Richard Kettlewell <invalid@invalid.invalid> writes:

> Dan Espen <dan1espen@gmail.com> writes:
>> Richard Kettlewell <invalid@invalid.invalid> writes:
>>> "26C.Z968" <26C.Z968@noaada.net> writes:
>>>> DO wanna know the latest trix in poking the keyboard buffer though
>>>> ... FINDING it is the first big trick. These days it's all very
>>>> 'abstracted' - local instances buried really deep with obscure names
>>>> and such. OK for most uses UNLESS you wanna automatically put text
>>>> into annoying prompts/forms over and over and over.
>>>
>>> If you want to insert characters for subsequent reading into your
>>> terminal’s input buffer, you want the TIOCSTI ioctl. See
>>> https://linux.die.net/man/4/tty_ioctl.
>>
>> An ioctl is called by the application.
>> Pretty sure you want to use X11's synthetic keystrokes.
>> Some applications accept them, some don't.
>
> That’s a good additional option to include. If OP has been more specific
> about their needs then I missed it.

Well, he missed a lot. Starting with, there's nothing wrong with the
way Python reads it's input now.

--
Dan Espen

On 2/8/23 1:34 PM, Dan Espen wrote:
> Richard Kettlewell <invalid@invalid.invalid> writes:
>
>> Dan Espen <dan1espen@gmail.com> writes:
>>> Richard Kettlewell <invalid@invalid.invalid> writes:
>>>> "26C.Z968" <26C.Z968@noaada.net> writes:
>>>>> DO wanna know the latest trix in poking the keyboard buffer though
>>>>> ... FINDING it is the first big trick. These days it's all very
>>>>> 'abstracted' - local instances buried really deep with obscure names
>>>>> and such. OK for most uses UNLESS you wanna automatically put text
>>>>> into annoying prompts/forms over and over and over.
>>>>
>>>> If you want to insert characters for subsequent reading into your
>>>> terminal’s input buffer, you want the TIOCSTI ioctl. See
>>>> https://linux.die.net/man/4/tty_ioctl.
>>>
>>> An ioctl is called by the application.
>>> Pretty sure you want to use X11's synthetic keystrokes.
>>> Some applications accept them, some don't.
>>
>> That’s a good additional option to include. If OP has been more specific
>> about their needs then I missed it.
>
> Well, he missed a lot. Starting with, there's nothing wrong with the
> way Python reads it's input now.

Nothing WRONG with it at all ... but limitations MAY be
encountered. Alternatives are always good. They can get
you out of that corner you inadvertently painted yerself
into :-)

On 2/7/23 22:28, 26C.Z968 wrote:
> On 2/7/23 9:46 PM, Jack Strangio wrote:
>> vjp2.at@at.BioStrategist.dot.dot.com writes:
>>>
>>> How do you put a partition inside a partition?
>>>
>>
>> A partition, or a filesystem??
>>
>> We tend to use both of these terms interchangeably, but they ARE NOT THE
>> SAME THING.
>
>   VERY true !
>
>> Normally we fill a disk-partition with just one filesystem, though  MBR
>> 'extended partitions' can hold several 'logical partitions'.
>
>
>   GPT has effectively ended the need for 'extended' partitions.
>   However it's still good for people to KNOW about them. I still
>   have a couple older boxes/systems that are MBR with extended
>   partitions and no UEFI. If you CAN do it the older, simpler,
>   way - why not ? Lots more documentation.

The older and simpler way did not look so to me but then I learned to
do partitioning on the Amiga 1000 with an expansion box
and GVP tools on a SCSI disk and the Amiga's own Rigid Disk Block
(which I am surprised I remembered the RDB proper name). It was more
like the present GTP than the MBR. The limitation of the MBR used
to drive me up the wall trying to figure out what was failing in
a Logical Partition. Of course the disks were a small fraction of
the present sizes and the AmigaOS was 32 bit much before the i386.

