Linux DRI (and Google Earth)

[Andrea Venturoli]

Hello.

What's the status of 3d hardware acceleration in Linux emulation?
I'm running 8.1R/i386 and I was finally able to get DRI working (with 
native software) on my Radeon HD 4200.
So I installe Google Earth, but it's warning that it will use software 
rendering and is, of course, slow.


%pkg_info|grep linux
linux-f10-dri-7.2_1 Mesa libGL runtime libraries and DRI drivers (Linux 
Fedora
linux-f10-expat-2.0.1 Linux/i386 binary port of Expat XML-parsing 
library (Linux
linux-f10-fontconfig-2.6.0 An XML-based font configuration API for X 
Windows (Linux Fe

linux-f10-xorg-libs-7.4_1 Xorg libraries (Linux Fedora 10)
linux_base-f10-10_2 Base set of packages needed in Linux mode for 
i386/amd64 (L


I had linux-dri-7.4_1 installed by default, but tried switching to 
linux-f10-dri-7.2_1; nothing changed.


 bye & Thanks
av.
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

sysinstall vs gmirror

[perryh]

How do I get sysinstall to recognize a gmirror?

I've created the mirror -- which currently has only one provider --
using Fixit#, followed by

Fixit# ln -s /dist/boot/kernel /boot
Fixit# gmirror load

after which /dev/mirror/gm0{,a,b} exist.  However, even after
rescanning the disks, sysinstall doesn't include gm0 in its
drive list.  I also tried:

Fixit# ( cd /dev && ln -s mirror/* . && ll gm* )
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 gm0@ -> mirror/gm0
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 gm0a@ -> mirror/gm0a
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 gm0b@ -> mirror/gm0b

in case sysinstall looks only in /dev itself and not in any
subdirectories, and that didn't help.  I even tried:

Fixit# ( cd /dev && ln -s mirror/gm0 ar0 \
   && for p in a b d e ; \ 
   do ln -s mirror/gm0$p ar0$p ; done && ll ar* )
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 ar0@ -> mirror/gm0
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 ar0a@ -> mirror/gm0a
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 ar0b@ -> mirror/gm0b
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 ar0d@ -> mirror/gm0d
lrwxr-xr-x  1 root  0  10 Sep  6 10:48 ar0e@ -> mirror/gm0e

in case sysinstall looks only for names of known disk drivers,
and that didn't help either.
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: gs-8-8.71 under 8.1-Release missing x11 devices

[Michael Powell]

Jin Guojun[VFF] wrote:

> gs 8-8.71 under FreeBSD 8.1-R seems missing x11 device.
> When use ghostview, it complains "/unknown device x11"
> 
> /By tracing around, I found it was caused by gs 8-8.71. As typing "gs
> --help", it shows much less
> devices supported than gs 8-8.62 under FreeSBD 6.4-R.
> 
> By searching on the Internet, one message says that this could be
> resulted by build config.
> Is this true? or can gs be dynamically configured to use x11 device?

Possibly, if the module was built at compile time when selected from the 
make config list. However, I suspect it is not needed and just in the way. 
There may be a .conf file somewhere where you could tell it not to load the 
X11 modules even though they may have been built.
 
> Hopefully, users do not have to recompile ghostscript.

When you run make config in the ghostscript port you should get a list with 
checkboxes to set build configuration. Unless there is some direct need you 
might consider clearing the X11 checkbox(es) and recompiling with make, then 
make deinstall, followed by make reinstall. 

If you want to completely clear the build config simply do make rmconfig and 
it will remove previously saved options. Then make should present you with 
the build config option screen with default options preselected. Doing make 
config allows to pull up the saved options for adjustment as needed. 

But the short answer is you probably need to rebuild the port without the 
X11 modules which are producing your errors and probably not needed anyway.

-Mike
 



___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: Upgrading packages - portupgrade confusion

[Michael Powell]

Kaya Saman wrote:

> Hi,
> 
> I have 2 servers one production and another test.
> 
> The test machine's packages however, seem to be older then the
> production machines one's even though I built the production system a
> few months ago.
> 
> I used the: portupgrade command in order to try to upgrade the ports nad
> re-install the packages only the same versions seem to be compiling???
> 
> I ran: portupgrade -ai
> 
> on the base system as the system where these packages are installed into
> is a FreeBSD jail.
> 
> The ports in question are these:
> 
> tomcat-6.0.29   Open-source Java web server by Apache, 6.x branch
> postgresql-client-8.2.17_1 PostgreSQL database (client)
> postgresql-server-8.2.17_1 The most advanced open-source database
> available anywhere
> 
> Which on my newer test system show up as such:
> 
> postgresql-client-8.2.13 PostgreSQL database (client)
> postgresql-server-8.2.13 The most advanced open-source database
> available anywhere
> tomcat-6.0.20_1 Open-source Java web server by Apache, 6.x branch
> 
> I don't understand this 100%???
> 
> I would like the versions to be the same as the production system since
> I have a postgres-Tomcat connector which doesn't work on the test setup
> as my Tomcat webapp isn't being displayed!!
> 
> Can I do anything about this??
> 
> I don't even know why it is like this although I must admit that it has
> been an exceptionally long day and am really suffering from fatigue now
> which might be a contributor but I can't tell.
> 
> Can anyone give me any advise??
> 

Have you refreshed the ports tree(s) with csup using the same supfile to 
ensure the ports trees are up to date ( and therefore identical)? Since you 
are using portugrade, as I do, this is what I do to see what needs to be 
done:

I cd to /usr/sup which is where I keep my supfiles and the housekeeping. 
Then using this command sequence will refresh the ports tree, the ports 
index database, and ensure the package database is clean and synced. 
Portversion then just tells you with a "<" symbol any that are old and in 
need of an update.

csup -L 2 ports && portsdb -uF && pkgdb -u && portversion

where "ports" above is my supfile for ports refresh and looks like this:

*default host=cvsup.nl.freebsd.org
*default base=/usr
*default prefix=/usr
*default release=cvs tag=.
*default delete use-rel-suffix compress
ports-all

Then a portupgrade -a as required. If all symbols in the right column are 
"=" everything is up to date and nothing is required. Adjust server location 
for mirror near you (or one that works best).

-Mike



___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

portupgrade -a stops at building gnome-menus

[Michael D. Norwick]

FreeBSD 8.1-RELEASE

#>uname -a
FreeBSD *...@.net 8.1-RELEASE FreeBSD 8.1 RELEASE FreeBSD 8.1-RELEASE
#0: Thu Aug 12 08:43:46 CDT 2010 
*...@.net:/usr/obj/src/sys/GENERIC i386


Running on VirtualBox Version 3.2.8 r64453 running on current Debian 
'lenny',  Pentium 4 2.4 GHz. 4G ram.

portsnap update on 09/08/2010.

Trying to do 'portupgrade -a',  initially had a portupgrade stop at 
'/usr/ports/graphviz'  an error about 'dot' and doxygen.  Built doxygen 
and graphiz from the individual /usr/ports/*** directories after 
'portsclean -DLP' and individual 'make clean' in the respective 
/usr/ports directories.  Had the "/usr/local/include/python2.6/pth.h 
link to /usr/local/include/pth/pth.h" issue.  Fixed that and portupgrade 
borked at building gnome-menus.  '/usr/bin/ld: cannot find -lpth'.  Did 
ln -s /usr/local/lib/pth/libpth.a /usr/local/lib/ and ln -s 
/usr/local/include/pth/pth.h /usr/local/include/python2.6/ - again.  Ran 
make install clean from /usr/ports/x11/gnome-menus after executing make 
clean.  Still no joy.
Yeah, I'm a 15 year linux guy, but, I've installed and used FreeBSD 
around the 5.0-RELEASE days so I don't think I'm totally clueless.  I'm 
running it as a virtual machine because I would like to install it on a 
new machine once I get past the test drive and checkout.
Tried to build a new kernel a week or two ago and that went awry.  
Deleted the VM and reinstalled from the RELEASE dvd.iso.

What am I doing wrong?

Michael D. Norwick

PS: I've R.T.F.M'd and Googled.  Filed a bug report on the graphviz 
issue but now I don't think it was a problem with the graphviz build.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Upgrading packages - portupgrade confusion

[Kaya Saman]

Hi,

I have 2 servers one production and another test.

The test machine's packages however, seem to be older then the 
production machines one's even though I built the production system a 
few months ago.


I used the: portupgrade command in order to try to upgrade the ports nad 
re-install the packages only the same versions seem to be compiling???


I ran: portupgrade -ai

on the base system as the system where these packages are installed into 
is a FreeBSD jail.


The ports in question are these:

tomcat-6.0.29   Open-source Java web server by Apache, 6.x branch
postgresql-client-8.2.17_1 PostgreSQL database (client)
postgresql-server-8.2.17_1 The most advanced open-source database 
available anywhere


Which on my newer test system show up as such:

postgresql-client-8.2.13 PostgreSQL database (client)
postgresql-server-8.2.13 The most advanced open-source database 
available anywhere

tomcat-6.0.20_1 Open-source Java web server by Apache, 6.x branch

I don't understand this 100%???

