Re: smbserver

41 views
Skip to first unread message
Message has been deleted

David Pitt

unread,
Apr 7, 2007, 6:29:02 AM4/7/07
to
In message <4ecfbf4b2einval...@invalid-domain.co.uk>
Paul Vigay <invalid-em...@invalid-domain.co.uk> wrote:

>In article <slrnf1elh...@compsoc.dur.ac.uk>,
> Herbert zur Nedden <HzN....@HQ.gag.de> wrote:
>> In agreement with Thomas I supply his version 0.08 of !smbserver
>> including source code on <URL:http://www.gag.de/software/smbserver> and
>> it will remain available there until UPP does supply a better version.
>
>Is this different to the version 0.08 of SmbServer available at
>http://riscossmbserver.sourceforge.net/
>
There is clearly something different, the version from sourceforge
works. The version on Herbert's site stiffs my Iyonix everytime. I shall
be interested to see if the pre-Milius version is more stable in the
longer term on the Iyonix. I am sure I have used it in the past and it
did eventually upset something, MPro?. Trying again then.

--
David Pitt.

Computing with RISC OS.

Thomas Milius

unread,
Apr 7, 2007, 8:48:29 AM4/7/07
to
In message <2f1dc2cf4...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

My Samba version is working fine on my Iyonix else I won't have
published a while ago. The original version from sourceforge
has a fatal definition bug which crashed my machine every 20 minutes
after recording some DVB-T video if setting TX/DX size to 16384.
There are some other not so heavy bugs too.

So it is interesting to hear that my version shall stiff an Iyonix.
I shall download it from Herberts side to ensure that I did not
provide something wrong. However it is linked against Castles
C-library. Does this might cause trouble? My RISC OS Version is 5.13.
But I did not had any problems on 5.11/5.12 and earlier too.

Regards

Thomas Milius

Thomas Milius

unread,
Apr 7, 2007, 9:13:02 AM4/7/07
to
In message <70e1cecf...@thomas-milius.t-online.de>
Thomas Milius <Thomas...@t-online.de> wrote:

Just downloaded. The setup seems odd. I don't know why.
I am personally using a bit more modern version which is
compiled for XScale only but that doesn't make a difference.

"stiff": I am using !DNSServer from Justin Fletcher (yes there
is a Public Domain DNS-server for the Iyonix ...) which was written
1997 in BASIC. As far as I renember sometimes the Samba server
tries to make usage of such a server (depends on device
connected). If this would be the case the "stiff" should
disappear after lets say 40 seconds or so. However I am ever
starting both programs together since two years so wasn't aware of
it.

Unfortunately I don't know whether !DNSServer is still public
available. I took it from an old CD as far as I renember.
!DNSServer is "... This application is Public
Domain Freeware. This means that it may be distributed, so long as no charge
other than copying costs are charged for it. The copyright remains with
myself (Justin Fletcher).". So I think that I can sent you a copy if
required. Would be glad if soneone could give a hint to the
original download link.

Regards

Thomas Milius

Thomas Milius

unread,
Apr 7, 2007, 9:41:43 AM4/7/07
to
In message <0321d1cf...@thomas-milius.t-online.de>
Thomas Milius <Thomas...@t-online.de> wrote:

> In message <70e1cecf...@thomas-milius.t-online.de>
> Thomas Milius <Thomas...@t-online.de> wrote:
>
> > In message <2f1dc2cf4...@pittdj.plus.com>
> > David Pitt <ne...@pittdj.co.uk> wrote:
> >

> ...


> > > There is clearly something different, the version from sourceforge
> > > works. The version on Herbert's site stiffs my Iyonix everytime. I shall
> > > be interested to see if the pre-Milius version is more stable in the
> > > longer term on the Iyonix. I am sure I have used it in the past and it
> > > did eventually upset something, MPro?. Trying again then.
> > >
> >

> > ...


>
> Just downloaded. The setup seems odd. I don't know why.

Inside the !smbserver.Resources.params file netbios name
must be set to netbios name,TRUE.

> I am personally using a bit more modern version which is
> compiled for XScale only but that doesn't make a difference.
>
> "stiff": I am using !DNSServer from Justin Fletcher (yes there
> is a Public Domain DNS-server for the Iyonix ...) which was written
> 1997 in BASIC. As far as I renember sometimes the Samba server
> tries to make usage of such a server (depends on device
> connected). If this would be the case the "stiff" should
> disappear after lets say 40 seconds or so. However I am ever
> starting both programs together since two years so wasn't aware of
> it.
>
> Unfortunately I don't know whether !DNSServer is still public available. I
> took it from an old CD as far as I renember. !DNSServer is "... This
> application is Public Domain Freeware. This means that it may be
> distributed, so long as no charge other than copying costs are charged for
> it. The copyright remains with myself (Justin Fletcher).". So I think that
> I can sent you a copy if required. Would be glad if soneone could give a
> hint to the original download link.

The !smbserver.Resources.smb/conf file does not contain a
"nameserver enabled" entry which seems to lead to an enabling.
Either it must be disabled manually in the configuration
or the entry

nameserver enabled = No

must be added to the

# Global parameters

section at the end or !DNSSserver must be installed.

Regards

Thomas Milius

David Pitt

unread,
Apr 7, 2007, 10:55:34 AM4/7/07
to
In message <09c1d3cf...@thomas-milius.t-online.de>
Thomas Milius <Thomas...@t-online.de> wrote:

>In message <0321d1cf...@thomas-milius.t-online.de>
> Thomas Milius <Thomas...@t-online.de> wrote:

>> In message <70e1cecf...@thomas-milius.t-online.de>
>> Thomas Milius <Thomas...@t-online.de> wrote:

>> > In message <2f1dc2cf4...@pittdj.plus.com>
>> > David Pitt <ne...@pittdj.co.uk> wrote:

>> > > There is clearly something different, the version from
>> > > sourceforge works. The version on Herbert's site stiffs my Iyonix
>> > > everytime.

[snip]

>> Just downloaded. The setup seems odd. I don't know why.
>
>Inside the !smbserver.Resources.params file netbios name
>must be set to netbios name,TRUE.
>
>> I am personally using a bit more modern version which is
>> compiled for XScale only but that doesn't make a difference.
>>
>> "stiff": I am using !DNSServer from Justin Fletcher (yes there
>> is a Public Domain DNS-server for the Iyonix ...) which was written
>> 1997 in BASIC. As far as I renember sometimes the Samba server
>> tries to make usage of such a server (depends on device
>> connected). If this would be the case the "stiff" should
>> disappear after lets say 40 seconds or so. However I am ever
>> starting both programs together since two years so wasn't aware of
>> it.

The problem here seems to have happened because on setting up !Samba I
managed not to notice that an interface is set by default in smb/conf,
usually this would be blank or at least it needs to be the same as is
set on the machine.

That corrected the new !Samba is now working on the Iyonix. (I have not
got it to work on the A9home yet. The WindowsXP laptop cannot find the
folder.)

I did get a few of these during setup on both the Iyonix and A9home, not
quite sure why :-

07 Apr 14:47:15 000 00800E00: Error from (unknown): Internal error, trap
while in trap handler: Internal error: abort on instruction fetch at
&00000002, pc = FFFFFFFE: registers at 0015DE58


>> Unfortunately I don't know whether !DNSServer is still public
>> available. I took it from an old CD as far as I renember. !DNSServer
>> is "... This application is Public Domain Freeware. This means that
>> it may be distributed, so long as no charge other than copying costs
>> are charged for it. The copyright remains with myself (Justin
>> Fletcher).". So I think that I can sent you a copy if required. Would
>> be glad if soneone could give a hint to the original download link.

A 32bit DNS server would be useful, I have some of my machines set to
with manual configuration just the name them. That was the only way to
get Filezilla to find FTPs on the Iyonix.

>The !smbserver.Resources.smb/conf file does not contain a
>"nameserver enabled" entry which seems to lead to an enabling.
>Either it must be disabled manually in the configuration
>or the entry
>
> nameserver enabled = No
>
>must be added to the
>
># Global parameters
>
>section at the end or !DNSSserver must be installed.

There is an oddity there, if the option is unticked then that line
appears in smb/conf on saving, if the option is ticked then the line is
absent.

Thanks for the help.

Thomas Milius

unread,
Apr 7, 2007, 12:10:17 PM4/7/07
to
In message <e283dacf4...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

> In message <09c1d3cf...@thomas-milius.t-online.de>
> Thomas Milius <Thomas...@t-online.de> wrote:
>
> >In message <0321d1cf...@thomas-milius.t-online.de>
> > Thomas Milius <Thomas...@t-online.de> wrote:
>
> >> In message <70e1cecf...@thomas-milius.t-online.de>
> >> Thomas Milius <Thomas...@t-online.de> wrote:
>
> >> > In message <2f1dc2cf4...@pittdj.plus.com>
> >> > David Pitt <ne...@pittdj.co.uk> wrote:
>
> >> > > There is clearly something different, the version from
> >> > > sourceforge works. The version on Herbert's site stiffs my Iyonix
> >> > > everytime.
>
> [snip]
>

> That corrected the new !Samba is now working on the Iyonix. (I have not
> got it to work on the A9home yet. The WindowsXP laptop cannot find the
> folder.)
>

May be this is due to the fact that I linked it aginst Castles
C-Library which is not available for the A9Home as far as I renember.
So it seems that I have to link it independently.
However I must see how to do so.

Regards

Thomas Milius

Thomas Milius

unread,
Apr 7, 2007, 12:58:54 PM4/7/07
to

> In article <slrnf1elh...@compsoc.dur.ac.uk>,
> Herbert zur Nedden <HzN....@HQ.gag.de> wrote:
> > In agreement with Thomas I supply his version 0.08 of !smbserver
> > including source code on <URL:http://www.gag.de/software/smbserver> and
> > it will remain available there until UPP does supply a better version.
>
> Is this different to the version 0.08 of SmbServer available at
> http://riscossmbserver.sourceforge.net/
>

The changes as far as I renember are as follows:

1. There was an attirbute problem with masking some bits
which I changed. Afterwards the silly dates for some
files generated by my DVB-T receiver vanished.
2. During tests with another selfwritten program I had to
detect that a networking result is not always correct
under RISC OS. So the according parts inside Samba
were changed.
3. The program crashed after 20 Minutes DVB-T recording regulary.
Reason was a wrong array definition. This has been corrected
and the program works for hours of recording or replaying
now in a stable way.
4. I added a simple read prediction to accelerate DVB-T replay
transfer speed. The options worked but did solve my personal
problem for it was based on old too slow dLAN adapters.
5. All is compiled with newer versions of Castles C/C++ Compiler.
This may have a positive speed effect.

All changes are marked with a "TM" inside the code.

