Which QNX OS do you wish to develop for?
You can’t “download” QNX 4 or Neutrino.
You will, however, soon be able to download the Realtime Platform when it is
released.
Andrew Hurley <ahurley@gcom.nnnneeettttt.au> wrote in message
news:3975c048@Internet02…
I’m wanting to develop applications for QNX, and would like to know where
I
could get a download of the OS or a CD. Any tips would be helpful.Thankyou,
Andrew.
Hi!
I am using the following scheme which works fine running a Modbus RTU
Master with QNX. The only disadvantage is that you have to access the
UART registers directlty.
Kind Regards,
Henrik W. Maier
Senior Software Engineer
Henrik W. Maier Diplom-Ingenieur (Professional Engineer)
Powercorp Pty Ltd
3406 Export Drive
Trade Development Zone
Northern Territory 0822 Australia
Tel: +61 (0)8 8947 0933
Fax: +61 (0)8 8947 0925
Email: henrik@pcorp.com.au
Web Site: www.pcorp.com.au
Stefan Scherrer wrote:
Thanks,
for all hints. The most stable version I found was setting ticksize to the
specified value und in the first int outputing the byte and in the next
switching of rts >mfg,
Stefan
Hi Jim,
I am Florian from Centura in France, I am in charge of embedded systems
and information appliances. I read you were looking for a DBMS on
Neutrino.(for the story I used to work for Empress!)
Centura has bought Centura Corporation last year and has got two DBMS
which are RDM5.0 and Velocis3.0. Both are based on a combined models,
network and relational.
RDM has a really small footprint (between 50 to 250 Ko)and low level C
and C++ API to give total control to developper. RDM is not available
on Neutrino but very easy to port as long as ANCI C source code is
available. We have already done many port such
VxWorks/LynxOS/QNX/Nucleus and even on proprietary OS. This DBMS is
used by many companies in the world has a black box.
Velocis is the client/Server version with high level API.
If you wish more information, please do not hesitate to contact me my
mail is florian.michaud@centurasoft.com.
Florian
\
David Ramos García <david_ramos@tempel.es> wrote in message
news:3962D90C.5CE165CA@tempel.es…
Dear all,
We are working in a project with C&T65545 VGA card (PC-104) and
we have several troubles to work properly with this VGA. In Photon we
can only work in 16 Colors and 640480. We´ll need at least 256 Colors
and 800600.In other project we work without troubles with C&T65550 and
C&T69000, but we need this card for this project. Does anybody know
anything about this trouble???THANKS ALL. DAVID.
Hello David,
I also have a project that is using the C&T 65545 on our processor board
(driving a 640x480 LCD) and need 256 colours. When I ran the Qnx DemoDisk,
it only supported 16 colours also. I assumed the full QNX would have source
code to update the Photon for less common chipsets. Is it true that QNX
does not support more than 16 colours on 65545 chips? This chip is built
into the WinSystem board we are using in our first generation product and we
can’t switch chips.
I read in XFree86 that the 65545 is supported which means we could look for
a XFreeQNX version or switch compleatly to Linux, but I would prefer to try
QNX if it supports the hardware.
Anybody have a suggestions?
Kevin Rutten
BBR Technologies Inc.
Hi,
I have recently worked with the Advantech’s single board PC (Biscuit line)
that had C&T65550. At the beginning I also had the same and very persistent
problem of getting video resolutions higher than 640x480. What saved me was
one DOS BIOS utility program that came with the PC, and which allowed me to
specify CRT output only (I believe that, by default, C&T video chips are set
both for CRT and flat monitors).
Regards,
Zeljko
“Kevin Rutten” <krutten@bbr.ca> wrote in message
news:mWZf5.9736$47.186445@news.bc.tac.net…
David Ramos García <> david_ramos@tempel.es> > wrote in message
news:> 3962D90C.5CE165CA@tempel.es> …Dear all,
We are working in a project with C&T65545 VGA card (PC-104) and
we have several troubles to work properly with this VGA. In Photon we
can only work in 16 Colors and 640480. We´ll need at least 256 Colors
and 800600.In other project we work without troubles with C&T65550 and
C&T69000, but we need this card for this project. Does anybody know
anything about this trouble???THANKS ALL. DAVID.
