RFC: QNX4 vs. QNX6

Kris Warkentin wrote:

Dave Gibbs put in a request earlier for ‘prefix’ which would show which
paths are owned by which resmgrs.

Since “sin na” no longer applies, “pidin na” could be “overloaded” to
mean “show owners of NAmespaces”.
Rennie

Kris Warkentin wrote:

You prefer CASD to CAD?

I think CASD is better, since “windohz peeple” (yes I am reserving that
name for my rock band :wink: expect CAD to bring up a task manager thingy
(not directly initiate a shutdown).

Rennie

I miss these very much:

sin (fd, fi, rt, ir)
ditto
modem

Thanks, Pavol Kycina


“Kris Warkentin” <kewarken@qnx.com> wrote in message
news:aedho6$q6o$1@nntp.qnx.com

Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced functionality
to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

cheers,

Kris

<andy@microstep-mis.com> wrote in message
news:aejs40$6oh$1@charon.microstep-mis.sk

[…]

my proposal: vedit or vp

For those of us not brought up on a diet of vi/jed/emacs etc, and who mourn
the loss of vedit/vplus, there may now be a suitable replacement. Check out
Workspace at pages.infinit.net/micbel. I have not used it in anger as yet,
but it seems to have all the facilities an ex-vplus user like me would ever
need.:sunglasses:

Jim

What about the situation when there are multiple OS partitions? I’m just
thinking that the timeout for the choice is quite short and it’s
considerably more inconvenient to have to use fdisk to set the active
partition every time than to just pick one or let it time out. There’s
nothing to stop you from changing the active partition to have a different
default. I guess I’m not really seeing why having the choice is a bad
thing. Perhaps an ‘expert mode’ or something to get this behaviour?

Kris

“Jens H Jorgensen” <jhj@remove-nospam-videk.com> wrote in message
news:aedk2b$o5j$1@inn.qnx.com

Here is general request for QNX boot manager:

If you have multiple partions with a valid QNX installation (.diskboot in
the root) the QNX boot manager comes out and ask which partion to use as
root partition. I would actually prefer if the QNX boot manager (or
certain
version of the boot manager) would simply use the partion which is marked
as
the active boot partion as the root partion, and not offer up the choice
to
the user.

The reason: in a multi partion setup an expect user can control which
partion to boot from and the novice user is not posed with the partion
selection question every time the system is booted.

Jens

“Kris Warkentin” <> kewarken@qnx.com> > wrote in message
news:aedho6$q6o$> 1@nntp.qnx.com> …
Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced
functionality
to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

cheers,

Kris
\

Dave Gibbs put in a request earlier for ‘prefix’ which would show which
paths are owned by which resmgrs.

cheers,

Kris

<kabe@sra-tohoku.co.jp> wrote in message news:aeif0t$6s4$1@inn.qnx.com

Was there any tool which shows
which resmgr is attaching to which /path/ ?

Especially when multiple resmgr is attaching to the same path
you can’t easily guess which resmgr is called first;
only procnto knows for sure.

(and digging through /proc/mount doesn’t seem the correct way)

kabe

I’m assuming that this gives you some way to connect to the machine via
modem. Is ‘modem’ a utility of some sort? What would talk to it on the
other end? Does it allow tcpip traffic or something else?

Kris

“Rodney Stevens” <rodney.stevens@minerals.csiro.au> wrote in message
news:aejv4e$a2o$1@inn.qnx.com

Ok Kris

What about
tinit -c "modem " -t /dev/serx etc

Use to work great on 4.25 a real pain on 6.x. Can I assume that you were
expecting everyone to use pppd/chat for remote modem connection.

Rod


“Kris Warkentin” <> kewarken@qnx.com> > wrote in message
news:aedho6$q6o$> 1@nntp.qnx.com> …
Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced
functionality
to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

cheers,

Kris
\

Here’s a follow-up question.