However in the moment I linked against Castles C-library so the
version may not be useable on the A9Home. I shall see wehther I
can change this. It is also compiled for StrongARM upwards
and optimized for Xscale. So old machines with ARM610/710
or older may get trouble.

Regards

Thomas Milius

David Pitt

unread,
Apr 7, 2007, 12:59:57 PM4/7/07
to
In message <295be1cf...@thomas-milius.t-online.de>
Thomas Milius <Thomas...@t-online.de> wrote:

>In message <e283dacf4...@pittdj.plus.com>
> David Pitt <ne...@pittdj.co.uk> wrote:
>
>> In message <09c1d3cf...@thomas-milius.t-online.de>
>> Thomas Milius <Thomas...@t-online.de> wrote:
>>
>> >In message <0321d1cf...@thomas-milius.t-online.de>
>> > Thomas Milius <Thomas...@t-online.de> wrote:
>>
>> >> In message <70e1cecf...@thomas-milius.t-online.de>
>> >> Thomas Milius <Thomas...@t-online.de> wrote:
>>
>> >> > In message <2f1dc2cf4...@pittdj.plus.com>
>> >> > David Pitt <ne...@pittdj.co.uk> wrote:
>>
>> >> > > There is clearly something different, the version from
>> >> > > sourceforge works. The version on Herbert's site stiffs my Iyonix
>> >> > > everytime.
>>
>> [snip]
>>
>> That corrected the new !Samba is now working on the Iyonix. (I have not
>> got it to work on the A9home yet. The WindowsXP laptop cannot find the
>> folder.)
>>
>
>May be this is due to the fact that I linked it aginst Castles
>C-Library which is not available for the A9Home as far as I renember.
>So it seems that I have to link it independently.
>However I must see how to do so.

I do not think so, I tried a 32bit version of Castle 32bit CLib,
provided in the Castle Tools, on the A9home and it still did not work.

The A9home does have a bug that stops NewsHound posting, this may be
similar.

The A9home is still beta.

The way round the CLib debacle is to compile against StubsG, if
possible.

I would recommend waiting until the A9home comes out of beta.

druck

unread,
Apr 7, 2007, 3:54:48 PM4/7/07
to
On 7 Apr 2007 David Pitt <ne...@pittdj.co.uk> wrote:
> The A9home does have a bug that stops NewsHound posting, this may be
> similar.

Anyone with a second machine and passing knowledge of NNTP tried
running a telnet server on the NNTP port and pointing NewsHound at it,
and seeing if its sending something wrong, or not responding?

---druck

--
The ARM Club Free Software - http://www.armclub.org.uk/free/
The 32bit Conversions Page - http://www.quantumsoft.co.uk/druck/

druck

unread,
Apr 7, 2007, 3:53:13 PM4/7/07
to
On 7 Apr 2007 Thomas Milius <Thomas...@t-online.de> wrote:

[snip]

> It is also compiled for StrongARM upwards and optimized for Xscale. So
> old machines with ARM610/710 or older may get trouble.

Compiling for StrongARM may generate halfword instructions which are
incompatible with the Risc PC's memory system. In both Norcroft and
GCC its best to specify a target architecture of v3 (so no
incompatible instructions are generated), but specify optimisation for
the XScale processor, as this will only effect instruction ordering.

Peter Naulls

unread,
Apr 7, 2007, 5:34:59 PM4/7/07
to
In message <58c4f5cf...@druck.freeuk.net>
druck <ne...@druck.freeuk.com> wrote:

> On 7 Apr 2007 Thomas Milius <Thomas...@t-online.de> wrote:
>
> [snip]
>
> > It is also compiled for StrongARM upwards and optimized for Xscale. So
> > old machines with ARM610/710 or older may get trouble.
>
> Compiling for StrongARM may generate halfword instructions which are
> incompatible with the Risc PC's memory system. In both Norcroft and
> GCC its best to specify a target architecture of v3 (so no
> incompatible instructions are generated), but specify optimisation for
> the XScale processor, as this will only effect instruction ordering.

http://www.riscos.info/index.php/GCC_processor

Unfortunately, Herbert hasn't made any attempt to contact me regarding
this, so it's rather hard to clear up whatever misunderstanding he
thinks might have happened. After all, I'm hosting sources for an
updated version, which addresses many of the stability isses mentioned
here (but is not really in a releasable state).


--
Peter Naulls - pe...@chocky.org | http://www.chocky.org/
----------------------------------------------------------------------------
RISC OS Community Wiki - add your own content | http://www.riscos.info/

Herbert zur Nedden

unread,
Apr 8, 2007, 2:14:27 AM4/8/07
to
In message <a519ffcf...@chocky.org>
Peter Naulls <pe...@chocky.org> wrote:

> In message <58c4f5cf...@druck.freeuk.net>
> druck <ne...@druck.freeuk.com> wrote:
>
> > On 7 Apr 2007 Thomas Milius <Thomas...@t-online.de> wrote:
> >
> > [snip]
> >
> > > It is also compiled for StrongARM upwards and optimized for Xscale. So
> > > old machines with ARM610/710 or older may get trouble.
> >
> > Compiling for StrongARM may generate halfword instructions which are
> > incompatible with the Risc PC's memory system. In both Norcroft and
> > GCC its best to specify a target architecture of v3 (so no
> > incompatible instructions are generated), but specify optimisation for
> > the XScale processor, as this will only effect instruction ordering.
>
> http://www.riscos.info/index.php/GCC_processor
>
> Unfortunately, Herbert hasn't made any attempt to contact me regarding
> this, so it's rather hard to clear up whatever misunderstanding he
> thinks might have happened. After all, I'm hosting sources for an
> updated version, which addresses many of the stability isses mentioned
> here (but is not really in a releasable state).

Hi Peter,

I don't see any misunderstanding here at all - well perhaps there is
one: I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
you plan to take some time off RISC OS).

There is no smbserver to be found on riscos.info (except some svn
repositories) and thus not much help for the normal user. Well to be
more precise: I used the search function on riscos.info and looked at
the UPP pages trying to locate smbserver and all I got was the svn
links.

Add to this It is just that in the recent past problems to access
Windows XP systems shares with LanMan etc populated the odd mailing list
and smbserver is one option to address the file sharing issue. Of the
different versions I found on my harddisc the Milius-variant was best.

I thus wanted the users to be able to download the Milius variant too -
since it does the job pretty well for Thomas and me (opposed to the odd
other smbserver version I found on my harddisc). And as I mentioned in
my announcement this is a temporary offer until UPP does supply an
enhanced version. As soon as that is the case I'm happy to take my page
off line.


--
Herbert

Herbert zur Nedden
German Archimedes Group http://www.gag.de

Herausgeber DER deutschen RISC OS-Zeitschrift.

Theo Markettos

unread,
Apr 8, 2007, 7:57:40 AM4/8/07
to
Herbert zur Nedden <HzN....@hq.gag.de> wrote:
> There is no smbserver to be found on riscos.info (except some svn
> repositories) and thus not much help for the normal user. Well to be
> more precise: I used the search function on riscos.info and looked at
> the UPP pages trying to locate smbserver and all I got was the svn
> links.
>
> Add to this It is just that in the recent past problems to access
> Windows XP systems shares with LanMan etc populated the odd mailing list
> and smbserver is one option to address the file sharing issue. Of the
> different versions I found on my harddisc the Milius-variant was best.

I didn't see this thread until just now, but I've created an smbserver page
on riscos.info. As a result of reading the thread I'm not quite sure which to
recommend to downloaders. How well has the original 0.08 been tested, and
how well has Thomas' 0.08? Currently it recommends Thomas', but perhaps it
ought to stress that the original is the one that's been more heavily
tested, and Thomas' is alpha/beta?

Theo

Peter Naulls

unread,
Apr 8, 2007, 9:42:46 AM4/8/07
to
In message <54a42ed...@hznipc.zurnedden.de> you wrote:

> I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
> you plan to take some time off RISC OS).

I haven't gone away. What I said was that I would take time off RISC OS
development due to all the nonsense. What you've done here is just a
bit too characteristic of that, I'm afraid.

> In message <a519ffcf...@chocky.org>
> Peter Naulls <pe...@chocky.org> wrote:


> > this, so it's rather hard to clear up whatever misunderstanding he
> > thinks might have happened. After all, I'm hosting sources for an
> > updated version, which addresses many of the stability isses mentioned
> > here (but is not really in a releasable state).
>
> Hi Peter,
>
> I don't see any misunderstanding here at all - well perhaps there is
> one: I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
> you plan to take some time off RISC OS).

Well, there is in your very own words:

"Unfortunately the only result of Thomas' decision was only that now his
enhanced version of !smbserver was unavailable since for some reason
beyond my understanding..."

> There is no smbserver to be found on riscos.info (except some svn
> repositories) and thus not much help for the normal user. Well to be
> more precise: I used the search function on riscos.info and looked at
> the UPP pages trying to locate smbserver and all I got was the svn
> links.

That's mostly all there is, as I said:

http://www.riscos.info/index.php/Smbserver


> I thus wanted the users to be able to download the Milius variant too -
> since it does the job pretty well for Thomas and me (opposed to the odd
> other smbserver version I found on my harddisc). And as I mentioned in
> my announcement this is a temporary offer until UPP does supply an
> enhanced version. As soon as that is the case I'm happy to take my page
> off line.

Yes, but guess what. We're now in exactly the same position as Firefox.
Two sets of information, one created by someone who didn't make any
effort to contact me as to the best thing to do, and providing a
version which often locks up machines. I really think it would have been
much less effort and much better for you to contact me.

David Pitt

unread,
Apr 8, 2007, 9:57:43 AM4/8/07
to
In message <5fb457d0...@chocky.org>
Peter Naulls <pe...@chocky.org> wrote:

>In message <54a42ed...@hznipc.zurnedden.de> you wrote:
>
>> I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
>> you plan to take some time off RISC OS).
>
>I haven't gone away. What I said was that I would take time off RISC
>OS development due to all the nonsense. What you've done here is just
>a bit too characteristic of that, I'm afraid.

Rudeness will just not do. Where is the nonsense. Herbert is simply
providing hosting for Thomas's work pending any developent of any better
version in the future.

Thomas Milius

unread,
Apr 8, 2007, 10:25:11 AM4/8/07
to
In message <5fb457d0...@chocky.org>
Peter Naulls <pe...@chocky.org> wrote:

> In message <54a42ed...@hznipc.zurnedden.de> you wrote:
>
> > I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
> > you plan to take some time off RISC OS).
>
> I haven't gone away. What I said was that I would take time off RISC OS
> development due to all the nonsense. What you've done here is just a
> bit too characteristic of that, I'm afraid.
>
> > In message <a519ffcf...@chocky.org>
> > Peter Naulls <pe...@chocky.org> wrote:
>
>
> > > this, so it's rather hard to clear up whatever misunderstanding he
> > > thinks might have happened. After all, I'm hosting sources for an
> > > updated version, which addresses many of the stability isses mentioned
> > > here (but is not really in a releasable state).
> >
> > Hi Peter,
> >
> > I don't see any misunderstanding here at all - well perhaps there is
> > one: I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
> > you plan to take some time off RISC OS).

