5:48 p.m.
Although the 'Installation' file captures the output of 'uname -a', it
currently doesn't record the compiler used for the build. Should it
do so, at least for compilers/systems where we know how to find out?
('gcc --version | head -1' is a good start). Some platforms (like
AIX) have multiple compilers - there's the XL series of compilers, the
VisualAge compilers, gcc, and probably a few that I've repressed the
horror... ;)
5:44 a.m.
>>>> "Valdis" == Valdis Kletnieks
<Valdis.Kletnieks(a)vt.edu> writes:
Valdis> Although the 'Installation' file captures the output of
Valdis> 'uname -a', it currently doesn't record the compiler used
Valdis> for the build. Should it do so, at least for
Valdis> compilers/systems where we know how to find out? ('gcc
Valdis> --version | head -1' is a good start).
ISTR you have access to a few different compilers, wanna start adding
to the list?
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
10:37 p.m.
On Sat, 26 Apr 2003 13:44:41 +0900, "Stephen J. Turnbull" said:
ISTR you have access to a few different compilers, wanna start
adding
to the list?
OK.. got hit by a combo of insomnia and too much classic British metal.. ;)
So far, it does:
For gcc, report version and specfile.
For vendor CC, report version for AIX, Solaris, and Tru64.
Reports libc version for RedHat/Mandrake, AIX, Solaris.
It still doesn't do *BSD etc. It doesn't DTRT for Suse/Debian or libc5-based
systems yet, and no libc version for Tru64 yet.
Comments? Anything else that would be useful while I'm at it?
/Valdis
How it looks so far:
Relevant output on my RedHat-9 based laptop:
XEmacs 21.5-b12 "carrot" configured for `i686-pc-linux'.
Compilation / Installation:
Source code location: /home/valdis/src/xemacs-21.5
Installation prefix: /usr/local
Operating system description file: `s/linux.h'
Machine description file: `m/intel386.h'
Compiler: gcc -Wall -Wno-switch -Winline -Wmissing-pr
ototypes -Wsign-compare -Wundef -Wstrict-prototypes -Wpacked -Wshadow -Wmissing-
declarations -Os -pipe -march=pentium4 -mfpmath=sse -fexpensive-optimizations
Compiler version: gcc (GCC) 3.2.2 20030313 (Red Hat Linux 3.2
.2-10)
Compiler specs file: /usr/lib/gcc-lib/i386-redhat-linux/3.2.2/sp
ecs
Relocating allocator for buffers: no
GNU version of malloc: yes
- Using Doug Lea's new malloc from the GNU C Library.
libc: glibc-2.3.2-27.9
On an AIX 5.1 box:
XEmacs 21.5-b11 "cabbage" (+CVS-20030220) configured for
`powerpc-ibm-aix5.1'.
Compilation / Installation:
Source code location: /home/valdis/src/xemacs-21.5
Installation prefix: /usr/local
Operating system description file: `s/aix4-2.h'
Machine description file: `m/ibmrs6000.h'
Compiler: xlc_r -O3 -qstrict -ma -I/usr/local/include
Compiler version: vac.C 5.0.2.5 - C for AIX Compiler
Relocating allocator for buffers: yes
GNU version of malloc: yes
libc: bos.rte.libc 5.1.0.28
On a Solaris 2.8 with a borked gcc install (I did the specs file just to
catch this, I didn't realize I'd be my own first victim - the admin of
the box has been flamed ;)
XEmacs 21.5-b12 "carrot" configured for `sparc-sun-solaris2.8'.
Compilation / Installation:
Source code location: /export/home/valdis/src/xemacs-21.5
Installation prefix: /usr/local
Runtime library search path: /usr/ccs/lib:/usr/dt/lib:/usr/local/lib
Operating system description file: `s/sol2.h'
Machine description file: `m/sparc.h'
Compiler: gcc -Wall -Wno-switch -Winline -Wmissing-pr
ototypes -Wsign-compare -Wundef -Wstrict-prototypes -Wshadow -Wmissing-declarati
ons -Wpointer-arith -g -fno-schedule-insns
Compiler version: 2.8.1
Compiler specs file: /usr/local/lib/gcc-lib/sparc-sun-solaris2.6
/2.8.1/specs
Relocating allocator for buffers: yes
GNU version of malloc: yes
libc: SUNWcsl 11.8.0,REV=2000.01.08.18.12
Same Solaris box, with 'CC=cc':
Compilation / Installation:
Source code location: /export/home/valdis/src/xemacs-21.5
Installation prefix: /usr/local
Runtime library search path: /usr/dt/lib:/opt/SUNWdt/lib
Operating system description file: `s/sol2.h'
Machine description file: `m/sparc.h'
Compiler: cc -v -g -fno-schedule-insns
Compiler version: cc: Sun WorkShop 6 update 2 C 5.3 2001/05/1
5
Relocating allocator for buffers: yes
GNU version of malloc: yes
libc: SUNWcsl 11.8.0,REV=2000.01.08.18.12
8
A Tru64 5.1A box:
XEmacs 21.5-b12 "carrot" configured for `alphaev56-dec-osf5.1'.