Hello David,
I also have a project that is using the C&T 65545 on our processor board
(driving a 640x480 LCD) and need 256 colours. When I ran the Qnx
DemoDisk,
it only supported 16 colours also. I assumed the full QNX would have
source
code to update the Photon for less common chipsets. Is it true that QNX
does not support more than 16 colours on 65545 chips? This chip is built
into the WinSystem board we are using in our first generation product and
we
can’t switch chips.I read in XFree86 that the 65545 is supported which means we could look
for
a XFreeQNX version or switch compleatly to Linux, but I would prefer to
try
QNX if it supports the hardware.Anybody have a suggestions?
Kevin Rutten
BBR Technologies Inc.
J?rg Kampmann <joerg.kampmann@ibk-consult.de> wrote:
this seems to be a very bad issue. I did it also once and - failed.
so you are not alone >> (
I posted an answer to this question somewhere else, but don’t remember
where.
In summary, don’t have a “/=4,f” prefix. Mount everything explicitly –
your ramdisk as /ramdisk, your harddrive partition, hd0t77 say, as /hd.
Then alias / appropriately –
/ramdisk/bin/prefix -R /=/ramdisk/
and
/hd/bin/prefix -R /=/hd/
Also, you can’t run prefix (utility) from a prefix (name) that is going away.
-David
Joel Hannan schrieb:
My question is, After prefixing another device
to /, how can I change the prefix the t77 partition
back to /.I have a task that requires me to disassociate the
hard drive and work from a ramdisk as root. When I’m
done I need to prefix root back to the hard drive -
but it doesn’t work.What I am doing is essentially:
/ramdisk/bin/prefix -R /=/ramdisk/ #prefix the ramdisk
mount /dev/hd0.0t77 /hd #make a mount point to reference
the hd
#Do a bunch of stuff . . .
prefix -R /=/hd/ #re-prefix the hd (!doesn’t really
work!)After this last command I am left in limbo with a
severely crippled directory structure. I have tried
several other points to prefix to (/ramdisk/hd, /dev/hd0.0t77,
etc.), but none restore the prefix correctly.I have had some success with prefix -R /=/ramdisk/hd/
but later, when I clean out the ramdisk I loose the prefix
point.How can I get / back from this condition
/=/ramdisk/
to an original condition
/=4,f
?Thanks
–
Joerg Kampmann
IBK-Consult - (embedded Systems)
WWW: > http://www.ibk-consult.de
Mario Charest wrote:
Is there a driver that will allow me to use the PS2 port as a
serial port under QNX4.22?
No
Actually, Input uses the alternate PS/2 port as a serial port (for either
keyboard or mouse). I would think that if you asked QSSL they might be
willing to give you the source. The Dev.ser source is public, and with
a little bit of hacking you should have a (limited capability) serial
port, over the PS/2 aux channel.
Note, the port is quite limited wrt baud rates, control lines etc…
My application definitely works when you install it on a normal desktop
with
two onboard serial ports. Is there any advice you could give me on
setting
up a second serial port on a laptop so that it appears as “comm2” to my
application. Is there a locally available PCMCIA card which provides a
serial port and has drivers for QNX4.22? Do you perhaps know of a locally
available laptop that comes with two “physical” serial ports onboard.Yes I’v seen som PCMCIA card that adds serial port and work with QNX
(from BlackBox). But 4.22 doesn’t have PCMCIA support, you might
be able to use the on in 4.25.
If your application is low volume (or the limited capability of the PS/2 port
is a problem) a PCMCIA serial port is your best bet. Any “dumb” serial port
device should work fine with QNX.