> ...


>
> Yes, but guess what. We're now in exactly the same position as Firefox.
> Two sets of information, one created by someone who didn't make any
> effort to contact me as to the best thing to do, and providing a
> version which often locks up machines. I really think it would have been
> much less effort and much better for you to contact me.
>

I don't hope that you are meaning me with this Peter.

My versions definitely cures some bugs. You have been able to compare
the changes I made as I marked them and send them towards to you
to enable you to insert them into newer versions in case that you would
have used David Buxtons source for a new Samba version.

Only a small reminder for you:

----- Original Message from 9 Apr 2006 -----

In message <7d7dbb14...@chocky.org> you wrote:

>
> Hi Guys.
>
> I've been frustrated for a long time, as I think have others, by its
> sometime insistance on locking up the whole machine. Well, today I
> managed to fix that problem with the available source (or perhaps just
> work around it). Anyway, it no longer crashes or locks up as it did
> previously. I noticed shortly after that Thomas had made a bunch of
> changes himself, mostly for performance reasons, and unrelated to my
> changes (although perhaps fixing the problem in a completely different
> way).
>
> I wanted to check that I really did have the latest version of the
> sources, but the CVS for it on source forge seems unavailable.
>
> In any case, are you planning to take this project forward? If not, I
> would very much like to maintain it on riscos.info under subversion. I
> know this is still very useful to many people, even if like myself, it's
> just being used as a printer server.
>
> Regards.
>

In the moment I am not having the time to do further fixes (unfortunately).
A central available code would make sense. If you would put this
to riscos.info it yould be ok from my view of sight. I would remove
it from my homepage and would place a link to the new loccation.

I hope that the code and compilation remains free for everone and
that it is running on all RISC OS machines. May be it is possible
to keep the source in such a way that it could be compiled by GCC
(which you are preferring AFAIR) and Castle/Norcroft C/C++ on
demand.

Regarding my changes:

All was based on dLAN trouble and DVB-T recording. AFAIR there was a
correction in blanking some flags in the attribbute settings. Then
I added read prediction which doesn't helped much in my case and
at the end I had to detect that there was a problem with the
array size definition which lead to regular crash after 20 minutes
of recordings if using block sizes above 8192 for it seems that
the one or two array definitions had been used for various concurrent
purposes.

Another change was to add read prediction and workgroup to the
configuration. Especally the missing workgroup caused trouble
without end to many users.

Would be be glad if you could ensure this on your new version for
my version is running very stable on my machines. That doesn't mean
that you must use any of the modifications I made. Only the results
count. I must say that I found the source a bit complicated to read.
Therefore it is likely that different people regarding the same problem
might come to different solutions ;-).

If you are examine the code in more detail there are the following
wishes still openstanding since a while from me and other persons:

1. There seems to be a problem with the amount of files in a directory.
It seems to depends on an odd buffer size. (Dicussed in the network
newsgroup)
2. In the moment I am not able to do timeshifting on my DVBT-T
receiver. I assume that this is based that the DVB-T recorder
tries to open the same file once for reading and once for writing
buit Samba does not recognize that it must use th same RISC OS
handle for this situation and therefore refuses to open the read
process file.
3. Password handling.

Under certain circumstances there have been hang ups. I cured this
on my machine by running a DNSserver (there is one old free one
written in BASIC by Justin Fletcher in 1997 which also runs on the
Iyonix). Under certain circumstances the samba server starts looking up
for names which leads to endless activity and delay without such a
tool.

Best Regards

Thomas Milius

----- Original Message from 9 Apr 2006 Ends -----

Everyone would be glad if there would be a new Samba version for RISC OS from
which one ever. I had problems with Davids version in 2005 as I used it
intensively for the first time and made the modifications which I wanted to
share with other people until there would be a new version from David. I took
it from my homepage as you contacted me (see above). And them? Nothing
happened, but a couple of people contacted me to ask where they could found
my improved version.

Herbert and I are relatively satisfied with my version. Of course
there could be better ones. But when and from which person?

Regards

Thomas Milius

Herbert zur Nedden

unread,
Apr 8, 2007, 12:03:37 PM4/8/07
to
In message <5fb457d0...@chocky.org>
Peter Naulls <pe...@chocky.org> wrote:

> In message <54a42ed...@hznipc.zurnedden.de> you wrote:
>
> > I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
> > you plan to take some time off RISC OS).
>
> I haven't gone away. What I said was that I would take time off RISC OS
> development due to all the nonsense. What you've done here is just a
> bit too characteristic of that, I'm afraid.

OK, let us not start this over again - we both know that we think
differently about patches, work arounds etc. (and that we both prefer
properly fixed versions).

> > There is no smbserver to be found on riscos.info (except some svn
> > repositories) and thus not much help for the normal user. Well to be
> > more precise: I used the search function on riscos.info and looked at
> > the UPP pages trying to locate smbserver and all I got was the svn
> > links.
>
> That's mostly all there is, as I said:
>
> http://www.riscos.info/index.php/Smbserver

Nice try - or shall I call it insulting!

That page has been created on April 4 (that is today) at 13:39.

Thus when I checked your website for smbserver to check if your
announced efforts for a new version resulted in something to download to
decide if I should put Thomas' version online or not nothing sensible
was there to be found!

> > I thus wanted the users to be able to download the Milius variant too -
> > since it does the job pretty well for Thomas and me (opposed to the odd
> > other smbserver version I found on my harddisc). And as I mentioned in
> > my announcement this is a temporary offer until UPP does supply an
> > enhanced version. As soon as that is the case I'm happy to take my page
> > off line.
>
> Yes, but guess what. We're now in exactly the same position as Firefox.
> Two sets of information, one created by someone who didn't make any
> effort to contact me as to the best thing to do, and providing a
> version which often locks up machines. I really think it would have been
> much less effort and much better for you to contact me.

Yes, but guess what: This is absolutely no surprise.

Quite some time ago you stated that you'd take over smbserver on UPP or
riscos.info. Your announcement did trigger Thomas to tell you about his
efforts, offer you his sources and then take his version offline to
avoid the different versions in different places to be there.

But since ever since then nothing visible did happen but the odd user
wanted a working smbserver (probably triggered to a certain extend to
the new issues with LanMan failing to access Windows XP in the recent
past). Since seaching for smbserver resulted in just revealing the svn
pages on riscos.info but no new version at all I decided to offer the
smbserver version of Thomas which does work fine for him and me for
download.

Thus indeed we're now in exactly the same position as Firefox. Some
alternative offer is online until a enhanced, fixed version of Firefox
resp. smbserver apprears on riscos.info.

Herbert zur Nedden

unread,
Apr 8, 2007, 12:23:24 PM4/8/07
to

Short update:

Thomas just sent me a version linked with StubsG and for ARM
architecture v3 which I put on http://www.gag.de/software/smbserver

Perhaps this has better chances on A9home or the like...

David Pitt

unread,
Apr 8, 2007, 1:56:10 PM4/8/07
to
In message <c16466d...@hznipc.zurnedden.de>

"Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:

>Thomas just sent me a version linked with StubsG and for ARM
>architecture v3 which I put on http://www.gag.de/software/smbserver
>
>Perhaps this has better chances on A9home or the like...

If only. Welcome to the wonderful world of the A9home. The new StubsG
compilation fails on the A9home with an old friend of an error message,
"Unable to start application (decompression failed)".

I manually xpanded !RunImage on the A9home and sadly I could not
persuade my laptop to find the share. I am reasonably persuaded that the
problem is because the A9home uses DHCP. My Iyonix, and RiscPC both have
manually set IP addresses, and Host names, and do allow the laptop to
connect. I think the A9home Host name is unknown to the laptop. If I
could manually configure the A9home I guess the problem would go away,
but that is another A9home foible, mine will not manually configure.

The version compiled with the 32bit Stubs did run on the A9home without
falling foul of the not decompressing issue. I assume it was squeezed
with an earlier version of the Castle Tools. It is the latest Castle
squeeze that ROL OSes dislike. RO6pv is similarly afflicted, the StubsG
version fails to decompress with it.

Thanks for all the good work.

Herbert zur Nedden

unread,
Apr 9, 2007, 12:34:32 AM4/9/07
to
In message <ffe26ed04...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

> In message <c16466d...@hznipc.zurnedden.de>
> "Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:
>
> >Thomas just sent me a version linked with StubsG and for ARM
> >architecture v3 which I put on http://www.gag.de/software/smbserver
> >
> >Perhaps this has better chances on A9home or the like...
>
> If only. Welcome to the wonderful world of the A9home. The new StubsG
> compilation fails on the A9home with an old friend of an error message,
> "Unable to start application (decompression failed)".
>
> I manually xpanded !RunImage on the A9home and sadly I could not
> persuade my laptop to find the share. I am reasonably persuaded that the
> problem is because the A9home uses DHCP. My Iyonix, and RiscPC both have
> manually set IP addresses, and Host names, and do allow the laptop to
> connect. I think the A9home Host name is unknown to the laptop. If I
> could manually configure the A9home I guess the problem would go away,
> but that is another A9home foible, mine will not manually configure.

Well even if it uses DHCP I am pretty sure that it will get the same IP
address all the time - especially if the other systems with a fixed
address get one outside the DHCP address pool.

You can try to connect to the A9home from Windows using the IP address
instead of the name, i.e.

in Windows explorer try "\\192.168.1.2\share" instead of the common
"\\a9home\share" - certainly replacing the ip address I show here be the
one of the a9home, the host name "a9home" by the real one and the share
name "share" by the real one.

Or try this command (again with correct ip address and share name):
"net use * \\192.168.1.2\share"

> The version compiled with the 32bit Stubs did run on the A9home without
> falling foul of the not decompressing issue. I assume it was squeezed
> with an earlier version of the Castle Tools. It is the latest Castle
> squeeze that ROL OSes dislike. RO6pv is similarly afflicted, the StubsG
> version fails to decompress with it.
>
> Thanks for all the good work.

Sorry I can't help more but perhaps Thomas has an idea.

Herbert zur Nedden

unread,
Apr 9, 2007, 12:37:20 AM4/9/07
to
In message <Sgh*Ty...@news.chiark.greenend.org.uk>
Theo Markettos <theom...@chiark.greenend.org.uk> wrote:

Hi Theo,

thanks for that page.

The Milius-0.08 works better on Thomas' and my IYONIX pc and is
basically the same as the standard 0.08 with some fixes. Since the older
ones (pre 0.08) as well as the standard 0.08 doesn't work as stable as
the Milius variant on my IYONIX pc I dare say it is better.

