[belenix-discuss] Some status updates

[Sriram Narayanan]

Hi everyone:

There are two updates from me for everyone.

a. More members for Belenix Development
Certain interested people have contacted Moinak and myself offlist and
have expressed an interest in having a standards compliant distro in
place.

I had a telephonic conversation with Moinak about this last night (He
had just returned from a family trip to someplace and was inaccessible
for some days). Moinak has said that he too considers this a
worthwhile effort.

I will not be forwarding the private mail thread to these discussion
lists since those discussions are private. However, I have requested
the participants in that thread to join belenix-dev and
belenix-discuss so that we can take the discussion ahead.

b. Belenix build efforts
On an exciting note, I've made some progress in the roadmap in terms
of build pipelines. I'll showcase this next weekend (and see how to
make a video and host it on youtube).
Some months ago, I'd conducted a session at the BOSUG meet at
Thoughtworks on how we could aim for continuous integration and build
pipelines for ON. It does seem feasible, and if this is in place,
we'll be able to replace components as and when we want (e.g. replace
some existing userland components with those form ATT's AST or from
the Hierloom project - something that had come up on illumos-discuss).

-- Sriram
Belenix: www.belenix.org
_______________________________________________
belenix-discuss mailing list
http://mail.opensolaris.org/mailman/listinfo/belenix-discuss
http://groups.google.com/group/belenix-discuss

[Lionel Cons]

Sounds good, except the correct spelling for Hierloom is actually
Heirloom :) and that I'd prefer to go with the AST toolchain first,
they have a proven good relationship and support with ON.

What are we doing about SPARC support?

Lionel

[Cedric Blancher]

I agree with Lionel on that.

> What are we doing about SPARC support?

Yes, what about SPARC support. Do we have any decent T1/T2 machines
which could used as build machines?

Ced
--
Cedric Blancher <cedric....@googlemail.com>
Institute Pasteur

[Joerg Schilling]

Sriram Narayanan <sri...@belenix.org> wrote:

> Some months ago, I'd conducted a session at the BOSUG meet at
> Thoughtworks on how we could aim for continuous integration and build
> pipelines for ON. It does seem feasible, and if this is in place,
> we'll be able to replace components as and when we want (e.g. replace
> some existing userland components with those form ATT's AST or from
> the Hierloom project - something that had come up on illumos-discuss).

Heirloom seems to be a dead project mostly based on OpenSolaris code.

There have only been very few changes during the past years and not even real
and important fixes added by Sun since 2005 have been applied. So be very
careful with that source. There is more activity in the "Schily" project, e.g.
with the Bourne Shell. I recently fixed all known bugs in the Bourne Shell that
have been documented since 1988.

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Lionel Cons]

On 15 May 2012 15:18, Joerg Schilling

<Joerg.S...@fokus.fraunhofer.de> wrote:
> Sriram Narayanan <sri...@belenix.org> wrote:
>
>> Some months ago, I'd conducted a session at the BOSUG meet at
>> Thoughtworks on how we could aim for continuous integration and build
>> pipelines for ON. It does seem feasible, and if this is in place,
>> we'll be able to replace components as and when we want (e.g. replace
>> some existing userland components with those form ATT's AST or from
>> the Hierloom project - something that had come up on illumos-discuss).
>
> Heirloom seems to be a dead project mostly based on OpenSolaris code.

That's not true. Heirloom has lots of interesting changes and features
which are not based on Opensolaris and were added later. Yes,
development has mostly ceased (same applies for SchillyX, which hadn't
any commits in 2012 and only a handful in 2011) but some of their
features are still worth taking.

> There have only been very few changes during the past years and not even real
> and important fixes added by Sun since 2005 have been applied. So be very
> careful with that source. There is more activity in the "Schily" project, e.g.
> with the Bourne Shell. I recently fixed all known bugs in the Bourne Shell that
> have been documented since 1988.

Seriously, who still wants the original Bourne shell?