>
>
>> This most likely what you wanted the answer to, but you CAN put a
>> 'partition' within a
>> partition by formatting a normal file as a filesystem.
>
>   Ummm ... that's REALLY kludgy and a "file" really
>   isn't a "file system". For 99.9% these days I'd
>   rec GPT and straight-up GPT partitions with single
>   filesystems on them. It's KISS.
>
>> So (for instance) you might have a 100MB partition that hold ten 10 MB
>> files, each of which holds a filesystem and each of those can be mounted
>> just like a disk-partition can be mounted.
>>
>> ====================================
>> #!/bin/bash
>> #jvs script "loopmount"
>> # mounts .iso image file on /cd
>> # mounts filesystem-file on mount-point
>> #
>> if [ "$2" != "" ]; then
>>          MOUNTPOUNT=$2
>> else
>>          MOUNTPOUNT=/cd
>> fi
>>
>> if [ "$3" != "" ]; then
>>          TYPE=$3
>> else
>>          TYPE=iso9660
>> fi
>>
>> if [ -z $1 ]; then
>>          echo "usage: loopmount   <imagefile-name>  ( will be mounted
>> on /cd )  (will be iso9660)"
>>          echo "       loopmount   <imagefile-name>
>> <mount-point>              <filesystem type>"
>> else
>>          LOOP=$(sudo losetup --find --show $1)
>>          echo "sudo /bin/mount -t${TYPE} -oloop $LOOP  $MOUNTPOUNT"
>>          sudo /bin/mount -t${TYPE} -oloop $LOOP $MOUNTPOUNT
>> fi
>> ====================================
>
>
>   AAAaauuugghhh !!!  :-)
>
>   CAN do and SHOULD do are NOT the same thing !
>
Indubitably.

bliss - on the ever-faithful Dell Latitude E7450, PCLinuxOS 2022
KDE Plasma 5.26.5 Kernel Version: 6.1.10-pclos1 (64-bit)
KDE Frameworks 5.102.0 - Qt Version: 5.15.6
Graphics : X11 - Mesa Intel® HD Graphics 5500
15.5 GiB of RAM CPU 4 × Intel® Core™ i7-5600U CPU @ 2.60GHz
Actually 2 real cores and 2 virtual cores.

--
bliss dash SF 4 ever at dslextreme dot com

On 2/8/23 10:55 PM, Bobbie Sellers wrote:
> On 2/7/23 22:28, 26C.Z968 wrote:
>> On 2/7/23 9:46 PM, Jack Strangio wrote:
>>> vjp2.at@at.BioStrategist.dot.dot.com writes:
>>>>
>>>> How do you put a partition inside a partition?
>>>>
>>>
>>> A partition, or a filesystem??
>>>
>>> We tend to use both of these terms interchangeably, but they ARE NOT THE
>>> SAME THING.
>>
>>    VERY true !
>>
>>> Normally we fill a disk-partition with just one filesystem, though  MBR
>>> 'extended partitions' can hold several 'logical partitions'.
>>
>>
>>    GPT has effectively ended the need for 'extended' partitions.
>>    However it's still good for people to KNOW about them. I still
>>    have a couple older boxes/systems that are MBR with extended
>>    partitions and no UEFI. If you CAN do it the older, simpler,
>>    way - why not ? Lots more documentation.
>
>     The older and simpler way did not look so to me but then I learned
> to do partitioning on the Amiga 1000 with an expansion box
> and GVP tools on a SCSI disk and the Amiga's own Rigid Disk Block
> (which I am surprised I remembered the RDB proper name). It was more
> like the present GTP than the MBR.  The limitation of the MBR used
> to drive me up the wall trying to figure out what was failing in
> a Logical Partition.  Of course the disks were a small fraction of
> the present sizes and the AmigaOS was 32 bit much before the i386.
>
>>
>>
>>> This most likely what you wanted the answer to, but you CAN put a
>>> 'partition' within a
>>> partition by formatting a normal file as a filesystem.
>>
>>    Ummm ... that's REALLY kludgy and a "file" really
>>    isn't a "file system". For 99.9% these days I'd
>>    rec GPT and straight-up GPT partitions with single
>>    filesystems on them. It's KISS.
>>
>>> So (for instance) you might have a 100MB partition that hold ten 10 MB
>>> files, each of which holds a filesystem and each of those can be mounted
>>> just like a disk-partition can be mounted.
>>>
>>> ====================================
>>> #!/bin/bash
>>> #jvs script "loopmount"
>>> # mounts .iso image file on /cd
>>> # mounts filesystem-file on mount-point
>>> #
>>> if [ "$2" != "" ]; then
>>>          MOUNTPOUNT=$2
>>> else
>>>          MOUNTPOUNT=/cd
>>> fi
>>>
>>> if [ "$3" != "" ]; then
>>>          TYPE=$3
>>> else
>>>          TYPE=iso9660
>>> fi
>>>
>>> if [ -z $1 ]; then
>>>          echo "usage: loopmount   <imagefile-name>  ( will be mounted
>>> on /cd )  (will be iso9660)"
>>>          echo "       loopmount   <imagefile-name>
>>> <mount-point>              <filesystem type>"
>>> else
>>>          LOOP=$(sudo losetup --find --show $1)
>>>          echo "sudo /bin/mount -t${TYPE} -oloop $LOOP  $MOUNTPOUNT"
>>>          sudo /bin/mount -t${TYPE} -oloop $LOOP $MOUNTPOUNT
>>> fi
>>> ====================================
>>
>>
>>    AAAaauuugghhh !!!  :-)
>>
>>    CAN do and SHOULD do are NOT the same thing !
>>
>     Indubitably.
>
> bliss - on the ever-faithful Dell Latitude E7450, PCLinuxOS 2022
> KDE Plasma 5.26.5 Kernel Version: 6.1.10-pclos1 (64-bit)
> KDE Frameworks  5.102.0 - Qt Version: 5.15.6
> Graphics : X11 - Mesa Intel® HD Graphics 5500
> 15.5 GiB of RAM CPU 4 × Intel® Core™ i7-5600U CPU @ 2.60GHz
> Actually 2 real cores and 2 virtual cores.