Do you consider it generally acceptable to have things on the third party
cd? Most of the suggestions are for things we would include with the core
but some things, (ie send/expect) would be taken from public sources. The
reason I ask is that there are issues with including public utils in the
core offering:

  1. Licensing: If it’s BSD style, fine but we still have to show the license
    on installation. If it’s GPL, there’s a whole can of worms because the
    customer doesn’t want to get ambushed by accidentally including a GNU
    utility in their shipped image.

  2. Support: Like it or not, if we include a 3rd party util with the core,
    we’re implicitly bound to support it to a certain extent. DDD is a good
    example of this problem - I spent many an hour chasing issues in it even
    though we weren’t technically supporting it.

What I’m saying is that it’s much easier to get things done on the third
party cd than the main offering but if this is okay with everybody then we
can go ahead and follow that route.

cheers,

Kris

“Kris Warkentin” <kewarken@qnx.com> wrote in message
news:aedho6$q6o$1@nntp.qnx.com

Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced functionality
to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

cheers,

Kris

It is true that it is more convenient for a user that knows what he/she is
doing to be able to choose root partition at boot time. But the problem is
that you cannot change the default root partition currently, so if a
non-expect user has to use a certain root partition then that user has to
press the correct key, which can be a big challenge for some users.

The advantage with doing it through fdisk is that an expert user can setup
which partition to use as root partition, and then the novice user is not
presented with the choice after that.

The current mode is basically an expert mode for the boot strap code, so if
there was a “use active partition” version of the boot strap code that would
be very nice.

Thanks
Jens



“Kris Warkentin” <kewarken@qnx.com> wrote in message
news:aekqsk$6ru$1@nntp.qnx.com

What about the situation when there are multiple OS partitions? I’m just
thinking that the timeout for the choice is quite short and it’s
considerably more inconvenient to have to use fdisk to set the active
partition every time than to just pick one or let it time out. There’s
nothing to stop you from changing the active partition to have a different
default. I guess I’m not really seeing why having the choice is a bad
thing. Perhaps an ‘expert mode’ or something to get this behaviour?

Kris

“Jens H Jorgensen” <> jhj@remove-nospam-videk.com> > wrote in message
news:aedk2b$o5j$> 1@inn.qnx.com> …
Here is general request for QNX boot manager:

If you have multiple partions with a valid QNX installation (.diskboot
in
the root) the QNX boot manager comes out and ask which partion to use as
root partition. I would actually prefer if the QNX boot manager (or
certain
version of the boot manager) would simply use the partion which is
marked
as
the active boot partion as the root partion, and not offer up the choice
to
the user.

The reason: in a multi partion setup an expect user can control which
partion to boot from and the novice user is not posed with the partion
selection question every time the system is booted.

Jens

“Kris Warkentin” <> kewarken@qnx.com> > wrote in message
news:aedho6$q6o$> 1@nntp.qnx.com> …
Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but
don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced
functionality
to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

cheers,

Kris


\

Kris Warkentin <kewarken@qnx.com> wrote:

What about the situation when there are multiple OS partitions? I’m just
thinking that the timeout for the choice is quite short and it’s
considerably more inconvenient to have to use fdisk to set the active
partition every time than to just pick one or let it time out. There’s
nothing to stop you from changing the active partition to have a different
default. I guess I’m not really seeing why having the choice is a bad
thing. Perhaps an ‘expert mode’ or something to get this behaviour?

I think the “friendly” way is have “default selection” with, say,
15 seconds timeout. (Just like you boot from cdrom).

The “default selection” is the “just installed root, fs package”,
or “last choice”.

This way, a power-reset on this kind of installation will always
bring you back where it used to be, without human attanding…

-xtang


Kris

“Jens H Jorgensen” <> jhj@remove-nospam-videk.com> > wrote in message
news:aedk2b$o5j$> 1@inn.qnx.com> …
Here is general request for QNX boot manager:

If you have multiple partions with a valid QNX installation (.diskboot in
the root) the QNX boot manager comes out and ask which partion to use as
root partition. I would actually prefer if the QNX boot manager (or
certain
version of the boot manager) would simply use the partion which is marked
as
the active boot partion as the root partion, and not offer up the choice
to
the user.