Lionel

[Joerg Schilling]

Lionel Cons <lionelc...@googlemail.com> wrote:

> > Heirloom seems to be a dead project mostly based on OpenSolaris code.
>
> That's not true. Heirloom has lots of interesting changes and features
> which are not based on Opensolaris and were added later. Yes,

Most of the code was ported to Linux in a half hearted way that introduced bugs.
Few programs indeed have been enhanced (e.g. nail - en enhanced cversion of
mailx). But I reported e.g a bug in the imap implementation of nail that caused
nail to han in some cases with mails without Subject: and the author signalled
that he is not interested in working on a fix. For this reason, I recommend to
be very careful with decifing to use code from that site.

> development has mostly ceased (same applies for SchillyX, which hadn't
> any commits in 2012 and only a handful in 2011) but some of their
> features are still worth taking.

It seems that you missinterpret things. SchilliX-ON added 350000 lines of code.
Do you like to call this "a handful"?

There is more activity that you don't seem to notice.

> Seriously, who still wants the original Bourne shell?

It seems that you did not look at the Bourne Shell for a long time.

The problem with the korn shell is that it is huge and that it does not support
separate root & usr filesystems as it needs libraries that are under /usr.

Installing the Bourne Shell in /sbin/sh helpd to fix that problem.

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Lionel Cons]

On 16 May 2012 11:26, Joerg Schilling

<Joerg.S...@fokus.fraunhofer.de> wrote:
> Lionel Cons <lionelc...@googlemail.com> wrote:
>
>> > Heirloom seems to be a dead project mostly based on OpenSolaris code.
>>
>> That's not true. Heirloom has lots of interesting changes and features
>> which are not based on Opensolaris and were added later. Yes,
>
> Most of the code was ported to Linux in a half hearted way that introduced bugs.
> Few programs indeed have been enhanced (e.g. nail - en enhanced cversion of
> mailx). But I reported e.g a bug in the imap implementation of nail that caused
> nail to han in some cases with mails without Subject: and the author signalled
> that he is not interested in working on a fix.

I think this was more about a personal issue with your person. The
subject was that he won't take ANY patches from YOUR person anymore.

> For this reason, I recommend to
> be very careful with decifing to use code from that site.

Yeah, yeah. Now you turn the problems you have with the maintainer
into an argument against him. This is NOT civilised behaviour.
Also, Joerg, please grow up.

>> development has mostly ceased (same applies for SchillyX, which hadn't
>> any commits in 2012 and only a handful in 2011) but some of their
>> features are still worth taking.
>
> It seems that you missinterpret things. SchilliX-ON added 350000 lines of code.
> Do you like to call this "a handful"?

Yes, since development has almost stopped for SchillyX. Your not even
keeping track what illumos does.

> There is more activity that you don't seem to notice.
>
>> Seriously, who still wants the original Bourne shell?
>
> It seems that you did not look at the Bourne Shell for a long time.
>
> The problem with the korn shell is that it is huge and that it does not support
> separate root & usr filesystems as it needs libraries that are under /usr.

Joerg, this is not true. The libraries have to be moved to /lib,
that's all. Your enhanced bourne shell would have the same problem if
it would be dynamically linked.
Again, please grow up and stop making such promotional arguments.

> Installing the Bourne Shell in /sbin/sh helpd to fix that problem.

Does your Bourne shell conform to the POSIX shell standard. Please
answer 'yes' or 'no'. Only one of these two words.

Lionel

[Joerg Schilling]

Lionel Cons <lionelc...@googlemail.com> wrote:

> I think this was more about a personal issue with your person. The
> subject was that he won't take ANY patches from YOUR person anymore.

The person behind Heirloom does not look civilized as he atacked me several
times (e.g. when he claimed that he was the first person who created a portable
version of SCCS - but his code did not even work on Linux where he claimed to
have done tests). Maybe, he will grow up some time... It seems that after his
attacks, he stopped working on his sccs port while my sccs version has been
significantly enhanced, bug-fixed and gained performance.