I would like the versions to be the same as the production system since 
I have a postgres-Tomcat connector which doesn't work on the test setup 
as my Tomcat webapp isn't being displayed!!


Can I do anything about this??

I don't even know why it is like this although I must admit that it has 
been an exceptionally long day and am really suffering from fatigue now 
which might be a contributor but I can't tell.


Can anyone give me any advise??


Many thanks and best regards,


Kaya
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

xvidtune and nouveau video driver

[Arthur Barlow]

Ever since Debian went to the "nouveau" video driver for Nvidia, I have not
been able to adjust my horizontal screen position with xvidtune. The
application runs, but when I try to reset the "HSyncStart" value I get an
error dialog box that says, "Sorry: You have requested a mode-line that is
not possible, or not supported by your hardware configuration."

This didn't happen with the old "nv" driver.  Does anyone know a fix other
than reinstalling "nv"?
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

gs-8-8.71 under 8.1-Release missing x11 devices

[Jin Guojun[VFF]]

gs 8-8.71 under FreeBSD 8.1-R seems missing x11 device.
When use ghostview, it complains "/unknown device x11"

/By tracing around, I found it was caused by gs 8-8.71. As typing "gs 
--help", it shows much less

devices supported than gs 8-8.62 under FreeSBD 6.4-R.

By searching on the Internet, one message says that this could be 
resulted by build config.

Is this true? or can gs be dynamically configured to use x11 device?

Hopefully, users do not have to recompile ghostscript.

-Jin
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: Compiling software with different compiler than cc or clang results in unusable output

[Mehmet Erol Sanliturk]

On Sat, Sep 11, 2010 at 10:05 AM, O. Hartmann <
ohart...@mail.zedat.fu-berlin.de> wrote:
>
> Hello.
> Well, the only other architectures I have access to are Linux boxes.
>
> clang ist a very nice compiler since its syntax checking is formidable. But
> its code is slow and there seems no OpenMP support at the moment.
>
> Oliver
>
>
The following pages may be useful :

www.oracle.com/technetwork/server-storage/solarisstudio/downloads/index-jsp-141149.html

www.oracle.com/technetwork/server-storage/solarisstudio/downloads/index-jsp-136197.html

www.oracle.com/technetwork/server-storage/solarisstudio/documentation/express-june2010-137081.html
( Please notice Support for OpenMP 3.0 features in the C, C++, and Fortran
compilers:  )

This means that you may use Oracle Solaris Studio on Linux  with OpenMP 3.0
support
immediately .

I do not know whether they can be used in FreeBSD as Linux programs or not ,
because I did not study such a  possibility . For me , using Linux directly
is easy .

Thank you very much .


Mehmet Erol Sanliturk
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

apropos returning same item twice

[Steven Friedrich]

Why does apropos list mysql(1) twice?

It doesn't return duplicates with apropos kde...




ad...@laptop2(/dev/pts/1)/usr/home/admin 106% apropos mysql 
   
mysql(1) - the MySQL command-line tool  
   
mysql.server(1)  - MySQL server startup script
mysql_config(1)  - get compile options for compiling clients
mysql_install_db(1)  - initialize MySQL data directory
mysql_tzinfo_to_sql(1)   - load the time zone tables
mysql_upgrade(1) - check tables for MySQL upgrade
mysql_waitpid(1) - kill process and wait for its termination
mysqladmin(1)- client for administering a MySQL server
mysqlbinlog(1)   - utility for processing binary log files
mysqlbug(1)  - generate bug report
mysqlcheck(1)- a table maintenance program
mysqld_safe(1)   - MySQL server startup script safe_mysqld - MySQL 
server startup script
mysqldump(1) - a database backup program
mysqlimport(1)   - a data import program
mysqlshow(1) - display database, table, and column information
mysqltest(1) - program to run test cases mysqltest_embedded - 
program to run embedded test cases
slapd-ndb(5) - MySQL NDB backend to slapd
mysql(1) - the MySQL command-line tool
mysql.server(1)  - MySQL server startup script
mysql_config(1)  - get compile options for compiling clients
mysql_install_db(1)  - initialize MySQL data directory
mysql_tzinfo_to_sql(1)   - load the time zone tables
mysql_upgrade(1) - check tables for MySQL upgrade
mysql_waitpid(1) - kill process and wait for its termination
mysqladmin(1)- client for administering a MySQL server
mysqlbinlog(1)   - utility for processing binary log files
mysqlbug(1)  - generate bug report
mysqlcheck(1)- a table maintenance program
mysqld_safe(1)   - MySQL server startup script safe_mysqld - MySQL 
server startup script
mysqldump(1) - a database backup program
mysqlimport(1)   - a data import program
mysqlshow(1) - display database, table, and column information
mysqltest(1) - program to run test cases mysqltest_embedded - 
program to run embedded test cases
slapd-ndb(5) - MySQL NDB backend to slapd