The reason: in a multi partion setup an expect user can control which
partion to boot from and the novice user is not posed with the partion
selection question every time the system is booted.

Jens

“Kris Warkentin” <> kewarken@qnx.com> > wrote in message
news:aedho6$q6o$> 1@nntp.qnx.com> …
Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced
functionality
to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

cheers,

Kris
\

Armin Steinhoff <a-steinhoff@web_.de> wrote:


Kris Warkentin wrote:

Attention QNX users.

I am compiling a wish-list for utilities development over the next 6 months
to a year. I’m mostly looking for things that existed in QNX4 but don’t in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced functionality to
existing utilities, now is the time to let me know so I can try to get them
on the road map.

Implement a distributed file system for PVM or MPI/RT (still to
implement)

I’ve done MPICH a long time ago (on QNX4). What I am more interested in
is a QNET based, “made by QNX”, distributed system. (With things
like remote spawn, remote fork, job scheduling …)

-xtang

Armin



cheers,

Kris

Okay. I’ll put it on the list.

Thank you.

Kris

“Jens H Jorgensen” <jhj@remove-nospam-videk.com> wrote in message
news:aeku4r$3ar$1@inn.qnx.com

It is true that it is more convenient for a user that knows what he/she is
doing to be able to choose root partition at boot time. But the problem is
that you cannot change the default root partition currently, so if a
non-expect user has to use a certain root partition then that user has to
press the correct key, which can be a big challenge for some users.

The advantage with doing it through fdisk is that an expert user can setup
which partition to use as root partition, and then the novice user is not
presented with the choice after that.

The current mode is basically an expert mode for the boot strap code, so
if
there was a “use active partition” version of the boot strap code that
would
be very nice.

Thanks
Jens



“Kris Warkentin” <> kewarken@qnx.com> > wrote in message
news:aekqsk$6ru$> 1@nntp.qnx.com> …
What about the situation when there are multiple OS partitions? I’m
just
thinking that the timeout for the choice is quite short and it’s
considerably more inconvenient to have to use fdisk to set the active
partition every time than to just pick one or let it time out. There’s
nothing to stop you from changing the active partition to have a
different
default. I guess I’m not really seeing why having the choice is a bad
thing. Perhaps an ‘expert mode’ or something to get this behaviour?

Kris

“Jens H Jorgensen” <> jhj@remove-nospam-videk.com> > wrote in message
news:aedk2b$o5j$> 1@inn.qnx.com> …
Here is general request for QNX boot manager:

If you have multiple partions with a valid QNX installation (.diskboot
in
the root) the QNX boot manager comes out and ask which partion to use
as
root partition. I would actually prefer if the QNX boot manager (or
certain
version of the boot manager) would simply use the partion which is
marked
as
the active boot partion as the root partion, and not offer up the
choice
to
the user.

The reason: in a multi partion setup an expect user can control which
partion to boot from and the novice user is not posed with the partion
selection question every time the system is booted.

Jens

“Kris Warkentin” <> kewarken@qnx.com> > wrote in message
news:aedho6$q6o$> 1@nntp.qnx.com> …
Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but
don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced
functionality
to
existing utilities, now is the time to let me know so I can try to
get
them
on the road map.

cheers,

Kris




\

Three cheers for ditto! (My vote on this too).

Kevin

“Kris Warkentin” <kewarken@qnx.com> wrote in message
news:aegvs5$bf6$1@nntp.qnx.com

Great list of suggestions. Comments below.
camz@passageway.com> > wrote in message news:aeeqh7$kho$> 1@inn.qnx.com> …
Kris Warkentin <> kewarken@qnx.com> > wrote:
So, if there are any utilities you wish you had or enhanced
functionality to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

You can add my vote for “sin fd” and “sin irq” to the list. I’d prefer
to
see
all the functionality of qnx4’s sin appear in qnx6’s pidin. It is just
too
confusing to have pidin AND sin in qnx6, especially when neither of them
do all
the things we expect of them.