I only put the Milius version for download since the other version
should be on the net anyhow...

David Pitt

unread,
Apr 9, 2007, 4:41:07 AM4/9/07
to
In message <7454a9d...@hznipc.zurnedden.de>

"Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:

>> In message <c16466d...@hznipc.zurnedden.de>
>> "Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:

>> >Thomas just sent me a version linked with StubsG and for ARM
>> >architecture v3 which I put on http://www.gag.de/software/smbserver

[snip]

>You can try to connect to the A9home from Windows using the IP address
>instead of the name, i.e.
>
>in Windows explorer try "\\192.168.1.2\share" instead of the common
>"\\a9home\share" - certainly replacing the ip address I show here be the
>one of the a9home, the host name "a9home" by the real one and the share
>name "share" by the real one.
>
>Or try this command (again with correct ip address and share name):
>"net use * \\192.168.1.2\share"

I did try the DHCP served IP with no success.

I have now manually set the IP on the A9home and still the laptop does
not see the share.

On attempting to connect to the share from the laptop Samba errors and
quits :-

100 ***Error***
100 Title : Error
100 Task : SmbServer
100 Message : Internal error, trap while in trap handler:
Internal error:
abort on instruction fetch at &FFFFFFFC,
pc = FFFFFFFC: registers at 0015F6D8
100


And now for the bad news. I tried the 32bit Samba on the Foundation CD
and that works, the laptop does see the A9home. This Samba 0.08 17:01:37
Nov 17 2002, which is the version on sourceforge.

Thomas Milius

unread,
Apr 9, 2007, 5:24:39 AM4/9/07
to
In message <f0e7bfd04...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

I just mailed Herbert a version which is not squeezed to exclude the A9Home
problems with it.

It is uncertain what happens. The message does not say much.
It may be that it is a consequence of the first problem that
you are not seeing the A9Home-Share on the laptop.

I must say that it logically doesn't fit into the changes I made.
For it works on your Risc PC and your Iyonix it must be something
different on the A9Home. The only thing I can guess are other libraries.

Is there another library than StubsG which might mismatch?
I am not an owner of an A9Home and so I did not follow the discussion
which parts might went wrong. I am using the Castle C/C++ development
environment and so I am not aware about ROL specialities.

What is your exact !smbserver setup (RISC PC/Iyonix/A9Home)? Are you using
DNServer meanwhile? Perhaps this could give me a hint what could be wrong.
I only renember that I had to play around with the settings until
my only "Windows" machine, a CE based portable, recognized my Samba
Server on the Iyonix.

Regards

Thomas Milius

Michael Emerton

unread,
Apr 9, 2007, 8:10:46 AM4/9/07
to


Hi

Oddly with Adjust, 0.08 Runs on my machine without a flaw (attempted to
install to a fresh install, and worked fine)

in the last 4 months it has never crashed.

hth
Michael

Herbert zur Nedden

unread,
Apr 9, 2007, 9:13:13 AM4/9/07
to thomas...@t-online.de, ne...@pittdj.co.uk
[Posted and mailed]


Thomas was busy again:

On http://www.gag.de/software/smbserver is now new

1. separate source archive
2. the StubsG version now not compressed which perhaps is better
for a9home


> I tried the 32bit Samba on the Foundation CD and that works, the
> laptop does see the A9home. This Samba 0.08 17:01:37 Nov 17 2002,
> which is the version on sourceforge.

Good news ... I added a link to sourceforge on my smbserver page.

Thus for smbserver it seems to depend on the system you run which is
best. Let's hope that UPP presents a common and better smbserver soon.

Peter Naulls

unread,
Apr 9, 2007, 11:14:48 AM4/9/07
to
In message <010e59d04...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

There is no rudeness on my part, thanks. I'm simply pointing out what
should be very obvious. I did some work on this to cure some serious
stability issues - issues which are still showing up in this thread by
users who don't have such fixes, and users who will most likely be very
confused by these issues.

My concern is very simple - Herbert made no attempt to contact me
regarding this, and now the situation is much worse. It's astonishing
that so many people have demanded that ROL and Castle work together, but
when it comes others, we are so ready to brush off collobative
development, and make up excuses as to why it's so much better to work
separately in different directions.

If Herbert was really interested in taking things forward, there's no
reason why he couldn't maintain the version on riscos.info and make a
release, etc, etc. At worst, post Thomas' version there, and state it
still has the stability issues of the original. Unfortunately, such
forward-thinking continues to be lacking in RISC OS. I look forward to
Herbert making steps to remedy the situation.

Herbert zur Nedden

unread,
Apr 9, 2007, 1:41:58 PM4/9/07
to thomas...@t-online.de
[Posted and mailed]

In message <8cf7e3d0...@chocky.org>
Peter Naulls <pe...@chocky.org> wrote:

> In message <010e59d04...@pittdj.plus.com>
> David Pitt <ne...@pittdj.co.uk> wrote:
>
> > In message <5fb457d0...@chocky.org>
> > Peter Naulls <pe...@chocky.org> wrote:
> >
> > >In message <54a42ed...@hznipc.zurnedden.de> you wrote:
> > >
> > >> I'm glad you're back Peter (in IYONIX smartgroup you mentioned that
> > >> you plan to take some time off RISC OS).
> > >
> > >I haven't gone away. What I said was that I would take time off RISC
> > >OS development due to all the nonsense. What you've done here is just
> > >a bit too characteristic of that, I'm afraid.
> >
> > Rudeness will just not do. Where is the nonsense. Herbert is simply
> > providing hosting for Thomas's work pending any developent of any better
> > version in the future.
>
> There is no rudeness on my part, thanks. I'm simply pointing out what
> should be very obvious. I did some work on this to cure some serious
> stability issues - issues which are still showing up in this thread by
> users who don't have such fixes, and users who will most likely be very
> confused by these issues.

Well then perhaps the question arises why you never offered that version
for downloading and testing - you more often than not blame quite a few
others for their lack of helping you. But if all that happens is that
you state that he will enhance smbserver and that is all how do you
expect anybody to help?

But, if due to deficiencies of software some patches or work-arounds
appear which do help the odd user to be able to use that software you
start blaming the ones offering that kind of help. Nobody will disagree
that a fixed version of some software - be it of smbserver or of Firefox
2 release 2 - is better than some kind of patch or the like.

But you seem to ignore that some users out there are not capable of
fixing the software but want to be able to use it and are more than
happy to use whatever patch makes it work to be able to use it now and
not wait for some indefinite time. True, quite a few users are used to
have to wait pretty long for some release but that doesn't mean that
they are happy to do so.

> My concern is very simple - Herbert made no attempt to contact me
> regarding this, and now the situation is much worse. It's astonishing
> that so many people have demanded that ROL and Castle work together, but
> when it comes others, we are so ready to brush off collobative
> development, and make up excuses as to why it's so much better to work
> separately in different directions.

Wow, true I didn't contact you - the last message I read from you was on
IYONIX mailinglist where you said you'd take some time off RISC OS
development so why should I then disturb you since I got the impression
that you did want a break.

BTW: Peter you are concerned that I made not attempt to contact you.
Well why didn't you simply drop me a line but start this discussion
here?

So now let's get back to the smbserver issue: You do know of Thomas'
version since he told you and gave you his source modifications. So why
should I tell you that it is there?

Everybody was happy considering that now someone capable (with that I
mean you/UPP) wanted to look into enhancing smbserver since currently
the version you should use depends on the system you run.

But *nothing* happened as for your smbserver work - sorry, nothing
visible to the outside.

Since only the smbserver version with the patches by Thomas Milius runs
fine for him and for me and some users do have problems with smbserver
the most obvious solution was to offer that version for download until
you have found time to offer a new version of smbserver.

As for the stability issues you claim smbserver has: Since you know
of them please fix them. I don't have any such issues on my IYONIX pc
with the version by Thomas Milius - otherwise it would probably not make
much sense to re-publish it and I thus would probably not have done so.

> If Herbert was really interested in taking things forward, there's no
> reason why he couldn't maintain the version on riscos.info and make a
> release, etc, etc. At worst, post Thomas' version there, and state it
> still has the stability issues of the original. Unfortunately, such
> forward-thinking continues to be lacking in RISC OS. I look forward to
> Herbert making steps to remedy the situation.

I am intersted in things being taken forward, but I am not capable of
maintaining smbserver due to lack of time (quite some time would be
needed to first of all remove the current lack of knowledge and then the
time to do the work) so I won't try.

I look forward to Peter making steps to finish off his version of
smbserver (and with that remedying the situation). As I already stated I
will take Thomas' smbserver version off my web site as soon as a newer
working smbserver version is available.

Peter Naulls

unread,
Apr 9, 2007, 2:23:11 PM4/9/07
to
In message <106cf1d...@hznipc.zurnedden.de>

"Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:

> Well then perhaps the question arises why you never offered that version
> for downloading and testing - you more often than not blame quite a few
> others for their lack of helping you. But if all that happens is that
> you state that he will enhance smbserver and that is all how do you
> expect anybody to help?

That I didn't do any further work is unfortunate, but I spent a hell of
a lot of time on Firefox and related work. I'm not going to apologise
for that, although it's true I could have made my stance on smbserver
clearer. As for helping me, the solution is quite obvious -
collaborative development. The sources are avialable and buildable from
riscos.info SVN.

> But, if due to deficiencies of software some patches or work-arounds
> appear which do help the odd user to be able to use that software you
> start blaming the ones offering that kind of help. Nobody will disagree
> that a fixed version of some software - be it of smbserver or of Firefox
> 2 release 2 - is better than some kind of patch or the like.

That isn't my argument. My point, underlined once again, is that
"patches" by people unwilling to look beyond short-term fixes are often
worse.

> But you seem to ignore that some users out there are not capable of
> fixing the software but want to be able to use it and are more than
> happy to use whatever patch makes it work to be able to use it now and
> not wait for some indefinite time.

This isn't about that. It's about your actions. If you were so
concerned about having fixes, as you apparently were, then again, you
should have contacted me.

> BTW: Peter you are concerned that I made not attempt to contact you.
> Well why didn't you simply drop me a line but start this discussion
> here?

The onus was entirely yours. Do not try to blame me. I had no idea of
your intentions until you made a public statement.

> Since only the smbserver version with the patches by Thomas Milius runs
> fine for him and for me and some users do have problems with smbserver
> the most obvious solution was to offer that version for download until
> you have found time to offer a new version of smbserver.

"obvious" perhaps, but not the most responsible, nor the best for
long-term development. Let's quote "guestx" from drobe:

"...if I were making RISC OS software available only to see other
people's contributions top out at rituals and lucky charms from a
community who largely doesn't aspire to be anything more than a
bunch of tinkerers, even though they may pretend to be so much more,
I think I'd blow my lid at some point, too.".