> >> development has mostly ceased (same applies for SchillyX, which hadn't
> >> any commits in 2012 and only a handful in 2011) but some of their
> >> features are still worth taking.
> >
> > It seems that you missinterpret things. SchilliX-ON added 350000 lines of code.
> > Do you like to call this "a handful"?
>
> Yes, since development has almost stopped for SchillyX. Your not even
> keeping track what illumos does.

This is not true. From my information, there seem to me more activity than with
Belenix. But as you are wrong with Schillix, I may be wrong with Belenix...

The problem with Illumos is that there have been many changes that you really
don't like and that Illumos does not create changesets that have isolated
features only. This makes it hard to use Illumos as upstream. SchilliX-ON on
the other side is very careful with isolating changesets so people can cherry
pick features or fixes.

Illumos also does not care about SVr4 packages while SchilliX-ON supports both
package systems and enhanced the SVr4 packages by supporting direct
installation from the network.

I believe it is more important to work on the base (as I currently do) before
looking at features. I like to create a Solaris version that looks similar to
the previous Solaris Express and follows the Solaris ideas rather than Linux
ideas - by giving sufficient Linux compatibility to people that really like
that.

> > The problem with the korn shell is that it is huge and that it does not support
> > separate root & usr filesystems as it needs libraries that are under /usr.
>
> Joerg, this is not true. The libraries have to be moved to /lib,
> that's all. Your enhanced bourne shell would have the same problem if
> it would be dynamically linked.
> Again, please grow up and stop making such promotional arguments.

This is not true: ksh needs more than one huge library located in /usr while
the Bourne Shell is happy with the libs that are already in /lib.

> > Installing the Bourne Shell in /sbin/sh helpd to fix that problem.
>
> Does your Bourne shell conform to the POSIX shell standard. Please
> answer 'yes' or 'no'. Only one of these two words.

It seems that you did not understand the ideas behind the POSIX standard.

POSIX with good reason does not standardize on pathnames (except for
/dev/null). This allows a POSIX compliant system to put a POSIX compliant shell
wherever it likes in order to keep a traditional Bourne Shell in /bin/sh.
I however even plan to have ksh in /bin and the Bourne Shell in /sbin/sh
even though the current version of ksh93 is not fully POSIX compliant.
You see, your question is not helpful but just provoking.


BTW: The current version of the Bourne Shell see:

ftp://ftp.berlios.de/pub/cdrecord/alpha/

has a POSIX compliant "times" built-in command - ksh93's times gives a
non-compliant output.

There is also pushd/popd/dirs in the current Bourne Shell which is still
missing in ksh93.

But what intention do you have with your mail?

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Sriram Narayanan]

Thanks. I myself feel that we should have a separate thread on what
POSIX compliance means to us, and how we intend to ascertain that we
are POSIX compliant.

>
>> What are we doing about SPARC support?
>
> Yes, what about SPARC support. Do we have any decent T1/T2 machines
> which could used as build machines?

If:
a) anyone would like to contribute a zone or two on a SPARC box (if
not an entire SPARC)
b) we have people who are interested in SPARC packages and are willing
to test SPARC packages
c) we have people who are interested in maintaining SPARC packages,

then we can consider adding SPARC support too.

I myself can contribute by helping integrate the SPARC zones into our
overall build eco-system.

>
> Ced
> --
> Cedric Blancher <cedric....@googlemail.com>
> Institute Pasteur

--
Belenix: www.belenix.org

[Irek Szczesniak]

On Wed, May 16, 2012 at 12:33 PM, Joerg Schilling
<Joerg.S...@fokus.fraunhofer.de> wrote:
>> > The problem with the korn shell is that it is huge and that it does not support
>> > separate root & usr filesystems as it needs libraries that are under /usr.
>>
>> Joerg, this is not true. The libraries have to be moved to /lib,
>> that's all. Your enhanced bourne shell would have the same problem if
>> it would be dynamically linked.
>> Again, please grow up and stop making such promotional arguments.
>
> This is not true: ksh needs more than one huge library located in /usr while
> the Bourne Shell is happy with the libs that are already in /lib.