Amiga-OS was indeed "ahead of its time" for PCs of
that era. Yes, you COULD buy some UNIX versions
like (MS) Xenix and SCO Unix, but the prices were
prohibitive.

The big horrible problem is that they released the
A-1000/AOS maybe a year too early - before the bugs
were worked-out. 'Marketing' decision fer sure. They
knew the PCs and MACs were moving in the same direction
and HAD to beat them.

And UNIX didn't do all those cool graphics - it was
kinda 'academic' at the time.

There was considerable debate at IBM over whether to
use the M68000 series or the Intel chips back in the
day. They SHOULD have gone with Motorola - but there
were some reasons for the 8088. First of all it needed
less wiring. Secondly its instruction set was similar
enough to the Z80's so it was easy to port CP/M. That
was their backwards-compatibility promise ... at the
time BIZ=CP/M and IBM was BIZ. It turned out DOS was
the much more popular system - but IBM just HAD to
offer CP/M-86. CP/M-68k came out a little later. Still
have one of those 86 disks somewhere and it WOULD boot
up until Intel decided not to support 8-bit code anymore.
I think the Core2-Quad was the last best chip CP/M-86
would boot with (still have a couple of those boards
and they ain't all so bad either). You can still run
CP/M-86 on VirtualBox however.

Ya know, I did like the original MACs - the "bricks".
They were 'cute' and did lotsa stuff that was really
hard to do on PCs. The mindset was kinda alien however
and I came to dislike it intensely. Also, these days,
I'd have to wear magnifying goggles to READ the damned
tiny little screen ! Had to do that while setting up
a bunch of iPads recently :-)

26C.Z968 <26C.Z968@noaada.net> wrote:
> There was considerable debate at IBM over whether to use the M68000
> series or the Intel chips back in the day. They SHOULD have gone
> with Motorola - but there were some reasons for the 8088. First of
> all it needed less wiring. Secondly its instruction set was
> similar enough to the Z80's so it was easy to port CP/M. That was
> their backwards-compatibility promise ... at the time BIZ=CP/M and
> IBM was BIZ.

The reasoning given not long (as in early 'history' articles) after the
release of the IBM PC indicated a much simpler reasoning.

Intel had an official licensed second source supplier for 808x chips
(that second source being AMD) and Motorola was, at least at the time
IBM's PC group was picking a CPU, single source (with a promise to have
a second source within X time, like a year or something).

And IBM's corporate policy was to always have two sources for all
critical components.

So while 'less wiring' and 'could run CP/M if need be' may have figured
into the choice, if these history articles were correct, the actual
reason was much simpler.

Wikipedia as of today
(https://en.wikipedia.org/wiki/IBM-PC#Design_process) cites that the
68000 was not production ready at the time:

Several CPUs were considered, including the Texas Instruments
TMS9900, Motorola 68000 and Intel 8088. The 68000 was considered
the best choice, but was not production-ready like the others. The
IBM 801 RISC processor was also considered, since it was
considerably more powerful than the other options, but rejected due
to the design constraint to use off-the-shelf parts.

IBM chose the 8088 over the similar but superior 8086 because Intel
offered a better price for the former and could provide more
units,[19] and the 8088's 8-bit bus reduced the cost of the rest of
the computer. The 8088 had the advantage that IBM already had
familiarity with the 8085 from designing the IBM System/23
Datamaster. The 62-pin expansion bus slots were also designed to
be similar to the Datamaster slots, and its keyboard design and
layout became the Model F keyboard shipped with the PC, but
otherwise the PC design differed in many ways.