> As for the stability issues you claim smbserver has: Since you know
> of them please fix them.

I have, in SVN, as I should have made clear by now. Additionally, if
you had contacted me, you would also know that I spent a good deal of
effort updating the documentation (which is dynamically generated) to
correct details, update links, etc, etc.

> I don't have any such issues on my IYONIX pc with the version by
> Thomas Milius - otherwise it would probably not make much sense to
> re-publish it and I thus would probably not have done so.

That is purely luck. As it stands in any released version, it has a
good chance of accessing zero page, and later using that value,
resulting in reported aborts or lock ups that we see here.

> I am intersted in things being taken forward, but I am not capable of
> maintaining smbserver due to lack of time (quite some time would be
> needed to first of all remove the current lack of knowledge and then the
> time to do the work) so I won't try.

I think you should have thought of this before making steps in this
direction. You have already spent more time now than you would have had
you contacted me. I would have hade no problems with an interim
release built from my sources, and incorporating Thomas's changes.
(Incidentally, I have no argument with Thomas, and I spent
considerable time with him and the original porter clarifying the
situation).

> I look forward to Peter making steps to finish off his version of
> smbserver (and with that remedying the situation). As I already stated I
> will take Thomas' smbserver version off my web site as soon as a newer
> working smbserver version is available.

There is unlikely to be any immediate version from me, for reasons I've
already detailed. If I was to do any RISC OS development right now, it
would probably be towards GCCSDK or Firefox, but I've stated my stace on
that too. Since it continues to be opaque to many, here it is again:

I'm unwilling to particpate in development for a platform where we
continue to give long-term development the short-shrift, pretend that
guestx's named magic charms are ok, and critisize those won't work
together (e.g. ROL vs Castle) but in our own efforts split development
just because we feel like it. If you are at all serious about RISC OS
going forward, then all this must change.

In this particular case, the correct course of action for you is:

- Decide if you are interested in any kind of role in smbserver
development.
- If so, then decide how and where a released version (which does _not_
have the named problems) should live.
- Decide what content should be on riscos.info regarding it.
Rememeber, riscos.info is far from my own content. Theo in
particular is keen to work with anyone needing specific content.

I recommend taking any specific technical discussion to the GCCSDK
mailing list. However, if you insist on maintaining the version you've
posted on your website (which I object to, and will cause only more
grief), then you should make it very clear that it is unstable, and a
corrected version will be [fill in your intentions here] and that you
are unable to provide support, etc, etc.

Please, try to think beyond short-term fixes and hacks and take some
kind of holistic view of RISC OS. If you don't, many others also won't.

Theo Markettos

unread,
Apr 9, 2007, 3:02:12 PM4/9/07
to
Herbert zur Nedden <HzN....@hq.gag.de> wrote:
> So now let's get back to the smbserver issue: You do know of Thomas'
> version since he told you and gave you his source modifications. So why
> should I tell you that it is there?
>
> Everybody was happy considering that now someone capable (with that I
> mean you/UPP) wanted to look into enhancing smbserver since currently
> the version you should use depends on the system you run.

I know nothing of the smbserver sources, so can I ask a couple of basic
questions:

How does the version in riscos.info subversion differ from the original?
How does it differ from Thomas'?
Why won't the subversion copy build at the moment, or why isn't it suitable
for general use?

Would it be worth introducing Thomas' changes from the original as a branch
in the SVN, with Peter's remaining the trunk? In a similar way that RISC OS
GCC 3.4 is forked from GCC 4 development, but some patches applied to 3.4
are also applied to the GCC 4 trunk.

That way the sources would be publically available, would be fairly
straightforward to build, and many bugfix patches for this branch can also
be applied to the trunk to update Peter's version. Then riscos.info could
host a 'stable' smbserver from that branch whilst development continued on
the trunk.

It might still be worth doing this even if Thomas' patches aren't 'stable'
enough to be promoted as such, since it's makes transferring patches between
branches easier.

> I am intersted in things being taken forward, but I am not capable of
> maintaining smbserver due to lack of time (quite some time would be
> needed to first of all remove the current lack of knowledge and then the
> time to do the work) so I won't try.

Do you or Thomas have the time to check things into SVN and generate
releases from that? Obviously it's far from ideal but it does mean those
sources are together?

Theo

David Pitt

unread,
Apr 9, 2007, 3:23:45 PM4/9/07
to
In message <22e4c3d0...@thomas-milius.t-online.de>
Thomas Milius <Thomas...@t-online.de> wrote:

>In message <f0e7bfd04...@pittdj.plus.com>
> David Pitt <ne...@pittdj.co.uk> wrote:
>
>> In message <7454a9d...@hznipc.zurnedden.de>
>> "Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:

[snip - Samba on the A9home]

>> On attempting to connect to the share from the laptop Samba errors
>> and quits :-
>>
>> 100 ***Error***
>> 100 Title : Error
>> 100 Task : SmbServer
>> 100 Message : Internal error, trap while in trap handler:
>> Internal error:
>> abort on instruction fetch at &FFFFFFFC,
>> pc = FFFFFFFC: registers at 0015F6D8
>> 100
>>
>>
>> And now for the bad news. I tried the 32bit Samba on the Foundation CD
>> and that works, the laptop does see the A9home. This Samba 0.08 17:01:37
>> Nov 17 2002, which is the version on sourceforge.
>>
>
>I just mailed Herbert a version which is not squeezed to exclude the
>A9Home problems with it.

That should be all that is necessary to get !Samba to run on the A9home.

>It is uncertain what happens. The message does not say much.
>It may be that it is a consequence of the first problem that
>you are not seeing the A9Home-Share on the laptop.
>
>I must say that it logically doesn't fit into the changes I made.
>For it works on your Risc PC and your Iyonix it must be something
>different on the A9Home. The only thing I can guess are other libraries.
>
>Is there another library than StubsG which might mismatch?
>I am not an owner of an A9Home and so I did not follow the discussion
>which parts might went wrong. I am using the Castle C/C++ development
>environment and so I am not aware about ROL specialities.

It comes as no great surprise that ROL's new OSes are not entirely
compatible with the output of the Castle Tools. Being a bit more
positive I understand it is ROL's intention to try to ensure
compatibility.

I am not convinced that this is a compilation issue.

The other way to ensure complete compatibility across our platform is
to use the SharedUnixLib which means using gcc. But not yet.

>What is your exact !smbserver setup (RISC PC/Iyonix/A9Home)? Are you
>using DNServer meanwhile? Perhaps this could give me a hint what could
>be wrong. I only renember that I had to play around with the settings
>until my only "Windows" machine, a CE based portable, recognized my
>Samba Server on the Iyonix.

I am not using !DNServer. I have tried !Samba with it DNS server option
in both states and with the A9home's "Act as DNS name server" option
turned off. Never does the laptop find the share with this !Samba.

I would recommend getting a second opinion, I do admit to be good at
making mistakes. But the fact that the older version worked first time
does lead me suppose I do know what I am doing.

I will try the newer !Samba in RO6pv tomorrow, that OS is a bit more
developed that the A9home's. I will also ask a question on the A9home
support list.

For now it is best to regard this as not urgent.

Herbert zur Nedden

unread,
Apr 10, 2007, 2:45:59 AM4/10/07
to
In message <i8o*Tn...@news.chiark.greenend.org.uk>
Theo Markettos <theom...@chiark.greenend.org.uk> wrote:


> I know nothing of the smbserver sources, so can I ask a couple of basic
> questions:
>
> How does the version in riscos.info subversion differ from the original?
> How does it differ from Thomas'?
> Why won't the subversion copy build at the moment, or why isn't it suitable
> for general use?
>
> Would it be worth introducing Thomas' changes from the original as a branch
> in the SVN, with Peter's remaining the trunk? In a similar way that RISC OS
> GCC 3.4 is forked from GCC 4 development, but some patches applied to 3.4
> are also applied to the GCC 4 trunk.
>
> That way the sources would be publically available, would be fairly
> straightforward to build, and many bugfix patches for this branch can also
> be applied to the trunk to update Peter's version. Then riscos.info could
> host a 'stable' smbserver from that branch whilst development continued on
> the trunk.
>
> It might still be worth doing this even if Thomas' patches aren't 'stable'
> enough to be promoted as such, since it's makes transferring patches between
> branches easier.
>

> Do you or Thomas have the time to check things into SVN and generate
> releases from that? Obviously it's far from ideal but it does mean those
> sources are together?

Hi,

sounds like a very good idea and perhaps Thomas' changes are in Peter's
source tree already since they did confer on these. I think there is no
need for a separate branch though since Peter did surely evaluate them
and either pick them up in his tree or address the respective issue
differently.

As for Thomas' changes: AFAIK they are flagged with "TM" in the soruces
he used to build his version and the sources are on my website next to
Thomas' build.

Unfortunately I can't check Thomas' sources into SVN - it would not be a
good idea for me to try never having looked at them or used SVN in the
first place which might result in more damage than good happening.

In any case, Peter Naulls is probably the only one who can answer this.

> Theo

Herbert zur Nedden

unread,
Apr 10, 2007, 2:37:05 AM4/10/07
to
[the odd snip]

> That I didn't do any further work is unfortunate, but I spent a hell of
> a lot of time on Firefox and related work. I'm not going to apologise
> for that, although it's true I could have made my stance on smbserver
> clearer. As for helping me, the solution is quite obvious -
> collaborative development. The sources are avialable and buildable from
> riscos.info SVN.

Nobody is blaming you on not having had time to work on smbserver...
but perhaps for a lack of information after the initial announcement
and then the long period of silence.
As for source based work: That is something only a few are capable of.

> That isn't my argument. My point, underlined once again, is that
> "patches" by people unwilling to look beyond short-term fixes are often
> worse.

True, but if the people in the know can't help due to lack of time or
the like I prefer a patch if it helps me use an application I want or
need to use.

> > BTW: Peter you are concerned that I made not attempt to contact you.
> > Well why didn't you simply drop me a line but start this discussion
> > here?
>
> The onus was entirely yours. Do not try to blame me. I had no idea of
> your intentions until you made a public statement.

Since as I wrote I had to assume you being away from RISC OS development
for some time and thus didn't want to disturb you in the first place.

Futhermore you could have simply dropped me a line to make me aware of
the fact that a) you're back and b) you're still working on smbserver to
then try to have a discussion offline - or simply ask for direct contact
here.

> > As for the stability issues you claim smbserver has: Since you know
> > of them please fix them.
> I have, in SVN, as I should have made clear by now.

You should have, indeed. And perhaps offered a pre-compiled test version
for download.