Joerg, please stop promoting your Joerg Bourne shell here. Thank you.
Also please stop changing wikipedia articles (e.g. "comparison of
command shells") for promotional purposes, too. Just saying that here
before you quote that article as "proof" for Bourne shell vs Korn
shell capabilities.

>> > Installing the Bourne Shell in /sbin/sh helpd to fix that problem.
>>
>> Does your Bourne shell conform to the POSIX shell standard. Please
>> answer 'yes' or 'no'. Only one of these two words.
>
> It seems that you did not understand the ideas behind the POSIX standard.
>
> POSIX with good reason does not standardize on pathnames (except for
> /dev/null). This allows a POSIX compliant system to put a POSIX compliant shell
> wherever it likes in order to keep a traditional Bourne Shell in /bin/sh.
> I however even plan to have ksh in /bin and the Bourne Shell in /sbin/sh
> even though the current version of ksh93 is not fully POSIX compliant.

Joerg, I ran the SUS test suite last week against ast-open.2012-05-04.
ksh93, part of this beta release, passed the tests. Could you please
send the bug to *this* list which makes ksh93 not fully POSIX
compliant? Thank you.

> There is also pushd/popd/dirs in the current Bourne Shell which is still
> missing in ksh93.

pushd/popd/dirs are NOT part of the original Bourne shell, these are
additions you made. Neither are these mandated by POSIX or SUS. I
still have the original SystemV sources on tape.

FYI, ksh93 has these as loadable plugins:
ls -l src/cmd/ksh93/fun/
total 12
-rwxr-xr-x 1 irek lab14 2259 May 10 2001 dirs
-rwxr-xr-x 1 irek lab14 2508 Nov 30 2001 popd
-rwxr-xr-x 1 irek lab14 2508 Nov 30 2001 pushd

Irek

[Joerg Schilling]

Irek Szczesniak <iszcz...@gmail.com> wrote:

> Joerg, please stop promoting your Joerg Bourne shell here. Thank you.
> Also please stop changing wikipedia articles (e.g. "comparison of
> command shells") for promotional purposes, too. Just saying that here
> before you quote that article as "proof" for Bourne shell vs Korn
> shell capabilities.

Some people live in a world that has been true in 1979 and these people don't
like information to be updated to even to the state of 1995.

Some people seem to believe that other people will believe them when they use
the term "promotional" for attempts to remove false claims.

I am not interested in these people, as I live in the present and as I don't
like false claims.

Given the fact that I added the command line history editor I wrote 1982..1984
for my "bsh" to the Bourne Shell in December 2006, it would not even be
promotional if I did add a related hint to the wiki article as we now have 2012.

> > It seems that you did not understand the ideas behind the POSIX standard.
> >
> > POSIX with good reason does not standardize on pathnames (except for
> > /dev/null). This allows a POSIX compliant system to put a POSIX compliant shell
> > wherever it likes in order to keep a traditional Bourne Shell in /bin/sh.
> > I however even plan to have ksh in /bin and the Bourne Shell in /sbin/sh
> > even though the current version of ksh93 is not fully POSIX compliant.
>
> Joerg, I ran the SUS test suite last week against ast-open.2012-05-04.
> ksh93, part of this beta release, passed the tests. Could you please
> send the bug to *this* list which makes ksh93 not fully POSIX
> compliant? Thank you.

Passing the SUS tests does not verify POSIX compliance. Look e.g. in special at
Mac OS X that passed the tests even though bash is the only half-way POSIX
shell that was definitely not POSIX compliant when the tests have been run.
Bash did not implement the -e option correctly (important for correct behavior
of make(1)) and bash most likely still does job-control on scripts that causes
nested make calls to continue in the background if you hit ^C.

> > There is also pushd/popd/dirs in the current Bourne Shell which is still
> > missing in ksh93.
>
> pushd/popd/dirs are NOT part of the original Bourne shell, these are
> additions you made. Neither are these mandated by POSIX or SUS. I
> still have the original SystemV sources on tape.