-- 
System Name:   laptop2.StevenFriedrich.org
Hardware:  2.80GHz Intel Pentium 4 (HTT) with 2 GB memory
OS version:FreeBSD 8.1-RELEASE i386 (6.9 MB kernel)
manager(s):kde4-4.5.1 
X windows: xorg-7.5X.Org X Server 1.7.5
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: How to Best Prevent Unwanted named installation

[Arthur Chance]

On 09/10/10 21:58, Martin McCormick wrote:

After successfully installing bind97 from a package on
to a new server, I do a cvs-sup of the system to get the latest
patches in to the kernel. After discovering that bind97 had been
replaced with bind9.6.1, I looked in /usr/src and there is a
contrib/bind9 directory. What is the safest way to disable that
build without adversly effecting the rest of the update?

The reason for doing these things in this order is that
I would like to get bind running as quickly as possible since it
takes a couple of hours or more to get the world built when we
could be doing DNS.

Since I am not using that version of bind, not getting
it built is no problem. I don't even care if it gets built so
long as it does not end up in /usr/sbin to clobber the new
bind9.7.


If your ports version of named is in /usr/sbin you must have enabled the 
REPLACE_BASE option in the port. From man src.conf



 WITHOUT_BIND
 Setting this variable will prevent any part of BIND from being
 built.  When set, it also enforces the following options:

[list of sub options snipped]

Add

WITHOUT_BIND= true

into /etc/src.conf, and the next time you rebuild the world the base 
system bind will be left out of it.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: gjournal+geli

[Adam Vande More]

On Sat, Sep 11, 2010 at 9:25 AM, RW  wrote:

>
> I'm planning to use gjournal+geli on a 2TB drive in a USB enclosure.
>
> What I've read about this suggest that the order should be:
> geli-gjournal-ufs. I was wondering if it's possible to do it in
> the order gjournal-geli-ufs, which should be much more efficient. I've
> read that ufs should go directly on gjournal, but I just wanted to
> check that that is needed.
>
> I was also wondering about the journal size, and whether there are any
> performance optimizations to be made to mitigate the extra
> encryption/decryption in the journal. The man page suggests a size of at
> least 2xmemory which would be 2x1.5GB now, or maybe 2x16GB to allow for
> potential upgrades. It seems very large. The disk will hold fairly
> static data so it will be mostly be long sustained writes as files are
> copied in. Currently coping from geli to geli with soft-updates is
> slightly cpu limited.
>

AFAIK, ufs must be on top of gjournal.  Specific changes were made to allow
ufs to be aware of the journal and I think sticking geli in-between would
destroy that relationship.

IME, gjournal is more sensitive to load as the man page also suggests.  I
have one production server with moderate load, and a 5 GB problem.  I think
something like 5 -10 GB journal would be more than enough for almost all
loads, but that's just a guess.  It's easy to test though, just run
blogbench or some other io benchmark for a sustained period of time.  If it
doesn't panic, you're golden.

-- 
Adam Vande More
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: Compiling software with different compiler than cc or clang results in unusable output

[O. Hartmann]

On 09/11/10 14:26, Mehmet Erol Sanliturk wrote:



On Sat, Sep 11, 2010 at 5:43 AM, Andrew Brampton
mailto:brampton%2bfree...@gmail.com>> wrote:

On 11 September 2010 10:28, O. Hartmann
mailto:ohart...@mail.zedat.fu-berlin.de>> wrote:
 >
 > Dear Sirs,
 >
 > you see me a kind of desperate. I wrote my own a small piece of
 software in
 > C, calculating the orbit and position of astronomical objects,
astroids, in
 > a heliocentric coordinate system from Keplerian orbital elements.
So far.
 > The software calculates the set of points of an ellipse based upon
 > ephemeridal datas taken from the Minor Planet Cataloge. Again, so
far,
 > everything all right. The set of points of an orbit is all right and
 > correct. But when it comes to positions at a specific time, then
I loose
 > hair!
 >
 > Compiling this piece of software with FreeBSD's gcc (V4.2) and
clang (clang
 > devel) on my private and lab's FreeBSD boxes (both most recent
FreeBSD
 > 8.1/amd64), this program does well, the calculated orbital
positions are
 > very close to professional applications or observational checks.
But when
 > compiling the sources with gcc44 or gcc45 (same source, same
CFLAG setting,
 > mostly no CFLAGS set), then there is a great discrepancy.
Sometimes when
 > plotting positions, the results plotted seconds before differs
from the most
 > recent. The ellipses are allways correct, but the position of a
single point
 > at a specific time isn't correct.
 >
 > I use the GNU autotools to build the package.
 >
 > I suspekt miscompilations in memory alloction or in some time- or
 > mathematical functions like sin, cos.
 >
 > before I digg deeper I'd like to ask the community for some hints
how to
 > hunt down such a problem.
 >
 > regards,
 > Oliver

Sounds a cool project. I suspect you are miss-using a feature of C or
are using uninitialised memory, and with gcc44/45's more aggressive
optimisations it is getting it wrong. I have three suggestions

1) Use valgrind to check if it finds anything wrong when running your
program. Check both the good and the bad builds.

2) If your program is made up of multiple C files, then try compiling
all of the C files with gcc42, but just one at a time with gcc44. This
way will help you track down exactly which C file has "the bug".

3) Finally do some printf debugging to find the first line of code
that is generating the wrong value.

I hope these suggestions help.
Andrew



Another check may be to use Sun Studio C and or Fortran  compilers .
These can be used in Linux ( Linux version of Sun Studio )Â  and/or
OpenSolaris or Solaris ( Solaris version of SunStudio ( both in x86 ,
x86_64 , Sparc )Â  ( all of them are ( Solaris , OpenSolaris , Sun
Studio , Linux  )  free ) . All of them are freely downloadable from
www.sun.com  and/or www.opensolaris.com
 ( these sites or their pages may be
redirected to www.oracle.com  owned pages ) .

Personally I tried GCC compilers , but I found that they are very
unreliable . Now I am using Sun Studio compilers in OpenSolaris and
Linux , and never GCC compilers .Â

Thank you very much .

Mehmet Erol Sanliturk



Â


Hello.
Well, the only other architectures I have access to are Linux boxes.

clang ist a very nice compiler since its syntax checking is formidable. 
But its code is slow and there seems no OpenMP support at the moment.


Oliver

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: How to Best Prevent Unwanted named installation

[RW]

On Fri, 10 Sep 2010 15:58:42 -0500
Martin McCormick  wrote:

>   After successfully installing bind97 from a package on
> to a new server, I do a cvs-sup of the system to get the latest
> patches in to the kernel. After discovering that bind97 had been
> replaced with bind9.6.1, 

Presumably that's because you explicitly configured the port version to
install in the same place as the system version. It doesn't do that by
default.
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

gjournal+geli

[RW]

I'm planning to use gjournal+geli on a 2TB drive in a USB enclosure.

What I've read about this suggest that the order should be:
geli-gjournal-ufs. I was wondering if it's possible to do it in
the order gjournal-geli-ufs, which should be much more efficient. I've
read that ufs should go directly on gjournal, but I just wanted to
check that that is needed.

I was also wondering about the journal size, and whether there are any
performance optimizations to be made to mitigate the extra
encryption/decryption in the journal. The man page suggests a size of at
least 2xmemory which would be 2x1.5GB now, or maybe 2x16GB to allow for
potential upgrades. It seems very large. The disk will hold fairly
static data so it will be mostly be long sustained writes as files are
copied in. Currently coping from geli to geli with soft-updates is
slightly cpu limited.



___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: Compiling software with different compiler than cc or clang results in unusable output

[O. Hartmann]

On 09/11/10 11:43, Andrew Brampton wrote:

On 11 September 2010 10:28, O. Hartmann
  wrote:


Dear Sirs,

you see me a kind of desperate. I wrote my own a small piece of  software in
C, calculating the orbit and position of astronomical objects, astroids, in
a heliocentric coordinate system from Keplerian orbital elements. So far.
The software calculates the set of points of an ellipse based upon
ephemeridal datas taken from the Minor Planet Cataloge. Again, so far,
everything all right. The set of points of an orbit is all right and
correct. But when it comes to positions at a specific time, then I loose
hair!