–
To succeed in this world it is not enough to be stupid; you must also be
well-mannered. - Voltaire
Tom,
You can go to http://www.stlport.org and down load the latest version. STL4.0
You will need to fuss-about with the stl_watcom.h if you are using watcom…
Also, if you are using Watcom, you will get a lot of E094 fatal out of mem
errors… Other than splitting up your c++ files, I cannot seem to cure this one,
not even with WCGMEM… so if you use it keep your contruct small or use lots of
little c++ files. compile a lot, and with you see the E094 error, split your
c++ files up even more.
Hope this helps…
On Wed, 12 Jul 2000, Tom Freedy wrote:
Hello,
Does anyone know of a STL library for
Watcom/QNX 4.25c?Regards
Tom Freedy.
Previously, Ed Chapman wrote in comp.os.qnx:
Tom,
You can go to > http://www.stlport.org > and down load the latest version. STL4.0
You will need to fuss-about with the stl_watcom.h if you are using watcom…
Also, if you are using Watcom, you will get a lot of E094 fatal out of mem
errors… Other than splitting up your c++ files, I cannot seem to cure this one,
Is this related to the debugging format used? I can’t recall if dwarf format
(-gd3) is needed just for debugging stl code, or when compiling as well.
Sam
not even with WCGMEM… so if you use it keep your contruct small or use lots of
little c++ files. compile a lot, and with you see the E094 error, split your
c++ files up even more.Hope this helps…
On Wed, 12 Jul 2000, Tom Freedy wrote:
Hello,Does anyone know of a STL library for
Watcom/QNX 4.25c?Regards
Tom Freedy.
–
Sam Roberts (sam@cogent.ca), Cogent Real-Time Systems (www.cogent.ca)
Sam Roberts <sam@cogent.ca> wrote:
Previously, Ed Chapman wrote in comp.os.qnx:
Tom,
You can go to > http://www.stlport.org > and down load the latest version. STL4.0
You will need to fuss-about with the stl_watcom.h if you are using watcom…
Also, if you are using Watcom, you will get a lot of E094 fatal out of mem
errors… Other than splitting up your c++ files, I cannot seem to cure this one,Is this related to the debugging format used? I can’t recall if dwarf format
(-gd3) is needed just for debugging stl code, or when compiling as well.
That would be “-g3d”. The debug information usually cause a segment error,
some message about segment size exceeding 64k, and it usually happened at
link time – not at compile time. (Possibly always, I can only remember
hearing about it as a link error.)
-David
David Gibbs wrote:
Sam Roberts <> sam@cogent.ca> > wrote:
Previously, Ed Chapman wrote in comp.os.qnx:
Tom,
You can go to > http://www.stlport.org > and down load the latest version. STL4.0
You will need to fuss-about with the stl_watcom.h if you are using watcom…
Also, if you are using Watcom, you will get a lot of E094 fatal out of mem
errors… Other than splitting up your c++ files, I cannot seem to cure this one,Is this related to the debugging format used? I can’t recall if dwarf format
(-gd3) is needed just for debugging stl code, or when compiling as well.That would be “-g3d”. The debug information usually cause a segment error,
some message about segment size exceeding 64k, and it usually happened at
link time – not at compile time. (Possibly always, I can only remember
hearing about it as a link error.)-David
I had it appear as a compile error.
Happened when I was using SWIG generated source files, approx 1.4 Meg.
The -g3d eventually stopped working too, when the source files grew more, so
if you’re not already close to the max anticipated size of a source file, -g3d
is not a permanent solution.
You have to split the file up.
-g3d just allows you to wait longet before the file split is required.
-Bruce.
Previously, Jim Lambert wrote in comp.os.qnx:
All of this has really got me interested in why a disk defragmenter is such
a hard program to write. Isn’t it as simple as just rearranging the files
so all the blocks are continuous? I would have thought that it would be
relatively easy for someone to write a disk defragmenter especially in QNX.