Correct, I implemented pushd/popd/dirs in spring 1985 for my "bsh" to make life
easier. I recently added the proven interfaces to the Bourne Shell too. This is
not to gain POSIX compliance but to achieve usability out of the box.

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Lionel Cons]

On 17 May 2012 15:10, Joerg Schilling

<Joerg.S...@fokus.fraunhofer.de> wrote:
> Irek Szczesniak <iszcz...@gmail.com> wrote:
>
>> Joerg, please stop promoting your Joerg Bourne shell here. Thank you.
>> Also please stop changing wikipedia articles (e.g. "comparison of
>> command shells") for promotional purposes, too. Just saying that here
>> before you quote that article as "proof" for Bourne shell vs Korn
>> shell capabilities.
>
> Some people live in a world that has been true in 1979 and these people don't
> like information to be updated to even to the state of 1995.
>
> Some people seem to believe that other people will believe them when they use
> the term "promotional" for attempts to remove false claims.
>
> I am not interested in these people, as I live in the present and as I don't
> like false claims.
>
> Given the fact that I added the command line history editor I wrote 1982..1984
> for my "bsh" to the Bourne Shell in December 2006, it would not even be
> promotional if I did add a related hint to the wiki article as we now have 2012.

*sigh*
You are diverting to other topics. See below.

>
>
>> > It seems that you did not understand the ideas behind the POSIX standard.
>> >
>> > POSIX with good reason does not standardize on pathnames (except for
>> > /dev/null). This allows a POSIX compliant system to put a POSIX compliant shell
>> > wherever it likes in order to keep a traditional Bourne Shell in /bin/sh.
>> > I however even plan to have ksh in /bin and the Bourne Shell in /sbin/sh
>> > even though the current version of ksh93 is not fully POSIX compliant.
>>
>> Joerg, I ran the SUS test suite last week against ast-open.2012-05-04.
>> ksh93, part of this beta release, passed the tests. Could you please
>> send the bug to *this* list which makes ksh93 not fully POSIX
>> compliant? Thank you.
>
> Passing the SUS tests does not verify POSIX compliance. Look e.g. in special at
> Mac OS X that passed the tests even though bash is the only half-way POSIX
> shell that was definitely not POSIX compliant when the tests have been run.
> Bash did not implement the -e option correctly (important for correct behavior
> of make(1)) and bash most likely still does job-control on scripts that causes
> nested make calls to continue in the background if you hit ^C.

Joerg, would you please come with with detailed proof that ksh (ksh93)
violates the POSIX standards. Please don't divert to other topics,
just present us the technical facts. bash, your Bourne shell, Mac OS X
or anything else does not matter in this discussion.

Lionel

[Sriram Narayanan]

No more discussion on this topic, for now.

Let us make progress one step at a time, and discuss compliance one command at a time.

Ram

belenix-dev mailing list
belen...@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/belenix-dev

[Joerg Schilling]

Lionel Cons <lionelc...@googlemail.com> wrote:

> > I am not interested in these people, as I live in the present and as I don't
> > like false claims.
> >
> > Given the fact that I added the command line history editor I wrote 1982..1984
> > for my "bsh" to the Bourne Shell in December 2006, it would not even be
> > promotional if I did add a related hint to the wiki article as we now have 2012.
>
> *sigh*
> You are diverting to other topics. See below.

It seems that you did not follow the discussion. I was just replying to a false
claim from Irek Szczesniak.

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Irek Szczesniak]

On Sun, May 20, 2012 at 10:08 PM, Joerg Schilling
<Joerg.S...@fokus.fraunhofer.de> wrote:
> Lionel Cons <lionelc...@googlemail.com> wrote:
>
>> > I am not interested in these people, as I live in the present and as I don't
>> > like false claims.
>> >
>> > Given the fact that I added the command line history editor I wrote 1982..1984
>> > for my "bsh" to the Bourne Shell in December 2006, it would not even be
>> > promotional if I did add a related hint to the wiki article as we now have 2012.
>>
>> *sigh*
>> You are diverting to other topics. See below.
>
> It seems that you did not follow the discussion. I was just replying to a false
> claim from Irek Szczesniak.