Noted.

ditto is pretty high up on the list. It’s essential for maintaining
remote QNX
servers (such as QNXZone), especially when updating).

Not sure who would be looking at this. I’ll check up on it.

sendmail needs to be ported, and the patches submitted back to the
authors. I
know this is somewhat in the 3rd party area, but EVERY unix OS ships
with a
working sendmail.

I believe this is on the third party CD. I don’t know if there will be
any
movement of packages from that distribution to the main one. Probably if
there was enough demand…

A big one, which would involve changes at the API level as well as the
utility
level would be PAM. I’d also like to see the source for “login”,
“passwd”, and
“su” appear in the public CVS repository. Those utilities are
“protected”
by
the requirement of needing root-privs to run, and the encryption that is
used.
Revealing their source does not create any security breach. If the
source
for
those two utilities the user community would have been able to add
support
for
thing like NIS/YellowPages, and PAM.
The PAM architechture is my main request though, since I beleive it
would
allow
for the implementation of things like NIS/YellowPages as well even if
QSSL
is
uncomfortable with releasing the source for the utilities I mentioned.
In a similiar vein, the source for phlogin would also be useful, there
are
many
instances where a developer needs to customize a login environment.

These are things that will be considered because part of the activity over
the next while will involve deciding which utils to release source for.
One
of the barriers to getting self hosted QNX6 accepted into corporate
environments is the ability to play nice with corporate infrastructure. I
imaging that NIS/YP would help that.

I would also like to see IBM’s MQ Series available as a runtime for
QNX6.
I
know it’s out there, but damned if I can find it on their site.

Don’t know anything about this. I’ll look into it.

send/expect would also be useful for scripting.

If this isn’t on the third party CD, we could probably make sure that it
is.
I think some of our testers use expect internally so it could be that it’s
already ported.

typescript is on my list as well, and would also be welcomed by QSSL’s
customers at the various educational institutions (univerisites and such
that
use Unix in their courses often require students to compile and run
their
programs within a typescript session and the submit that file along with
their
assignment.

I’ve used the ‘script’ command for quite a while on QNX6. It seems to
work
fine.

I can also give you a list of issues/features with fs-pkg, but that
belongs in
another thread.

If you’d like to start a thread, I’m sure someone will be listening.
We’re
all working on our long and short term goals.

logical disk volume management would be nice too, so that you could
overlay
something like fs-lvm.so on top of existing fs layouts to “merge”
multiple
disks into one large logical disk. HP-UX and AIX have excellent
implementations
of this sort of thing for you to examine.

Some things probably won’t make a roadmap simply because our emphasis is
really on the single developer targetting an embedded board rather than a
server OS. Stuff like LVM is definitely more useful on big iron.

I’m sure I’ll think of more over time, I’ll let you know as I think of
them.

Great. Keep them coming.

cheers,

Kris

Jim Douglas wrote:

My wish is that you would fix QNet to be as good as the QNX4 native
networking -

  1. Make it reliable. At present there are too many errors and loss of
    service.

  2. Make it work over a serial link WITHOUT having to use TCP/IP and PPP as
    well.

  3. Make the ndp reliable and predictable and make it work over any type of
    link. At present it does not work at all over a serial link.

  4. Come up with a name location scheme (or similar) that is as good as, or
    better than nameloc so that it is technically possible to configure a
    logical system without regard for the physical network it runs on.

This is just an outline - if you want me to go into more detail just ask…

Hmm … what you are requesting is a cluster middleware. User PVM :slight_smile:

Armin


Jim

“Kris Warkentin” <> kewarken@qnx.com> > wrote in message
news:aedho6$q6o$> 1@nntp.qnx.com> …
Attention QNX users.

I am compiling a wish-list for utilities development over the next 6
months
to a year. I’m mostly looking for things that existed in QNX4 but don’t
in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced functionality
to
existing utilities, now is the time to let me know so I can try to get
them
on the road map.

cheers,

Kris

Kris Warkentin wrote:

Attention QNX users.