Wouldnt the only thing you have to worry about be the boot files and any
application files that expect a certain block (a crude copy protection)?
Nope. You also have to worry about the cached information that Fsys
keeps. As time went on, Fsys started caching (in multiple ways) heavier
and heavier. If you make changes under the feet if Fsys it tries hard
to invalidate its cache, and I think it will catch everything, but …
Come on all you QNX gurus, give me a disk defragmenter already! >
Someday when I’m so rich I don’t have to work for a living and I feel
like having fun I’ll pull out all the notes of things I meant to do
while at QSSL and work on them. Someday …
New Year’s Resolution: SaddleSore 1000 ™ or bust!
http://www.ironbutt.com
Previously, Mike Taillon wrote in comp.os.qnx:
fragmentation doesn’t affect the qnx4 filesystem as much as other filesystems.
on qnx.com server, the most fragmented file (last time a checked) was /.inodes,
which isn’t really a file (its not open()'d, lseek()'d directly usually) with
only 285 fragments (which could have been advoided using dinit -i)
So what is the 2nd most fragmented file? As you noted, it really isn’t
fair to use /.inodes as a measure of anything.
New Year’s Resolution: SaddleSore 1000 ™ or bust!
http://www.ironbutt.com
Previously, Robert Krten wrote in comp.os.qnx:
Perhaps not “as much as other filesystems”, but extent fragmentation still presents
a noticeable hit. You may ask why I get extent fragmentation >> I download
a whole whack of data, and append one line to about 8,000 files each day.
Thus, each of the 8,000 data files grows by one text line every day >
gets real messy real fast. So on Saturday mornings I run a “find/cp/rm/mv”
script to at least defragment the extents. It was quite noticeable. >
This is the case for which the “persistant pregrown files” were
implemented. I think they appeared in 4.25A. Are you sure you aren’t
mounting with the ‘-g’ option?
New Year’s Resolution: SaddleSore 1000 ™ or bust!
http://www.ironbutt.com
William A. Flowers (waflowers@insightcontrol.com) wrote:
: Previously, Robert Krten wrote in comp.os.qnx:
: > Perhaps not “as much as other filesystems”, but extent fragmentation still presents
: > a noticeable hit. You may ask why I get extent fragmentation I download
: > a whole whack of data, and append one line to about 8,000 files each day.
: > Thus, each of the 8,000 data files grows by one text line every day This
: > gets real messy real fast. So on Saturday mornings I run a “find/cp/rm/mv”
: > script to at least defragment the extents. It was quite noticeable.
: This is the case for which the “persistant pregrown files” were
: implemented. I think they appeared in 4.25A. Are you sure you aren’t
: mounting with the ‘-g’ option?
I am definitely mounting with the -g option. The problem with mounting
without the -g option is that my brand new, empty, 23G disk ran out of disk
space because of all the pregrown files. There was apparently no recovery
code that would go back and reclaim the pregrown data areas – once I
mounted with the -g and did a chkfsys, the disk went back to being about
1/2 full again (which is what the actual data space usage was).
I’m not trying to crap all over the -g/non-g idea, but at the time that
I tried it it definitely had some problems
Cheers,
-RK
–
Robert Krten, PARSE Software Devices; email my initials at parse dot com
Consulting, Systems Architecture / Design, Drivers, Training, QNX 4 & Neutrino
Check out our new QNX 4 and Neutrino books at http://www.parse.com/
+1 613 599 8316 or toll free +1 877 PARSE-SW (727-7379)
Previously, Robert Krten wrote in comp.os.qnx:
I am definitely mounting with the -g option. The problem with mounting
without the -g option is that my brand new, empty, 23G disk ran out of disk
space because of all the pregrown files. There was apparently no recovery
code that would go back and reclaim the pregrown data areas – once I
mounted with the -g and did a chkfsys, the disk went back to being about
1/2 full again (which is what the actual data space usage was).
I thought chkfsys would recover the space regardless of mount option.
Maybe not.