Compiling this piece of software with FreeBSD's gcc (V4.2) and clang (clang
devel) on my private and lab's FreeBSD boxes (both most recent FreeBSD
8.1/amd64), this program does well, the calculated orbital positions are
very close to professional applications or observational checks. But when
compiling the sources with gcc44 or gcc45 (same source, same CFLAG setting,
mostly no CFLAGS set), then there is a great discrepancy. Sometimes when
plotting positions, the results plotted seconds before differs from the most
recent. The ellipses are allways correct, but the position of a single point
at a specific time isn't correct.

I use the GNU autotools to build the package.

I suspekt miscompilations in memory alloction or in some time- or
mathematical functions like sin, cos.

before I digg deeper I'd like to ask the community for some hints how to
hunt down such a problem.

regards,
Oliver


Sounds a cool project. I suspect you are miss-using a feature of C or
are using uninitialised memory, and with gcc44/45's more aggressive
optimisations it is getting it wrong. I have three suggestions

1) Use valgrind to check if it finds anything wrong when running your
program. Check both the good and the bad builds.

2) If your program is made up of multiple C files, then try compiling
all of the C files with gcc42, but just one at a time with gcc44. This
way will help you track down exactly which C file has "the bug".

3) Finally do some printf debugging to find the first line of code
that is generating the wrong value.

I hope these suggestions help.
Andrew


Hello Andrew.

Thanks for your comments, they are worth trying out. I will do so ...

item 2) oh, yes, a very good idea ...

item 3) I did already, the whole software is built up by those printf's.

The problem boiled down to be some problem in the UNIX time routines. I 
use localtime(3), time(3) and a strftime(3) and strptime(3).


I use a 'wikipedia'-algorithm converting the actual time string into an 
'epoch' used in astronomical calculations. Compiling this routine with 
gcc42 and clang everything is all right, compiling it with gcc44 or 
gcc45 it returns 10 times higher values. I use very 'primitive' cutoffs 
for casting a double value into an int - I need the integrale value, not 
the remainings after the decimal point. I will check this again and look 
forward for a cleaner solution. But isn't this a 'bug'?


I'll try the BETA of the new FreeBSD PathScale compiler if I get some.

Well, I'll report ...

Oliver

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: this is probably a little touchy to ask...

[Dave]

On 10 Sep 2010 at 18:20, Jason C. Wells wrote:

Subject:Re: this is probably a little touchy to ask...

> On 09/10/10 07:29, Randal L. Schwartz wrote:
> >
> > I repeat... Java had its day.  Time to move on.
> >
> >
> Java is not just for browsers.
> 
> Regards,
> Jason C. Wells
> 

I can't help wondering if half of you are talking about the 
"JavaScripting" language that runs in a browser, while the rest are 
talking about the "Java Run Time Engine" that some (cross platform) 
standalone app's (and some browser apps) use.

All I know is they are very different beasts.

dit dit.

Dave B.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: this is probably a little touchy to ask...

[Dave]

On 10 Sep 2010 at 18:20, Jason C. Wells wrote:

Subject:Re: this is probably a little touchy to ask...

> On 09/10/10 07:29, Randal L. Schwartz wrote:
> >
> > I repeat... Java had its day.  Time to move on.
> >
> >
> Java is not just for browsers.
> 
> Regards,
> Jason C. Wells
> 

I can't help wondering if half of you are talking about the 
"JavaScripting" language that runs in a browser, while the rest are 
talking about the "Java Run Time Engine" that some (cross platform) 
standalone app's (and some browser apps) use.

All I know is they are very different beasts.

dit dit.

Dave B.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: Compiling software with different compiler than cc or clang results in unusable output

[Mehmet Erol Sanliturk]

On Sat, Sep 11, 2010 at 5:43 AM, Andrew Brampton

> wrote:

> On 11 September 2010 10:28, O. Hartmann
>  wrote:
> >
> > Dear Sirs,
> >
> > you see me a kind of desperate. I wrote my own a small piece of  software
> in
> > C, calculating the orbit and position of astronomical objects, astroids,
> in
> > a heliocentric coordinate system from Keplerian orbital elements. So far.
> > The software calculates the set of points of an ellipse based upon
> > ephemeridal datas taken from the Minor Planet Cataloge. Again, so far,
> > everything all right. The set of points of an orbit is all right and
> > correct. But when it comes to positions at a specific time, then I loose
> > hair!
> >
> > Compiling this piece of software with FreeBSD's gcc (V4.2) and clang
> (clang
> > devel) on my private and lab's FreeBSD boxes (both most recent FreeBSD
> > 8.1/amd64), this program does well, the calculated orbital positions are
> > very close to professional applications or observational checks. But when
> > compiling the sources with gcc44 or gcc45 (same source, same CFLAG
> setting,
> > mostly no CFLAGS set), then there is a great discrepancy. Sometimes when
> > plotting positions, the results plotted seconds before differs from the
> most
> > recent. The ellipses are allways correct, but the position of a single
> point
> > at a specific time isn't correct.
> >
> > I use the GNU autotools to build the package.
> >
> > I suspekt miscompilations in memory alloction or in some time- or
> > mathematical functions like sin, cos.
> >
> > before I digg deeper I'd like to ask the community for some hints how to
> > hunt down such a problem.
> >
> > regards,
> > Oliver
>
> Sounds a cool project. I suspect you are miss-using a feature of C or
> are using uninitialised memory, and with gcc44/45's more aggressive
> optimisations it is getting it wrong. I have three suggestions
>
> 1) Use valgrind to check if it finds anything wrong when running your
> program. Check both the good and the bad builds.
>
> 2) If your program is made up of multiple C files, then try compiling
> all of the C files with gcc42, but just one at a time with gcc44. This
> way will help you track down exactly which C file has "the bug".
>
> 3) Finally do some printf debugging to find the first line of code
> that is generating the wrong value.
>
> I hope these suggestions help.
> Andrew
>
>

Another check may be to use Sun Studio C and or Fortran  compilers . These
can be used in Linux ( Linux version of Sun Studio )  and/or OpenSolaris or
Solaris ( Solaris version of SunStudio ( both in x86 , x86_64 , Sparc )  (
all of them are ( Solaris , OpenSolaris , Sun Studio , Linux  )  free ) .
All of them are freely downloadable from www.sun.com and/or
www.opensolaris.com ( these sites or their pages may be redirected to
www.oracle.com owned pages ) .

Personally I tried GCC compilers , but I found that they are very unreliable
. Now I am using Sun Studio compilers in OpenSolaris and Linux , and never
GCC compilers .

Thank you very much .

Mehmet Erol Sanliturk
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: Compiling software with different compiler than cc or clang results in unusable output

[Andrew Brampton]

On 11 September 2010 10:28, O. Hartmann
 wrote:
>
> Dear Sirs,
>
> you see me a kind of desperate. I wrote my own a small piece of  software in
> C, calculating the orbit and position of astronomical objects, astroids, in
> a heliocentric coordinate system from Keplerian orbital elements. So far.
> The software calculates the set of points of an ellipse based upon
> ephemeridal datas taken from the Minor Planet Cataloge. Again, so far,
> everything all right. The set of points of an orbit is all right and
> correct. But when it comes to positions at a specific time, then I loose
> hair!
>
> Compiling this piece of software with FreeBSD's gcc (V4.2) and clang (clang
> devel) on my private and lab's FreeBSD boxes (both most recent FreeBSD
> 8.1/amd64), this program does well, the calculated orbital positions are
> very close to professional applications or observational checks. But when
> compiling the sources with gcc44 or gcc45 (same source, same CFLAG setting,
> mostly no CFLAGS set), then there is a great discrepancy. Sometimes when
> plotting positions, the results plotted seconds before differs from the most
> recent. The ellipses are allways correct, but the position of a single point
> at a specific time isn't correct.
>
> I use the GNU autotools to build the package.
>
> I suspekt miscompilations in memory alloction or in some time- or
> mathematical functions like sin, cos.
>
> before I digg deeper I'd like to ask the community for some hints how to
> hunt down such a problem.
>
> regards,
> Oliver

Sounds a cool project. I suspect you are miss-using a feature of C or
are using uninitialised memory, and with gcc44/45's more aggressive
optimisations it is getting it wrong. I have three suggestions

1) Use valgrind to check if it finds anything wrong when running your
program. Check both the good and the bad builds.

2) If your program is made up of multiple C files, then try compiling
all of the C files with gcc42, but just one at a time with gcc44. This
way will help you track down exactly which C file has "the bug".

3) Finally do some printf debugging to find the first line of code
that is generating the wrong value.

I hope these suggestions help.
Andrew
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Compiling software with different compiler than cc or clang results in unusable output

[O. Hartmann]

Dear Sirs,

you see me a kind of desperate. I wrote my own a small piece of  
software in C, calculating the orbit and position of astronomical 
objects, astroids, in a heliocentric coordinate system from Keplerian 
orbital elements. So far. The software calculates the set of points of 
an ellipse based upon ephemeridal datas taken from the Minor Planet 
Cataloge. Again, so far, everything all right. The set of points of an 
orbit is all right and correct. But when it comes to positions at a 
specific time, then I loose hair!