> > I don't have any such issues on my IYONIX pc with the version by
> > Thomas Milius - otherwise it would probably not make much sense to
> > re-publish it and I thus would probably not have done so.
> That is purely luck. As it stands in any released version, it has a
> good chance of accessing zero page, and later using that value,
> resulting in reported aborts or lock ups that we see here.

So I am lucky - fine with me.

Well, please allow for an end-user point of view on this one. If I have
a version of an application which does the task I want it to then that
is all I want. If other parts are broken or errors inside don't result
in it not working, it is a pity but not an issue for me in this case.

> I think you should have thought of this before making steps in this
> direction. You have already spent more time now than you would have had
> you contacted me. I would have hade no problems with an interim
> release built from my sources, and incorporating Thomas's changes.

Great, then please do so and I'm happy to take my website offline.

> There is unlikely to be any immediate version from me, for reasons I've
> already detailed.

Oops, I'm not sure I can follow this one. First you write that you have
had no problems with an interim release and now you state that there
will be none from you.

> I recommend taking any specific technical discussion to the GCCSDK
> mailing list.

Good idea.

> However, if you insist on maintaining the version you've posted on
> your website (which I object to, and will cause only more grief), then
> you should make it very clear that it is unstable, and a corrected
> version will be [fill in your intentions here] and that you are unable
> to provide support, etc, etc.

I am not maintaining it but just offering it for download. But I did add
a note to state that and to mention that is has some issues.

David Pitt

unread,
Apr 10, 2007, 3:38:20 AM4/10/07
to
In message <9cbdfad04...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

[snip]

>I would recommend getting a second opinion, I do admit to be good at
>making mistakes. But the fact that the older version worked first time
>does lead me suppose I do know what I am doing.
>
>I will try the newer !Samba in RO6pv tomorrow, that OS is a bit more
>developed that the A9home's. I will also ask a question on the A9home
>support list.

I have now tried smbserver 0.08 14:07:30 Aug 21 2005, the 32bit Stubs
variant, on my RiscPC, the results are decidedly odd. My laptop would
not find the share from the RiscPC whether it was running RO6pv or
OS4.04. The only machine that this version of smbserver does work fully
on is the Iyonix.

It is not an A9home only issue, so I will not raise it there.

John Cartmell

unread,
Apr 10, 2007, 4:24:34 AM4/10/07
to
In article <8efe3dd14...@pittdj.plus.com>,

> [snip]

It may still be worth bringing to Matt's attention. If he can solve the
problem for the A9 that should also solve it for Select inc RO6.

--
John Cartmell jo...@finnybank.com 0845 006 8822 or 0161 969 9820
Qercus magazine FAX +44 (0)8700-519-527 www.finnybank.com
Qercus - the best guide to RISC OS computing

Peter Naulls

unread,
Apr 10, 2007, 9:56:02 AM4/10/07
to
In message <e96238d...@hznipc.zurnedden.de>

"Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:

[snip excuse]

>
> > That isn't my argument. My point, underlined once again, is that
> > "patches" by people unwilling to look beyond short-term fixes are often
> > worse.
>
> True, but if the people in the know can't help due to lack of time or
> the like I prefer a patch if it helps me use an application I want or
> need to use.

Another excuse.

> > The onus was entirely yours. Do not try to blame me. I had no idea of
> > your intentions until you made a public statement.
>
> Since as I wrote I had to assume you being away from RISC OS development
> for some time and thus didn't want to disturb you in the first place.

Another excuse.

> Futhermore you could have simply dropped me a line to make me aware of
> the fact that a) you're back

I was never away.

> b) you're still working on smbserver to then try to have a discussion
> offline - or simply ask for direct contact here.

I'm not working on it. Another excuse.

> > > As for the stability issues you claim smbserver has: Since you know
> > > of them please fix them.
> > I have, in SVN, as I should have made clear by now.
>
> You should have, indeed. And perhaps offered a pre-compiled test version
> for download.

No, I shouldn't have. I have no time.

> > That is purely luck. As it stands in any released version, it has a
> > good chance of accessing zero page, and later using that value,
> > resulting in reported aborts or lock ups that we see here.
>
> So I am lucky - fine with me.

And very unlucky for everyone else. Look beyond your own requirements.

> Well, please allow for an end-user point of view on this one. If I have
> a version of an application which does the task I want it to then that
> is all I want. If other parts are broken or errors inside don't result
> in it not working, it is a pity but not an issue for me in this case.

You've made no attempt to obtain one. Another excuse.

>
> > I think you should have thought of this before making steps in this
> > direction. You have already spent more time now than you would have had
> > you contacted me. I would have hade no problems with an interim
> > release built from my sources, and incorporating Thomas's changes.
>
> Great, then please do so and I'm happy to take my website offline.
>
> > There is unlikely to be any immediate version from me, for reasons I've
> > already detailed.
>
> Oops, I'm not sure I can follow this one. First you write that you have
> had no problems with an interim release and now you state that there
> will be none from you.

Correct. You keep wrongly assuming that I actually intend or want to do
anything on it. I don't - not with this nonsense. That doesn't mean
you or someone else can't.

> I am not maintaining it but just offering it for download. But I did add
> a note to state that and to mention that is has some issues.

Too late. By putting it up for download, you implictly take some kind
of maintainership - that's not my doing, but your own, and you will get
plenty of emails about it.

Here's the bottom line. You *should* have contacted me. There is
nothing you can say that excuses that. You have made the situation much
worse, and you've not indicated what steps you are going to do to remedy
this. At this point the easist solution for me is simple to jettison
the whole thing, delete any sources I have and pretend I never have any
involvement with that. That is far better than dealing with another
situation where my efforts will be compromised because you cannot look
beyond a short-term goal.

Tim Powys-Lybbe

unread,
Apr 10, 2007, 10:39:10 AM4/10/07
to
In message of 10 Apr, Peter Naulls <pe...@chocky.org> wrote:

> Here's the bottom line. You *should* have contacted me. There is
> nothing you can say that excuses that. You have made the situation
> much worse, and you've not indicated what steps you are going to do to
> remedy this. At this point the easist solution for me is simple to
> jettison the whole thing, delete any sources I have and pretend I
> never have any involvement with that. That is far better than dealing
> with another situation where my efforts will be compromised because
> you cannot look beyond a short-term goal.

I regret that I simply cannot follow that reasoning.

There is no reason why anyone should have contacted you.

You had announced publicly after a furious exchange that you were not
going to do any more work and were retiring. There may have been some
qualification to this but it gave me the extremely strong impression
that you were best left alone in case we upset you further and
unnecessarily. Many people after a furious outburst are best left alone
and we all hope they recover.

So you were left alone.

A problem had developed with Samba and a perfectly sensible solution was
followed of putting an update in a place that others could download it.
This was not an offer to maintain, it was an offer to help others. The
person making the offer did not expect to be abused, he expected to be
thanked by those that used the facilities provided. To find himself
abused for not getting in touch with you when you had clearly indicated
that you wished to be left alone is just a shattering experience.

We are here on these newsgroups to help one another. You have your own
standards of what is helpful, and which standards imply that hardly
anyone else is to provide any help at all. But I do not think you have
the right to impose these standards on everyone. If others can offer
help, however imperfect, and this solves short-terms problems, then this
is good. You should be generous enough to recognise this and allow them
to provide this sort of help.

To say, as you have, above:

"You *should* have contacted me. There is nothing you can say that
excuses that."

is being ungenerous and showing a lack of understanding of how ordinary
people live and work together. It creates a blame culture and blame
cultures always indicate failing organisations.

There is a strong chance that the standards you have set for yourself
(not for others) are very good and will result in successful solutions.
If so, I will continue to support your work as I have done in the past
and as I have repeated privately to you. I gather that there are many
other people of the same view. Please believe us, please set yourself
high standards of work and please do not be offended by lesser mortals
who can help in their own, but not so exacting ways.

--
Tim Powys-Lybbe                                          t...@powys.org
             For a miscellany of bygones: http://powys.org/

Theo Markettos

unread,
Apr 10, 2007, 1:40:39 PM4/10/07
to
Herbert zur Nedden <HzN....@hq.gag.de> wrote:
> As for Thomas' changes: AFAIK they are flagged with "TM" in the soruces
> he used to build his version and the sources are on my website next to
> Thomas' build.
>
> Unfortunately I can't check Thomas' sources into SVN - it would not be a
> good idea for me to try never having looked at them or used SVN in the
> first place which might result in more damage than good happening.

When I have some time I can do the SVN stuff... I've just done a diff of
Thomas' code and the original 0.08 and indeed there aren't too many changes.
I'd be happy to pull those into a separate 'stable' SVN branch, but I feel
that they need more careful testing given the reported problems. Are the
problems due to A9 issues or the changes themselves? Can people run the
same setups side by side (with the same configuration files) and compare
them?

Theo

David Pitt

unread,
Apr 10, 2007, 3:58:10 PM4/10/07
to
In message <piv*fm...@news.chiark.greenend.org.uk>
Theo Markettos <theom...@chiark.greenend.org.uk> wrote:

I said earlier that the problems I encountered were not limited to the
A9home. The decompression failure is common to both the A9home and
RO6pv. The not connecting issue was common to both the A9home and RiscPC.

The decompression failure is only relevant to absolutes squeezed with
the latest Castle Tools.

I did eventually get the StubsG version to work on the A9home. There
seem to be a number of conditions that need to be met.

The A9home's IP address needs to be in !Samba's Global variables,
interfaces.

Do not have both encrypted password and null passwords selected at once.

Name server needs to be enabled in Global variables

I would not put too much money on these, there may be further erraticity
somewhere. I have had a torrid few hours.

I also have to say that the new !Samba, compiled with the 32bit Stubs,
stiffs my Iyonix something shocking.

Alan Calder

unread,
Apr 10, 2007, 4:04:54 PM4/10/07
to
In article <1c9760d1...@chocky.org>, Peter Naulls <pe...@chocky.org>
wrote:

> In message <e96238d...@hznipc.zurnedden.de> "Herbert zur Nedden"
> <HzN....@HQ.gag.de> wrote:

[Snip]

> > I am not maintaining it but just offering it for download. But I did
> > add a note to state that and to mention that is has some issues.

> Too late. By putting it up for download, you implictly take some kind
> of maintainership - that's not my doing, but your own, and you will get
> plenty of emails about it.

Too use your own phrase (much used in the section of the message I have
snipped) - Another excuse.

Where does it say in the Great Book of the IT Gods that putting something
onto a website automatically means you become the maintainer? If I put
up, with the author's permission, an old BBC program does that make me the
maintainer? Think not.

Note that Herbert did put a note on the site saying as much.

> Here's the bottom line. You *should* have contacted me. There is
> nothing you can say that excuses that. You have made the situation much
> worse, and you've not indicated what steps you are going to do to remedy
> this. At this point the easist solution for me is simple to jettison
> the whole thing, delete any sources I have and pretend I never have any
> involvement with that. That is far better than dealing with another
> situation where my efforts will be compromised because you cannot look
> beyond a short-term goal.