Joerg, this was not a false claim. The POSIX and SUS test suites are
very important and there is little to do if you do not pass it. My
mistake, by accidental omission, was to say that the written standard
and the standard interpretations done by Open Group and other standard
bodies have more importance (and precedence, before anything else).

Furthermore I have to say that I find your behavior very offensive and
aggressive (mainly your self promotion of SchillyX tools, which no one
is really using in real life). I think you have been told about that
by other standards boffins like Don Cragun (I think you will remember
him from incidents like
http://mail.opensolaris.org/pipermail/ogb-discuss/2007-August/001348.html)
or Andrew Josey (director of standards within the Open Group) that you
have to work on your attitude in email communication. Not that this
will change your behavior, but I still keep trying.

Irek

[Joerg Schilling]

Irek Szczesniak <iszcz...@gmail.com> wrote:

> On Sun, May 20, 2012 at 10:08 PM, Joerg Schilling
> <Joerg.S...@fokus.fraunhofer.de> wrote:
> > Lionel Cons <lionelc...@googlemail.com> wrote:
> >
> >> > I am not interested in these people, as I live in the present and as I don't
> >> > like false claims.
> >> >
> >> > Given the fact that I added the command line history editor I wrote 1982..1984
> >> > for my "bsh" to the Bourne Shell in December 2006, it would not even be
> >> > promotional if I did add a related hint to the wiki article as we now have 2012.
> >>
> >> *sigh*
> >> You are diverting to other topics. See below.
> >
> > It seems that you did not follow the discussion. I was just replying to a false
> > claim from Irek Szczesniak.
>
> Joerg, this was not a false claim. The POSIX and SUS test suites are
> very important and there is little to do if you do not pass it. My

I originally planned not to reply to this thread again, as asked by Sriram
Narayanan, but you again send incorrect claims that confuse people if not
corrected, so here is my hopefully last reply to this topic:

As mentioned before, you don't seem to understand POSIX, so let me try again to
help you:

- The Bourne Shell does not need to be POSIX compliant, so for the Bourne
Shell, the POSIX test suite is not important.

- POSIX intentionally does not require that /bin/sh is a POSIX shell.

- A POSIX shell is started by calling:

PATH=`getconf PATH`
sh

- The hacked ksh88 in /usr/xpg4/bin/sh is closer to the POSIX standard
than ksh93. Unfortunately, /usr/xpg4/bin/sh is not legally available in
source but the binary is redistributable. So as long as noone
introduces a new ISA, OpenSolaris distros can use that binary. With the
hope that ksh93 will cloaser approach POSIX in the future, ksh93 may be
discussed again soon.

- A failed POSIX compliance test verifies a POSIX non-compliance.
A passed POSIX compliance test does not verify POSIX compliance.

- The output from "times" is not the only non-compliance in ksh93. I see
deviations from time to time while doing tests in order to decide on
implementation details in the Bourne Shell or my "bsh", but I don't write
them down. This may however change in case the group of people that
promotes ksh93 for OpenSolaris will become more friendly.

> mistake, by accidental omission, was to say that the written standard
> and the standard interpretations done by Open Group and other standard
> bodies have more importance (and precedence, before anything else).

In case you don't know, I am actively working in the POSIX standard committee,
so I am one of the persons that decide on how to interpret the standard.

...
....offensive content censored....
...
.... and I am sorry to see that you again send a very offensive mail. Please
work on your discussion style.

And finally: in case you are not interested in writing things that a wider
audience is interesded in, you should better not reply to this thread in this
list again.

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Sriram Narayanan]

All:

I've said earlier on this thread that we can discuss things on a per
command basis. Let us please arrive at a common understanding of what
POSIX means to us.