On 2023-02-10 15:08, Rich wrote:
> The reasoning given not long (as in early 'history' articles) after the
> release of the IBM PC indicated a much simpler reasoning.
>
> Intel had an official licensed second source supplier for 808x chips
> (that second source being AMD) and Motorola was, at least at the time
> IBM's PC group was picking a CPU, single source (with a promise to have
> a second source within X time, like a year or something).
>
> And IBM's corporate policy was to always have two sources for all
> critical components.

I heard of this.

--
Cheers, Carlos.

On 2/10/23 9:08 AM, Rich wrote:
> 26C.Z968 <26C.Z968@noaada.net> wrote:
>> There was considerable debate at IBM over whether to use the M68000
>> series or the Intel chips back in the day. They SHOULD have gone
>> with Motorola - but there were some reasons for the 8088. First of
>> all it needed less wiring. Secondly its instruction set was
>> similar enough to the Z80's so it was easy to port CP/M. That was
>> their backwards-compatibility promise ... at the time BIZ=CP/M and
>> IBM was BIZ.
>
> The reasoning given not long (as in early 'history' articles) after the
> release of the IBM PC indicated a much simpler reasoning.
>
> Intel had an official licensed second source supplier for 808x chips
> (that second source being AMD) and Motorola was, at least at the time
> IBM's PC group was picking a CPU, single source (with a promise to have
> a second source within X time, like a year or something).
>
> And IBM's corporate policy was to always have two sources for all
> critical components.

The 8088 was just cheaper and "safer" in several respects,
so that's what they went with. CP/M compatibility WAS a
significant issue as well. CP/M-68k didn't come out for
another year or two (Radio Shack DID have a dual-board
computer that'd run CP/M-68k, once it arrived).

> So while 'less wiring' and 'could run CP/M if need be' may have figured
> into the choice, if these history articles were correct, the actual
> reason was much simpler.

BIG company, LOTS of issues - real or imagined - LOTS
of meetings and meetings about meetings about .....

A terrible thing they DIDN'T go with the 68 series
though, it WAS a much better chip - which Apple soon
exploited. By now we'd have the 680990 chip and Intel
would still be making accessory hardware.

> Wikipedia as of today
> (https://en.wikipedia.org/wiki/IBM-PC#Design_process) cites that the
> 68000 was not production ready at the time:
>
> Several CPUs were considered, including the Texas Instruments
> TMS9900, Motorola 68000 and Intel 8088. The 68000 was considered
> the best choice, but was not production-ready like the others. The
> IBM 801 RISC processor was also considered, since it was
> considerably more powerful than the other options, but rejected due
> to the design constraint to use off-the-shelf parts.
>
> IBM chose the 8088 over the similar but superior 8086 because Intel
> offered a better price for the former and could provide more
> units,[19] and the 8088's 8-bit bus reduced the cost of the rest of
> the computer. The 8088 had the advantage that IBM already had
> familiarity with the 8085 from designing the IBM System/23
> Datamaster. The 62-pin expansion bus slots were also designed to
> be similar to the Datamaster slots, and its keyboard design and
> layout became the Model F keyboard shipped with the PC, but
> otherwise the PC design differed in many ways.
>

I have a little experience with the TMS-9000 series. AT THE
TIME it was a 'clever' chip - hardware-support for multi-user
environments. However it was big and expensive and TI was
super-anal - a reason the TI Home Computer died so soon. Also
the 9900's only easily supported 64kb "workspaces" for users
apps which would have been too limiting given the way the
future was developing.

I have no experience with the IBM-801 so I can't say much.
Their PowerPC chip probably evolved from it and was kinda
successful (most Linux's still have a PPC version).

In any case, a lot of factors went into the Final Decision.

In the latter 80s there were SAGE computers - intended for
the same niche as the IBM-PCs and LOOKED almost exactly
like them. They used the 68000 series though. Alas they,
like the Amiga's, used a proprietary OS that just was
not THAT much better and ultimately could not displace
MS/IBM or MAC. There were also production issues, one
developer BOLTED a board to his desk so they couldn't
sneak in and steal it for retail :-) SAGE still has a
sort of cult following, they WERE good computers. Alas
I could never afford, or likely GET, one.

VAST number of good chips/systems have come - and DIED -
since the late 70s. Some maybe WERE the Better Idea, but
IBM/Apple just overwhelmed them. There was SOME old chip
that had a few ultra-high-speed serial links to other
chips so you could implement almost unlimited multi-
processor/multi-processing ... just scale up forever.
Can't remember the name now - but they'd have been a
different, perhaps advantageous, way to approach the
issue.