You must have some enormous number of files which are growing that way
to cause Fsys to consume all that space. Hmmm, I thought there was a
limit on the pre-grown allocation per file.
My problem is I am working from my admittedly faulty memory with no way
to check the details. And even if I could check, there is no way I
could do anything about it. In fact, you are probably in a better
position to do something about it than I – at least you can go in
person and complain to the people now resposible. I’d have a long trip
to make first.
I’m not trying to crap all over the -g/non-g idea, but at the time that
I tried it it definitely had some problems >
It was a rather new concept when I left and probably could have used
some refinements. It is unfortunate that no one appears to have taken
the concept further (e.g. automatic reclamation of preallocated space if
the disk starts getting full).
New Year’s Resolution: SaddleSore 1000 ™ or bust!
http://www.ironbutt.com
William A. Flowers <waflowers@insightcontrol.com> wrote:
Previously, Mike Taillon wrote in comp.os.qnx:
fragmentation doesn’t affect the qnx4 filesystem as much as other filesystems.
on qnx.com server, the most fragmented file (last time a checked) was /.inodes,
which isn’t really a file (its not open()'d, lseek()'d directly usually) with
only 285 fragments (which could have been advoided using dinit -i)So what is the 2nd most fragmented file? As you noted, it really isn’t
fair to use /.inodes as a measure of anything.
sorry, don’t remember the details, but am pretty sure the next most
fragmented file didn’t even have half what .inodes had – was a big archive
under /updates.
(Just wondering how well things are still working.)
considering we (purposely) keep it at least 98% full
all the time, i would say fragmentation is minimal…
–
William A. (Bill) Flowers Phone: +1-727-669-9999
Insight Control Systems Fax: +1-727-669-8999New Year’s Resolution: SaddleSore 1000 ™ or bust!
http://www.ironbutt.com
Previously, Mike Taillon wrote in comp.os.qnx:
sorry, don’t remember the details, but am pretty sure the next most
fragmented file didn’t even have half what .inodes had – was a big archive
under /updates.
I can believe it.
considering we (purposely) keep it at least 98% full
all the time, i would say fragmentation is minimal…
Not perfect, but not bad either.
New Year’s Resolution: SaddleSore 1000 ™ or bust!
http://www.ironbutt.com
At some point, a generic SCSI driver will need to be written or ported to
QnX. Once this happens,
then cdrecord will be usable under QnX. The generic driver allows much
SCSI control to be exported
into user space, allowing exotic devices (like cd burners) to be supported
without having to muck about with the vendor-specific drivers. It’s not a
generalized solution - ill-behaved applications can
lock up your scsi bus. Still, it’s better than recoding support for such
devices into each and every
vendor-specific driver.
vboyko@my-deja.com wrote:
An approach is to use Windo$e to burn your CD.
Get the mkisofs from the QNX free S/W area.
It creates a file much like tar does, but in ISO9660 format suitable
for CD burning.Steps:
- Create image in QNX of files and directories using mkisofs.
- Copy image to windo$e machine, via FTP, or if on same machine to the
DOS partition.- Burn CD in windo$e using the image.
Vince
In article <8j8d3b$1s5$> 1@inn.qnx.com> >,
“Scott Merz” <> scottm@gorge.net> > wrote:
hii was wondering if anybody has heard of software used for utilizing a
cd
burner in qnx…im talkin software that would allow you to burn cd’s
from
inside qnx w/ qnx filesif anybody has any info on this
id appreciate itScott Merz
Solution Engineering
Sent via Deja.com > http://www.deja.com/
Before you buy.
In article <8nc28j$bvq$1@nnrp1.deja.com>,
cateyes5612@my-deja.com wrote:
What is the difference between the -L and -rpath-link options
in ld? Thanks.Sent via Deja.com > http://www.deja.com/
Before you buy.
I mean qcc’s -L flag and ld’s -rpath-link.
Sent via Deja.com http://www.deja.com/
Before you buy.