While I have remained silent on this thread for the past two days, I
have also received numerous mails asking that we kickban Joerg from
Belenix-discuss, that we enforce moderation, people feeling
disappointed and wanting to leave, etc. I have chosen to wait a few
days and to then respond, since there is nothing to gain by losing
cool and clamping down.

Everyone needs to understand something important: the BOSUG and
Belenix culture is to work in a civil manner with everyone. We are
culturally very very different, and work on Belenix out of interest,
love for it, and respect for how Belenix has been the foundation of
the OpenSolaris distro.

Joerg: On the one hand, you write authoritatively about POSIX, and at
least Moinak and I like some of your tools. but on the other hand,
there are accusations about you trying to push in your tools, about
your lying about POSIX standards, etc. While I personally welcome your
advice, there are others on this group (infact this very thread), who
are skeptical about your statements. My request: Please let me set up
the right discussion threads, and then provide inputs. I will
arbitrate and ensure that the discussion is civil, but would also like
us all to collectively weigh and understand what we need and which
tools suit the groups' needs.

Irek, and others who have opinions different from Joerg: Please be
patient. We are not taking the final call on anything POSIX.

I personally am now very very curious as to why there is so much of
debate over the years, when the POSIX compliance tests are so easily
available (accompanied with some different levels of understanding of
how much these cost).

It seems like we need a project manager to facilitate discussions -
I'll set up something around this by mid next week.

-- Ram

[Joerg Schilling]

Sriram Narayanan <srir...@gmail.com> wrote:

> Joerg: On the one hand, you write authoritatively about POSIX, and at
> least Moinak and I like some of your tools. but on the other hand,
> there are accusations about you trying to push in your tools, about
> your lying about POSIX standards, etc. While I personally welcome your

The only person, who could be identified as lying about POSIX is Irek.

If you have questions about the POSIX standard, please ask...

If you like to have civil discussions, please find a way to restrict those who
attack others.

Some people in this list have a very offensive discussion style. I am not sure
whether this is something that is inherent to people with english as
mothertongue or whether this is specific to some persons. Fortunately, the
indian people I know from the OpenSolaris core developer meetings seem to be
very friendly.

Back to "my" tools:

- Sun made a definitve decision to include Star in Solaris in 2004 and
only the make file framework for integration was missing. This is the
past now since 2 years, as I wrote the makefiles for the OpenSolaris
build framework (after it turned out that Roland Mainz who promised to
write them wil never do it), so let us stop to discuss this - star
integration was done.

- Most other software that is discussed here (that I am working on) is
Solaris sources that I enhanced over the years. At the same time,
other people (such as Irek) try to promote software they are working
on and that is not originally from Solaris.

- Illumos added a od(1) replacement from Garret even though it has known
proven bugs. It is your decision whether you like to have a slow
and buggy od(1) from Garret or whether you like a POSIX compliant,
SVID compliant and faster od(1) from me.

I hope that there is a way to have tecnically based discussons in the future
and that people who usually pop up when they believe there is something they
can use for a dispute will stay elsewhere.

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Sriram Narayanan]

Please Joerg, no more updates on this mail thread.

No more updates on this thread, anyone.

I'm trying hard to not crack down on anyone, nor invoke moderation. There are just too many personal attacks on this list now, and enough is enough.

Ram

[Joerg Schilling]

Sriram Narayanan <srir...@gmail.com> wrote:

> Please Joerg, no more updates on this mail thread.
>
> No more updates on this thread, anyone.
>
> I'm trying hard to not crack down on anyone, nor invoke moderation. There
> are just too many personal attacks on this list now, and enough is enough.

If you did not claim that I was lying, my last mail had not been written.

So try to avoid writing things that look ofensive if you like this thread to
calm down.

Jörg

--
EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
j...@cs.tu-berlin.de (uni)
joerg.s...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily

[Sriram Narayanan]

Joerg, my second public request to you - please stop responding to this thread.

Let's discuss offline. I have not called you a liar. I see a communication issue here. I will take this offline with you.

Ram