On 11/02/2023 06:50, 26C.Z968 wrote:
> On 2/10/23 9:08 AM, Rich wrote:
>> 26C.Z968 <26C.Z968@noaada.net> wrote:
>>>    There was considerable debate at IBM over whether to use the M68000
>>>    series or the Intel chips back in the day.  They SHOULD have gone
>>>    with Motorola - but there were some reasons for the 8088.  First of
>>>    all it needed less wiring.  Secondly its instruction set was
>>>    similar enough to the Z80's so it was easy to port CP/M.  That was
>>>    their backwards-compatibility promise ...  at the time BIZ=CP/M and
>>>    IBM was BIZ.
>>
>> The reasoning given not long (as in early 'history' articles) after the
>> release of the IBM PC indicated a much simpler reasoning.
>>
>> Intel had an official licensed second source supplier for 808x chips
>> (that second source being AMD) and Motorola was, at least at the time
>> IBM's PC group was picking a CPU, single source (with a promise to have
>> a second source within X time, like a year or something).
>>
>> And IBM's corporate policy was to always have two sources for all
>> critical components.
>
>   The 8088 was just cheaper and "safer" in several respects,
>   so that's what they went with. CP/M compatibility WAS a
>   significant issue as well. CP/M-68k didn't come out for
>   another year or two (Radio Shack DID have a dual-board
>   computer that'd run CP/M-68k, once it arrived).
>
>> So while 'less wiring' and 'could run CP/M if need be' may have figured
>> into the choice, if these history articles were correct, the actual
>> reason was much simpler.
>
>   BIG company, LOTS of issues - real or imagined - LOTS
>   of meetings and meetings about meetings about .....
>
>   A terrible thing they DIDN'T go with the 68 series
>   though, it WAS a much better chip - which Apple soon
>   exploited. By now we'd have the 680990 chip and Intel
>   would still be making accessory hardware.
>
>> Wikipedia as of today
>> (https://en.wikipedia.org/wiki/IBM-PC#Design_process) cites that the
>> 68000 was not production ready at the time:
>>
>>      Several CPUs were considered, including the Texas Instruments
>>      TMS9900, Motorola 68000 and Intel 8088.  The 68000 was considered
>>      the best choice, but was not production-ready like the others.  The
>>      IBM 801 RISC processor was also considered, since it was
>>      considerably more powerful than the other options, but rejected due
>>      to the design constraint to use off-the-shelf parts.
>>
>>      IBM chose the 8088 over the similar but superior 8086 because Intel
>>      offered a better price for the former and could provide more
>>      units,[19] and the 8088's 8-bit bus reduced the cost of the rest of
>>      the computer.  The 8088 had the advantage that IBM already had
>>      familiarity with the 8085 from designing the IBM System/23
>>      Datamaster.  The 62-pin expansion bus slots were also designed to
>>      be similar to the Datamaster slots, and its keyboard design and
>>      layout became the Model F keyboard shipped with the PC, but
>>      otherwise the PC design differed in many ways.
>>
>
>   I have a little experience with the TMS-9000 series. AT THE
>   TIME it was a 'clever' chip - hardware-support for multi-user
>   environments. However it was big and expensive and TI was
>   super-anal - a reason the TI Home Computer died so soon. Also
>   the 9900's only easily supported 64kb "workspaces" for users
>   apps which would have been too limiting given the way the
>   future was developing.
>
>   I have no experience with the IBM-801 so I can't say much.
>   Their PowerPC chip probably evolved from it and was kinda
>   successful (most Linux's still have a PPC version).
>
>   In any case, a lot of factors went into the Final Decision.
>
>   In the latter 80s there were SAGE computers - intended for
>   the same niche as the IBM-PCs and LOOKED almost exactly
>   like them. They used the 68000 series though. Alas they,
>   like the Amiga's, used a proprietary OS that just was
>   not THAT much better and ultimately could not displace
>   MS/IBM or MAC. There were also production issues, one
>   developer BOLTED a board to his desk so they couldn't
>   sneak in and steal it for retail :-) SAGE still has a
>   sort of cult following, they WERE good computers. Alas
>   I could never afford, or likely GET, one.
>
>   VAST number of good chips/systems have come - and DIED -
>   since the late 70s. Some maybe WERE the Better Idea, but
>   IBM/Apple just overwhelmed them. There was SOME old chip
>   that had a few ultra-high-speed serial links to other
>   chips so you could implement almost unlimited multi-
>   processor/multi-processing ... just scale up forever.
>   Can't remember the name now - but they'd have been a
>   different, perhaps advantageous, way to approach the
>   issue.

In the end what wins is good enough, a bit cheaper, and readily available.
Just look at all the wives.

--
"First, find out who are the people you can not criticise. They are your
oppressors."
- George Orwell