, but sometimes I will try and fix something I
> find irksome, just for my own satisfaction.

Ok, but it is cc64 that I need (the thing you released to
the public domain), and that is generated code, and you
can no longer generate it, and it's nominally not
maintainable.

> than F()), I know about. The increment one is new. However, I can't
> reproduce it. This program:

I confirmed the problem is still occurring.

> rec x = {999, NULL};
> rec* p = &x;
>
> x.fbuf = &a[2];
>
> printf("*p->fbuf = %d\n", *p->fbuf);
>
> *p->fbuf++ = 77; // From your example

I am doing a function call. It's not a simple struct.
I also don't know whether you are using the same
cc64 that I am.

> gives the same results with both bcc and gcc, whatever T is chosen. What
> is the actual type of fbuf? And at what offset is it within __stdin? I
> tried with and without that dummy field.

char * and 32.

C:\devel\pdos\pdpclib>grep fbuf stdio.h | grep char
stdio.h: char *fbuf; /* file buffer - this is what all the routines

C:\devel\pdos\pdpclib>git diff .
diff --git a/pdpclib/pdptest.c b/pdpclib/pdptest.c
index b877649e..348b587b 100644
--- a/pdpclib/pdptest.c
+++ b/pdpclib/pdptest.c
@@ -14,6 +14,7 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
+#include <stddef.h>

static char buf[6144]; /* arbitrary buffer size */

@@ -37,6 +38,7 @@ int main(int argc, char **argv)
#endif

printf("welcome to pdptest\n");
+ printf("%d\n", offsetof(FILE, fbuf));
#ifndef SEGHACK

#if defined(__SUBC__) && defined(__64BIT__)

C:\devel\pdos\pdpclib>pdptest
welcome to pdptest
32
main function is at 00409020
allocating 10 bytes
m1 is 0009AA48
allocating 20 bytes
m2 is 000920D8
stack is around 0061FDD8
usage: pdptest [-bb/-tt/-tb/-bt] <infile> <outfile>
default is text to text copy

C:\devel\pdos\pdpclib>

__PDPCLIB_HEADFUNC FILE **__gtin(void);

#define __stdin (*(__gtin()))

Relevant code is here:

https://sourceforge.net/p/pdos/gitcode/ci/master/tree/pdpclib/

And to reproduce the problem do this:

C:\devel\pdos\pdpclib>git diff .
diff --git a/pdpclib/start.c b/pdpclib/start.c
index a2662643..e5a6c412 100644
--- a/pdpclib/start.c
+++ b/pdpclib/start.c
@@ -523,7 +523,7 @@ __PDPCLIB_API__ int CTYP __start(char *p)
__stdin->bufTech = _IOLBF;
__stdin->intBuffer = buffer1;
__stdin->fbuf = __stdin->intBuffer + 2;
-#ifdef __CC64__
+#ifdef __XCC64__
*__stdin->fbuf = '\0';
__stdin->fbuf++;
*__stdin->fbuf = '\0';

C:\devel\pdos\pdpclib>

C:\devel\pdos\pdpclib>pdmake -f makefile.c64

C:\devel\pdos\pdpclib>pdptest

C:\devel\pdos\pdpclib>

(crashes without displaying anything)

But reproducing it requires the supporting utilities from
PDOS/386 plus:

C:\devel\pdos\src>pdmake -f makek32.s64

before running the above, in order to produce kernel32.lib.
You might be able to get kernel32.lib from elsewhere.

Also during development I discovered some problems with
pdld and those are only in the source tree, not the PDOS
disk, but I don't think they are required for the above (and
that is why I didn't mention makek32.n64).

BFN. Paul.