I am compiling a wish-list for utilities development over the next 6 months
to a year. I’m mostly looking for things that existed in QNX4 but don’t in
QNX6 but I’m also taking general utils feature requests.

So, if there are any utilities you wish you had or enhanced functionality to
existing utilities, now is the time to let me know so I can try to get them
on the road map.

Implement a distributed file system for PVM or MPI/RT (still to
implement)

Armin


cheers,

Kris

Ah…good idea.

Kris
“Rennie Allen” <rallen@csical.com> wrote in message
news:3D0DA255.9030605@csical.com

Kris Warkentin wrote:
Dave Gibbs put in a request earlier for ‘prefix’ which would show which
paths are owned by which resmgrs.


Since “sin na” no longer applies, “pidin na” could be “overloaded” to
mean “show owners of NAmespaces”.
Rennie

Paul D. Smith <pausmith@nortelnetworks.com> wrote:

%% “Kris Warkentin” <> kewarken@qnx.com> > writes:

nkw> We actually rolled a native ppc gdb for someone (was it you?) at
nkw> one point so it’s quite possible that native debugging could
nkw> become a reality.

Yes, it was us, but the binary we received wouldn’t even start; runtime
linker rejected it: incorrect C library I think we discovered it was.
We were using 6.1.0a at the time and I believe the binary must have been
compiled against the 6.2 prerelease tree or something.

Hi. I did the ntoppc native GDB. I was told some time ago that the binary
failed at startup, and I had suggested that you try re-building it on your
system, using the target files (libc.so, etc…) for your ntoppc target.
I never heard back, so I thought no news was good news. Apparently that
is not the case.

There is a script I had given you with the source (“build-ntoppc” or
something like that). Did you try that? I think a re-build in your
environment should fix this.

Regards,
GP


Native debugging will be nice and now that 6.2 is out we’ll try harder
to get it. But mainly I’m interested in getting the new stuff in GDB
5.x available, and getting shared library debugging to work in some sort
of sane fashion on QNX.

nkw> Certainly if our changes are applied to the public GNU source
nkw> then people could roll their own. It may not be completely
nkw> desireable in the future though since the Eclipse/qconn/pdebug
nkw> combination is becoming the “Ultimate Debugging Machine…dum dum
nkw> dum…™” and just using gdb by itself may be too limiting.

I’m not interested in running Eclipse, until/unless you support it on
Linux. And maybe not even then. I don’t have any Windows or Intel/QNX
systems (or SCM toolset is ClearCase so we’ll never move to hosted
QNX–until and unless ClearCase supports it–and maybe not even then > :slight_smile:> ).
And, bulky Java/GUI toolkits like this just run like crap on
Solaris/SPARC, in my experience.

I’ve been using Emacs with GDB et.al. as my IDE for over 14 years and
I’m perfectly happy with that… when it works right > :slight_smile:> .

nkw> Note that the toolchain that we ship and any submissions to the
nkw> GNU people are completely unrelated. Getting our support in the
nkw> public tree is to benefit those who want to use the latest and
nkw> greatest and also to avoid future back-porting headaches.
nkw> Obviously our shipped toolchain must undergo intense scrutiny and
nkw> the decision to move to a new version is not a trivial one.

Ditto for us: once we get to a stable point we’ll almost certainly stick
with whatever compiler it’s using for quite a while. Just for starters,
we always build with -Werror and any new version of compiler entails
bunches of work fixing the new set of warnings > :slight_smile:> . Plus, obviously,
complete regression testing to be sure the compiler didn’t invent any
new bugs. The same, of course, for QNX itself.



Thanks for the reply, Kris…

Paul D. Smith <> pausmith@nortelnetworks.com> > HASMAT–HA Software Mthds & Tools
“Please remain calm…I may be mad, but I am a professional.” --Mad Scientist

These are my opinions—Nortel Networks takes no responsibility for them.