Compiling this piece of software with FreeBSD's gcc (V4.2) and clang 
(clang devel) on my private and lab's FreeBSD boxes (both most recent 
FreeBSD 8.1/amd64), this program does well, the calculated orbital 
positions are very close to professional applications or observational 
checks. But when compiling the sources with gcc44 or gcc45 (same source, 
same CFLAG setting, mostly no CFLAGS set), then there is a great 
discrepancy. Sometimes when plotting positions, the results plotted 
seconds before differs from the most recent. The ellipses are allways 
correct, but the position of a single point at a specific time isn't 
correct.


I use the GNU autotools to build the package.

I suspekt miscompilations in memory alloction or in some time- or 
mathematical functions like sin, cos.


before I digg deeper I'd like to ask the community for some hints how to 
hunt down such a problem.


regards,
Oliver
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: this is probably a little touchy to ask...

[Polytropon]

On Sat, 11 Sep 2010 03:17:49 -0500, Joshua Isom  wrote:
> So to configure your router, you need a java enabled browser, and odds 
> are you get the jar file from the router, so it has an http server, and 
> probably another server just to process configuration requests?  Now 
> your router has two servers running, one to get the jar, one to deal 
> with config, instead of one http server with one cgi script.

Yes, such complicated devices exist. Accessing it with Java
switched off, you can't do anything. Very overcomplicated,
and slow.



> Java has/had its uses, but I don't recall the last time I ran something 
> using java. 

As it has been mentioned, Java is often required in online
banking, but as far as I've noticed, it's also less and less
important in those fields. I'm not using online banking
myself so my opinion is very little substanciated.



> At the moment when it comes to the browser, flash is more 
> important and that's only for all the websites that want to stream 
> instead of give you a file like they used to.

Not only that. Whole suites of development tools are arranged
around "Flash" in order to replace dealing with HTML at all.
Navigational elements, as well as non-AV content is enclosed
in "Flash" to limit accessibility (which of course makes the
web less barrier-free, but who cares except cripples - they
don't count, majority wins). Also "content protection" is a
field where "Flash" is heavily used, like "No, you can't
select this text and copy it somewhere else!" What animated
GIFs were in the past, that's "Flash" today, but much more
ressource-intensive, proprietary, dangerous, and annoying.



> I remember years and years ago starting to learn java. 

It was hard for me to "learn" Java at university when I had
already years of C experience. :-)



> I got really 
> frustrated by spending a few hours going through documentation to find 
> the "proper" way to read a text file. 

I didn't know there was one. :-)



> Writing the gui seemed easy, the 
> rest wasn't.

That's the basic idea: Make it "look good" on the outside, so
it appeals to users using the "first sight effect". Don't care
for the internals, nobody can see them anyway. :-)



-- 
Polytropon
Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: this is probably a little touchy to ask...

[Joshua Isom]

On 9/11/2010 1:10 AM, Polytropon wrote:

Let me add another field: There are applicances like "all-in-one
DSL modem telephone splitter router DHCP server NAT firewall boxes"
that are very common in german households. Those usually use Java
to present their control elements to the user; "Applet loading"
is often seen when connected to that box in order to change some
setting. I think the initial developers found it better to put
a Java applet in there than some PHP generated HTML served by
a little web server... they could have used an efficient and
professional programming language, too, but that's something you
won't find in home consumer crap devices.:-)




So to configure your router, you need a java enabled browser, and odds 
are you get the jar file from the router, so it has an http server, and 
probably another server just to process configuration requests?  Now 
your router has two servers running, one to get the jar, one to deal 
with config, instead of one http server with one cgi script.


Java has/had its uses, but I don't recall the last time I ran something 
using java.  At the moment when it comes to the browser, flash is more 
important and that's only for all the websites that want to stream 
instead of give you a file like they used to.


I remember years and years ago starting to learn java.  I got really 
frustrated by spending a few hours going through documentation to find 
the "proper" way to read a text file.  Writing the gui seemed easy, the 
rest wasn't.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"

Re: ipfw fwd and ipfw allow

[perryh]

Victor Sudakov  wrote:

> ... the 'fwd ... keep-state' statement does create a useful
> dynamic rule. It contradicts the ipfw(8) man page but works ...

Hopefully someone who understands all this will submit a patch
for the man page :)
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"