On 30/08/2023 15:06, Paul Edwards wrote:
> Hi Bart.
>
> I have encountered the below problem.
>
> mingw64 works, but cc64 crashes on the fprintf.
>
> Any idea? I will answer your other post now.
>

> C:\devel\pdos\pdpclib>pdld -s -nostdlib --no-insert-timestamp --image-base 0x400000 -o pdptest.exe w32start.obj pdptest.obj msvcrt.lib
>
> C:\devel\pdos\pdpclib>pdptest
> xhello 5
>
> C:\devel\pdos\pdpclib>type pdptest.c
>
> int fprintf(void *stream, const char *format, ...);
> int printf(const char *format, ...);
>
> extern char *__imp__iob;
>
> int main(void)
> {
> printf("xhello %d\n", 5);
> fprintf(__imp__iob + 48, "yhello, world\n");
> return (0);
> }

It's not clear which C library is being linked in here. BCC is designed
to use MSVCRT.DLL (that is reflected in headers like stdio.h).

Because '__imp__iob' is not exported by msvcrt.dll; '_iob' is as a
variable, but that crashes too (even after fiddling with the generated ASM).

(Or maybe that extra '__imp_' prefix is an artefact of the linking
method that is being used, if using msvcrt.lib to access the same .dll.)

I'm not sure how well BCC links to variables exported from DLLS rather
than functions, but your test doesn't appear to use my internal linker
anyway.

I managed to get your example working by using a function instead of a
variable. See that version below. Note that in my stdio.h, 'stdout' is
defined as:

#define stdout ((FILE*)(__iob_func()+sizeof(FILE)))

sizeof(FILE) is 48.

--------------------------------------------

int fprintf(void *stream, const char *format, ...);
int printf(const char *format, ...);
extern char* __iob_func(void);

int main(void)
{ printf("xhello %d\n", 5);
fprintf(__iob_func() + 48, "yhello, world\n");
return (0);
}

On 30/08/2023 15:51, Paul Edwards wrote:
> On Wednesday, August 30, 2023 at 6:38:30 PM UTC+8, Bart wrote:
>
>> I can't guarantee anything, but sometimes I will try and fix something I
>> find irksome, just for my own satisfaction.
>
> Ok, but it is cc64 that I need (the thing you released to
> the public domain), and that is generated code, and you
> can no longer generate it, and it's nominally not
> maintainable.
>
> > than F()), I know about. The increment one is new. However, I can't
>> reproduce it. This program:
>
> I confirmed the problem is still occurring.
>
>> rec x = {999, NULL};
>> rec* p = &x;
>>
>> x.fbuf = &a[2];
>>
>> printf("*p->fbuf = %d\n", *p->fbuf);
>>
>> *p->fbuf++ = 77; // From your example
>
> I am doing a function call. It's not a simple struct.

I missed that completely. I'll have to do a bigger test!

On 30/08/2023 16:09, Bart wrote:
> On 30/08/2023 15:51, Paul Edwards wrote:
>> On Wednesday, August 30, 2023 at 6:38:30 PM UTC+8, Bart wrote:
>>
>>> I can't guarantee anything, but sometimes I will try and fix something I
>>> find irksome, just for my own satisfaction.
>>
>> Ok, but it is cc64 that I need (the thing you released to
>> the public domain), and that is generated code, and you
>> can no longer generate it, and it's nominally not
>> maintainable.
>>
>>   > than F()), I know about. The increment one is new. However, I can't
>>> reproduce it. This program:
>>
>> I confirmed the problem is still occurring.
>>
>>> rec x = {999, NULL};
>>> rec* p = &x;
>>>
>>> x.fbuf = &a[2];
>>>
>>> printf("*p->fbuf = %d\n", *p->fbuf);
>>>
>>> *p->fbuf++ = 77; // From your example
>>
>> I am doing a function call. It's not a simple struct.
>
> I missed that completely. I'll have to do a bigger test!
>
>

It took me nearly half an hour to sort out the right combinations of
"*", "." and "->" operations for this test.

BCC's code goes wrong here in this equivalent code:

*(*F())->fbuf++ = 77;

It's going to be too much effort to fix. I set up the same test in my
main non-C language, where it looks like this:

F().fbuf++^:=77

(Two of the three necessary derefs are applied automatically.)

Here it works OK. I think any future efforts in the BCC code generator,
if I ever get back to it, are better spent in porting the back end from
the other compiler.