Why, precisely, *should* Herbert have contacted you? You weren't the owner
of the software in question.

The whole thing has become ridiculous, Peter. I hate conspiracy theories
(sorry, Paul :-) ) but it really does seem that you are looking for an
excuse to leave the RISC OS scene - to be able to say you were 'driven
away' by peoples' attitudes to your efforts. Unfortunately the only
'attitude' around here seems to be yours. Can it be an effect of being in
the US of A where everything outside is a dangerous and hatefilled place?

Please calm down.

Alan

--
Alan Calder, Milton Keynes, UK.

Thomas Milius

unread,
Apr 10, 2007, 3:46:47 PM4/10/07
to

If anyone could tell me where I am finding subversion (svn!?) for RISC OS
I would try to check out to rebuild myself.

As mentioned before in another mail my changes are of minor nature.
If running TX/RX size 8192 the old definition is ok. Only at higher
values (typically 16384) one buffer definiton is wrong.

One change affected an attribut mask. Don'T think that this causes this
trouble.

A couple of changes was done in evaluating the transmission
results which might lead to problems due to RISC OS internal
problems in this area under very seldom circumstances (independently
of Samba).

The read predction is optimal if it is not enabled it should not have
an effect.

I never said that I kicked out all bugs. However I hoped to correct
someones and not to add new ones. The question is whether the last
part is the case. For Samba works fine on my machine am I don't think so
but I can't prove.

Regards

Thomas Milius

Thomas Milius

unread,
Apr 10, 2007, 4:20:12 PM4/10/07
to
In message <8efe3dd14...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

> In message <9cbdfad04...@pittdj.plus.com>
> David Pitt <ne...@pittdj.co.uk> wrote:
>
> [snip]
>
> >I would recommend getting a second opinion, I do admit to be good at
> >making mistakes. But the fact that the older version worked first time
> >does lead me suppose I do know what I am doing.
> >
> >I will try the newer !Samba in RO6pv tomorrow, that OS is a bit more
> >developed that the A9home's. I will also ask a question on the A9home
> >support list.
>
> I have now tried smbserver 0.08 14:07:30 Aug 21 2005, the 32bit Stubs

????
Stopp please, the StubsG Variant was from 08 Apr 2007. What has happend
there? The unsqueezed version was from 09 Appr (!RunImage). If Castles
libraries are not installed the Stubs variant might fail even on the
Risc PC and as mentioned below you are using Select. Is this disabling
automatically the other libraries?

> variant, on my RiscPC, the results are decidedly odd. My laptop would
> not find the share from the RiscPC whether it was running RO6pv or
> OS4.04. The only machine that this version of smbserver does work fully
> on is the Iyonix.
>
> It is not an A9home only issue, so I will not raise it there.

I reinstalled yesterday the squeezed version on my RiscPC and
my RiscStation (RISC OS 4.02/4.03) and unfortunately I wasn't
able to detect any error as described. I could connect to
every machine with my Windows CE portbale.

However the whole thing is not really stable during configuration.
At my Risc PC I got hangs at the first attempt after installation
due to a port which was used by another appliaction (LANMANFS !?!).
Afterwards all worked fine.

My problem is that I have no really abilities to test with a Windows
PC. I am having only a Windows CE 4.2 based portable and it works
fine with it. Perhaps I could try to test with my old 486 PC-Card
and Windows 95. I am also using LINUX based DBV-T receiver which
was the reason for the my changes and it works fine with it
even with transmissions over hours in both directions.

I think we are facing a couple of possible problems:
- Did I add new bugs with my changes?
- Are there a couple of other bugs in worst case based on
uninitialized variables which would lead to absolutely
unreproduceable behaviour depending on the machine
and what the user did on it before?
- Had Samba and has it perhaps problems in conjunction
with different versions of Windows?
- Are there problems based on Castle/ROL (5.X vs Select/Adjust)
differences?
- Overall I had to find out that Samba (the original David
Buxton version included) shows sometimes odd behaviour
during configuration. One wrong parameter or copied but unchanged
resources files might cause trouble without end.

As discussed in a thread earlier there seems to be also the problem
that independetly of the version the directory dispaly seems to limited.
As I tried to trace it down I gave up as I came to the conclusion that
the problem is based on the size of a buffer. It must fit entirely
into it.

So there is the question whether it makes sense to continue using
the old source or whether it would be better to do a new port
based on the actual samba code even this would be much more work.

Regards

Thomas Milius

David Pitt

unread,
Apr 10, 2007, 6:11:53 PM4/10/07
to
In message <92be83d1...@thomas-milius.t-online.de>
Thomas Milius <Thomas...@t-online.de> wrote:

>In message <8efe3dd14...@pittdj.plus.com>
> David Pitt <ne...@pittdj.co.uk> wrote:
>
>> In message <9cbdfad04...@pittdj.plus.com>
>> David Pitt <ne...@pittdj.co.uk> wrote:
>>
[snip]

>> >I will try the newer !Samba in RO6pv tomorrow, that OS is a bit more


>> >developed that the A9home's. I will also ask a question on the
>> >A9home support list.
>>
>> I have now tried smbserver 0.08 14:07:30 Aug 21 2005, the 32bit Stubs
> ????
>Stopp please, the StubsG Variant was from 08 Apr 2007. What has happend
>there? The unsqueezed version was from 09 Appr (!RunImage). If Castles
>libraries are not installed the Stubs variant might fail even on the
>Risc PC and as mentioned below you are using Select. Is this disabling
>automatically the other libraries?

I was talking about the 32bit Stubs version running on the RiscPC. The
RiscPC, a Kinetic300, uses the ROL ROM 32bit CLib in RO6pv and the
Castle 32bit CLib when running OS4.04. RO6v errors, or at least objects,
if the Castle CLib is softloaded but it can be tried. I now believe the
problem is related to configuration settings. To be continued tomorrow.

The A9home runs ROL's ROM Clib. It can be persuaded to softload a 32bit
version of the 32bit Castle CLib.

So far I now do have the StubsG version running on the A9home, and did
that take some doing. I will try again with the normal 32bit Stubs
version with both CLibs tomorrow.

[snip]

>> It is not an A9home only issue, so I will not raise it there.
>
>I reinstalled yesterday the squeezed version on my RiscPC and
>my RiscStation (RISC OS 4.02/4.03) and unfortunately I wasn't
>able to detect any error as described. I could connect to
>every machine with my Windows CE portbale.

It appears to be a setup issue. Having found what works on the A9home I
will try that on the RiscPC.

>However the whole thing is not really stable during configuration.
>At my Risc PC I got hangs at the first attempt after installation
>due to a port which was used by another appliaction (LANMANFS !?!).
>Afterwards all worked fine.

That at least we can agree upon. With bad luck the zips contain an
unfortunate combination of settings in smb/conf. I have posted my
(interim) findings earlier this evening. I will have another look at
this tomorrow and if I get consistent answers I will post my suggestion
for smb/conf.

None of my machines had LanManFS running.

I do agree that once up and running Samba the setup instability stops.

I think it is too soon to abort your current work. There is a set up
issue, once that is cleared up a judgement could be made on the relative
stabilities of the versions.

My biggest concern is the poor stability on the Iyonix. I will revisit
that once I have the best setup sorted on the A9home and RiscPC. I am
aware that both Thomas and Herbert report good performance on the
Iyonix.

Peter Naulls

unread,
Apr 10, 2007, 9:44:37 PM4/10/07
to
In message <878564d...@south-frm.demon.co.uk>
Tim Powys-Lybbe <t...@powys.org> wrote:

> In message of 10 Apr, Peter Naulls <pe...@chocky.org> wrote:
>
> > Here's the bottom line. You *should* have contacted me. There is
> > nothing you can say that excuses that. You have made the situation
> > much worse, and you've not indicated what steps you are going to do to
> > remedy this. At this point the easist solution for me is simple to
> > jettison the whole thing, delete any sources I have and pretend I
> > never have any involvement with that. That is far better than dealing
> > with another situation where my efforts will be compromised because
> > you cannot look beyond a short-term goal.
>
> I regret that I simply cannot follow that reasoning.

That is too bad, but I have already expended too much energy to no
useful end.

>
> There is no reason why anyone should have contacted you.

[snip snip snip]

I have no wish to participate further in this charade. If you and
Herbert and others want to pretend that you know how to go about
software development (that is most assuredly what you are doing, even
though it's not programming), but ignore everything that has been gained
about how to go about collarobative development over the last 10 or so
years, then so be it. But I will have no part of such an unprofessional
and short-sighted view. I will remove my sources, and you can all
continue with your blinkers on and forget I did anything to improve the
intorerable instability of smbserver. It's better that way.

Peter Naulls

unread,
Apr 10, 2007, 10:24:59 PM4/10/07
to
In message <4ed1825819...@orpheusmail.co.uk>
Alan Calder <alan_...@orpheusmail.co.uk> wrote:

> In article <1c9760d1...@chocky.org>, Peter Naulls <pe...@chocky.org>

> Where does it say in the Great Book of the IT Gods that putting something


> onto a website automatically means you become the maintainer? If I put
> up, with the author's permission, an old BBC program does that make me the
> maintainer? Think not.

I'll pass that onto all the people who email you about it.

> Why, precisely, *should* Herbert have contacted you? You weren't the owner
> of the software in question.

Entirely presumption on your part. In fact, I *am* the maintainer of
the RISC OS version. Or rather was, since now I can't be bothered. I
was careful to make that clear with the original maintainer.

> The whole thing has become ridiculous, Peter. I hate conspiracy theories
> (sorry, Paul :-) ) but it really does seem that you are looking for an
> excuse to leave the RISC OS scene - to be able to say you were 'driven
> away' by peoples' attitudes to your efforts.

Why would I need to - there have been reasons not to use RISC OS for
many years. I don't have to make up excuses now.

> Unfortunately the only 'attitude' around here seems to be yours. Can
> it be an effect of being in the US of A where everything outside is a
> dangerous and hatefilled place?

Totally uncalled for. There is bigotry in any country, especially in
certain parts of the UK. Take this to somewhere more appropriate -
perhaps csa.misc.

> Please calm down.

I'm plenty calm. But I don't have to deal with people who can't see
beyond the end of their noses. There's really nothing in it for me.

Herbert zur Nedden

unread,
Apr 11, 2007, 2:35:13 AM4/11/07
to

Hi Peter,

> > I regret that I simply cannot follow that reasoning.
>
> That is too bad, but I have already expended too much energy to no
> useful end.

I dearly hope you understand at some time that others think differently.


> I have no wish to participate further in this charade.

Thanks!