Kris Warkentin <kewarken@qnx.com> wrote:

I’ve got “who” in our CVS tree. I also got a few things fixed up in tinit
and telnetd (login was already good) to make sure that utmp and wtmp are
kept up to date properly. These are mostly just a matter of getting it
packed and shipped. The other thing would be to wire some utmp/wtmp
accounting into pterms and phlogin. That would be something for the
development roadmap.

The problem with utmp/wtmp is that although they are supported, they are
NOT enabled by default. The utils that update them are written so that if
they are not enabled, then they do not get created. The reason beihnd this
is solid and quite valid, these files can get large after a while, and
this would be undesirable for an embedded system. Since the whole point of
QNX is to build embedded systems, it seems wrong to stray from this.

Someone else (cdm maybe?) had written a who that figured it out by some
other mechanism but, since all the wtmp/utmp stuff is in place and is the
standard unix way, it seemed best to do it this way.

That would be me. :slight_smile:
See my above note on why wtmp/utmp isn’t the best way for QNX.

One thought: if
whoever has the non-utmp who would give me the code, I could wire it into
who as a fallback if utmp/wtmp aren’t there.

I would even go so far as to give the ability to ignore utmp/wtmp as an
option on the command line. Every unix/linux system that I have used is
notoriously bad in its ability to maintain an accurate utmp/wtmp. After
a while you ALWAYS have garbage in there, and who winds up showing more
users logged in than there really are. The only solution it to truncate
the files, and then until all the currently online users logout, the who
output is STILL wrong, since it does not show those users (since they no
longer have an entry in utmp/wtmp).

The who detection algorithm that I am using will occasionally show users
that are not there, but when it does… it ALWAYS represents a stray /
orphaned process that IS there and needs to be cleaned up.

Cheers,
Camz.

Kris Warkentin <kewarken@qnx.com> wrote:

So, if there are any utilities you wish you had or enhanced functionality to
existing utilities, now is the time to let me know so I can try to get them
on the road map.

BIG ONE that I almost forgot.

Please, please, please, for the sanity of all developers everywhere, add
support for the keychord CTRL-ALT-SHIFT-DELETE.

This is one of the major annoyances with Qnx6 when developing an embedded
system.

Cheers,
Camz.

Kris Warkentin <kewarken@qnx.com> wrote:

ditto is pretty high up on the list. It’s essential for maintaining
remote QNX servers (such as QNXZone), especially when updating).

Not sure who would be looking at this. I’ll check up on it.

Last I heard, the delay was waiting for io-char or something like that, some
chunk of architecture that wouldimpact other utils and possibly some of the
devc-* drivers. (Hope that helps you check up on it).

sendmail needs to be ported, and the patches submitted back to the
authors. I know this is somewhat in the 3rd party area, but EVERY
unix OS ships with a working sendmail.

I believe this is on the third party CD. I don’t know if there will be any
movement of packages from that distribution to the main one. Probably if
there was enough demand…

It’s table stakes. Qnx6 doesn’t look like a serious Unix without it, doesn’t
matter if anyone uses it. Personally, I think that QSSL should provide some
active support / ownership for sendmail working on QNX, which places it firmly
on the core CD, not the 3rd Party CD. I don’t expect QSSL to provide support
to it’s customers on how to configure or use it beyond the supplied example
sendmail.cf files.

These are things that will be considered because part of the activity over
the next while will involve deciding which utils to release source for. One
of the barriers to getting self hosted QNX6 accepted into corporate
environments is the ability to play nice with corporate infrastructure. I
imaging that NIS/YP would help that.

Indeed. Providing the PAM architechture would be the ideal way to accomplish
this too, since it would also allow for an unsupported corporate authentication
infrastruture to be built by a customer.


typescript is on my list as well, and would also be welcomed by QSSL’s

I’ve used the ‘script’ command for quite a while on QNX6. It seems to work
fine.

I didn’t know it even existed. There are so many thing in QNX6 that have
just “appeared” and NEVER made it into a release document, sometimes that
makes it a real bear to actually find this stuff, let alone know what has
changed. A good example of this is the cmd-line installer that is now in
6.2, you will note that it isn’t mentioned anywhere.

Cheers,
Camz.