> I will remove my sources, and you can all continue with your blinkers
> on and forget I did anything to improve the intorerable instability of
> smbserver. It's better that way.

Well I do hope that someone managed to pick up your sources and the
changes you did first since I must say that your behaviour is that of a
spoilt child.

Herbert zur Nedden

unread,
Apr 11, 2007, 2:34:43 AM4/11/07
to
In message <74af80d1...@thomas-milius.t-online.de>
Thomas Milius <Thomas...@t-online.de> wrote:

Thanks Theo and Thomas.

I hope you have the sources of Peter by now since he plans to delete
them...

BTW, if for some reason a place to offer things for download on the web
is needed, I'm glad to help.

Herbert zur Nedden

unread,
Apr 11, 2007, 2:16:42 AM4/11/07
to
Hi Peter,

until now I really tried to remain polite and tried to reply outlining
my views which differ from yours and due to your request I even added
some disclaimer to my page and a link to yours but you get more and more
insulting and now you blew it!

There is only so much a man can take!


[the odd snip]

> > > The onus was entirely yours. Do not try to blame me. I had no idea of
> > > your intentions until you made a public statement.
> > Since as I wrote I had to assume you being away from RISC OS development
> > for some time and thus didn't want to disturb you in the first place.
> Another excuse.
> > Futhermore you could have simply dropped me a line to make me aware of
> > the fact that a) you're back
> I was never away.

Well then you did lie on IYONIX mailinglist and you did lie here too
when you stated that you wanted to take some time off RISC OS
development. And when someone believes what you write you blame him.

Oh, sorry I have to excuse myself: My error for not having read exactly
since you wanted to take time off RISC OS development only but not from
insulting...


> > b) you're still working on smbserver to then try to have a discussion
> > offline - or simply ask for direct contact here.
> I'm not working on it. Another excuse.

Good to know.. amazing that you excuse yourself for not working on it!


> > > > As for the stability issues you claim smbserver has: Since you know
> > > > of them please fix them.
> > > I have, in SVN, as I should have made clear by now.
> > You should have, indeed. And perhaps offered a pre-compiled test version
> > for download.
> No, I shouldn't have. I have no time.

...since you spend your time with postings like these and when others


> > > That is purely luck. As it stands in any released version, it has a
> > > good chance of accessing zero page, and later using that value,
> > > resulting in reported aborts or lock ups that we see here.
> > So I am lucky - fine with me.
> And very unlucky for everyone else. Look beyond your own requirements.

What a stupid remark. Why unlucky for anybody else? Nobody needs to use
it but for some it works. And as it seems some others are now trying to
work on smbserver since the person who claimed to want to work on it a
while ago is busy with other things and now even thinks of behaving
like a small child not gettigs his wish fulfilled.


> > Well, please allow for an end-user point of view on this one. If I have
> > a version of an application which does the task I want it to then that
> > is all I want. If other parts are broken or errors inside don't result
> > in it not working, it is a pity but not an issue for me in this case.
> You've made no attempt to obtain one. Another excuse.

Wow, you do seem to have a very short memory or you just needed another
excuse to make another stupid remark!

I told you that I did search on riscos.info for a newer version but
didn't find anything apart from SVN pages, that is no download versions
and no hints of any activity (you even agreed that you should have
added some hint on progress). The only smbservers I found were the
differnt official versions and the one by Thomas and of these just one
version works fine on my system and that version was not online anymore
so I put it online (after asking the author) in case it helps others
too.


> > Oops, I'm not sure I can follow this one. First you write that you have
> > had no problems with an interim release and now you state that there
> > will be none from you.
> Correct. You keep wrongly assuming that I actually intend or want to do
> anything on it. I don't - not with this nonsense.

Well if you obviously are not even interested in doing anything as for
smbserver why did you never state that clearly and thus perhaps
encourage someone else to look into it. But when the message is still
that Peter is working on it why work on it and thus risk to collide with
him (or worse but probable: be insulted, blamed, ...)?

> > I am not maintaining it but just offering it for download. But I did add
> > a note to state that and to mention that is has some issues.
> Too late. By putting it up for download, you implictly take some kind
> of maintainership - that's not my doing, but your own, and you will get
> plenty of emails about it.

Another excuse - and another blatantly stupid remark.

And your right that my page is not your doing but my own - how did you
find out that one? And as for email I get about it: What's wrong with
that? I can reply and I have no problem with that and did already as a
matter of fact.

But one person seems to need a broad audience since he doesn't send me
an email but uses the newsgroup to drop me a line...


> Here's the bottom line. You *should* have contacted me. There is
> nothing you can say that excuses that.

Well there is now a pretty obvious excuse: You!

Considering how you act, insult, blame etc. someone like me instead of
simply asking kindly (in case this word is not on your list, try these:
friendly, polite, not-attacking, ...) that we get together and find some
solution how can you seriously expect someone to want to contact you?


> You have made the situation much worse, and you've not indicated what
> steps you are going to do to remedy this.

Another excuse - your very unfriendly behaviour is making things worse!

I am very sorry about this and for Peter.

Peter, you're very good programming and you did much good for RISC OS
and we all hope for more. Peter, please try to accept that other users
do think differently!

Herbert zur Nedden

unread,
Apr 11, 2007, 2:32:34 AM4/11/07
to
[snip]

> > Why, precisely, *should* Herbert have contacted you? You weren't the owner
> > of the software in question.
>
> Entirely presumption on your part. In fact, I *am* the maintainer of
> the RISC OS version. Or rather was, since now I can't be bothered. I
> was careful to make that clear with the original maintainer.

Hold it there! All that you stated long ago is plans to look into
smbserver but you didn't even manage to put up any information page on
your wiki (it being a wiki to enable just that, quick and simple
addintion of pages) or supply and progress report. But nothing, nil,
nada, zero!

You gotta be joking! And I did look on your web site but not finding
anything useful I figured the obvious and that you have to accept.


> > Please calm down.
>
> I'm plenty calm.

Wow, you are plenty calm and in that attitude insulting others in such a
manner. Then you don't even have any rage as excuse!

David Pitt

unread,
Apr 11, 2007, 5:54:13 AM4/11/07
to
In message <67f88dd14...@pittdj.plus.com>
David Pitt <ne...@pittdj.co.uk> wrote:

[snip]

> To be continued tomorrow.

>So far I now do have the StubsG version running on the A9home, and did
>that take some doing. I will try again with the normal 32bit Stubs
>version with both CLibs tomorrow.

It is now tommorrow and here are some answers.

A9home
------
I set up the 32bit Stubs variant 0.08 14:07:30 Aug 21 2005 as per the
StubsG version.

When acceptably setup both versions run on the A9home.

Acceptably set means :-

Do not select both "encrypt passwords" and "null passwords".

Enable name server. (I am not running !DNServer.)

It is not necessary to enter the A9home's IP address as I said
yesterday.

RiscPC
------
The 32bit Stubs version was similarly set up, as above, on the RiscPC, a
Kinetic300, and worked with OS4.04 and Castle's 32bit Clib. It would not
work with RO6pv until the Castle 32bit CLib was softloaded over ROL's
32bit ROM resident CLib.

The expanded StubsG variant was started and found to work after setting
a "netbios name" in smbserver's Global Variables, (ensure that in the
Resources.params file that 'netbios name' is set TRUE). For all other
OSes that can be left blank and the name taken is the Host name in the
Resolver configuration.

Iyonix
------
I find the 'nameserver enabled' option does not need to be ticked. So
far today there have been no Iyonix freezes.

Message has been deleted

Peter Naulls

unread,
Apr 11, 2007, 9:14:02 AM4/11/07
to
In message <fc5abad...@hznipc.zurnedden.de>

"Herbert zur Nedden" <HzN....@HQ.gag.de> wrote:

> > Another excuse.
> > > Futhermore you could have simply dropped me a line to make me aware of
> > > the fact that a) you're back
> > I was never away.
>
> Well then you did lie on IYONIX mailinglist and you did lie here too
> when you stated that you wanted to take some time off RISC OS
> development. And when someone believes what you write you blame him.

I've not lied anywhere. Don't try and cover your the flaw in your
presumption with bluster.

[huge snip]

The rest of this demonstrates you're far more interested in arguing (is
your name John Cartmell?) than doing anything of substance. I've given
you ample chance to take a professional approach to this, but you've
rejected it on all counts. I won't make any further attempt to sway you
from your fantasy that you are doing the right thing.

> [snip]

You snipped attributes. Be more careful - Alan Calder said this:

>
> > > Why, precisely, *should* Herbert have contacted you? You weren't the owner
> > > of the software in question.
> >
> > Entirely presumption on your part. In fact, I *am* the maintainer of
> > the RISC OS version. Or rather was, since now I can't be bothered. I
> > was careful to make that clear with the original maintainer.
>
> Hold it there! All that you stated long ago is plans to look into
> smbserver but you didn't even manage to put up any information page on
> your wiki (it being a wiki to enable just that, quick and simple
> addintion of pages) or supply and progress report. But nothing, nil,
> nada, zero!

Well, you can pretend I'm not the maintainer, just because you weren't
aware of the presise details (because you didn't contact me) if you
like, that doesn't make it not true. I'm certainly not required to
document everything I intend to do on the wiki.

But since you so insist on this, there _is_ a page on riscos.info about
my intentions (which I apologies for not mentioning earlier), and
previously linked from the news section, dated 26th May 2006. It's not
the page that Theo created. It's immediately obvious from a Google
search, so please, no further excuses.

> You gotta be joking! And I did look on your web site but not finding
> anything useful I figured the obvious and that you have to accept.

I accept that you are making a mockery of collaborative development.
That's your decision, not mine.

Message has been deleted
Message has been deleted

Theo Markettos

unread,
Apr 11, 2007, 12:09:08 PM4/11/07
to
Thomas Milius <Thomas...@t-online.de> wrote:
> If anyone could tell me where I am finding subversion (svn!?) for RISC OS
> I would try to check out to rebuild myself.

http://www.cp15.org/versioncontrol/

I ought to add that Peter's now linked in his more comprehensive wiki page
to the riscos.info front page. His original page is here:
http://www.riscos.info/index.php/RISC_OS_Samba_Server
As that's more comprehensive I'm happy for him to delete my page, which he's
done, and it now redirect's to Peter's page.

Theo

Herbert zur Nedden

unread,
Apr 11, 2007, 12:28:45 PM4/11/07
to

In message <7395e0d1...@chocky.org>
Peter Naulls <pe...@chocky.org> wrote:

[snip a lot of bullshit and thus no comment on the snipped PNcrap]
[but I had to comment a few things]

> The rest of this demonstrates you're far more interested in arguing (is
> your name John Cartmell?) than doing anything of substance.

Well first of all who started this discussion.

Secondly the fact that you now throw in the name of someone completely