Compilation / Installation:
Source code location: /home/valdis/src/xemacs-21.5
Installation prefix: /usr/local
Runtime library search path: /usr/dt/lib
Operating system description file: `s/decosf4-0.h'
Machine description file: `m/alpha.h'
Compiler: cc -std1 -arch ev56 -fast
Compiler version: Compaq C V6.4-014 on Compaq Tru64 UNIX V5.1
A (Rev. 1885) Compiler Driver V6.4-215 (sys) cc Driver
Relocating allocator for buffers: yes
GNU version of malloc: yes
libc: system-provided libc on alphaev56-dec-osf5.
1
6:32 a.m.
|--==> "VK" == Valdis Kletnieks <Valdis.Kletnieks(a)vt.edu> writes:
VK> So far, it does:
VK> For gcc, report version and specfile.
OK, what does knowing where the spec file is at tell us?
VK> For vendor CC, report version for AIX, Solaris, and Tru64.
VK> Reports libc version for RedHat/Mandrake, AIX, Solaris.
Excellent! But why make the distinction of Linux distribution? I
don't know of any distro that doesn't use GNU libc. So surely if you
can get the libc version in RedCrap/Mandunny, you can get it for any
distro, no?
--
|---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---|
| XEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<youngs(a)xemacs.org>---|
9:31 a.m.
>>>> "Valdis" == Valdis Kletnieks
<Valdis.Kletnieks(a)vt.edu> writes:
Valdis> Comments? Anything else that would be useful while I'm at
Valdis> it?
I like it. Patch welcome any time. Don't worry about getting
everything you know how to do; a minimal comment that "we can do
better on Debian/SuSE" points people in the right direction, and "this
info is in Some File In /etc" or similar would probably make somebody
familiar with the system in question gag-and-implement---almost as
good as if you did it yourself.
Tee hee ... now that I know your weaknesses, you'd better make sure
port 8000 is blocked at the firewall!
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
9:05 p.m.
On Tue, 29 Apr 2003 15:32:00 +1000, Steve Youngs <youngs(a)xemacs.org> said:
|--==> "VK" == Valdis Kletnieks
<Valdis.Kletnieks(a)vt.edu> writes:
VK> So far, it does:
VK> For gcc, report version and specfile.
OK, what does knowing where the spec file is at tell us?
On systems where gcc is not the native compiler, the spec file path
encodes the system state at the time gcc was built.
XEmacs 21.5-b12 "carrot" configured for `sparc-sun-solaris2.8'.
^^^
Compiler specs file: /usr/local/lib/gcc-lib/sparc-sun-solaris2.6/2.8.1/specs
^^^
Yep. this one almost certainly has at *least* a 'fixincludes' problem...
I've thwapped the guy in the next cubicle already.
*THAT* was why I included the specfile path.. ;)
Excellent! But why make the distinction of Linux distribution? I
don't know of any distro that doesn't use GNU libc. So surely if you
can get the libc version in RedCrap/Mandunny, you can get it for any
distro, no?
Well.. for RedHat/Mandrake, I just cheat and do 'rpm -q glibc' to get
something a *bit* more specific than 'libc.so.6' (for starters, it's still
called libc.so.6 even under glibc 2.3.2 which has all the TLS stuff).
A co-worker is going to feed me the Debian and Suse checks, and if it's
none of those, I'll probably resort to 'ldd /bin/ls | grep' of some sort.
1:55 a.m.
>>>> "Valdis" == Valdis Kletnieks
<Valdis.Kletnieks(a)vt.edu> writes:
Valdis> A co-worker is going to feed me the Debian and Suse
Valdis> checks, and if it's none of those, I'll probably resort to
Valdis> 'ldd /bin/ls | grep' of some sort.
For Debian and Mac OS X "fink" `dpkg -l' is a good start, although you
can probably do better than that. But at least dpkg -l gives you the
Debian revision.
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
4:06 a.m.
|--==> "VK" == Valdis Kletnieks <Valdis.Kletnieks(a)vt.edu> writes:
VK> On Tue, 29 Apr 2003 15:32:00 +1000, Steve Youngs <youngs(a)xemacs.org>
said:
>Excellent! But why make the distinction of Linux distribution?
I
>don't know of any distro that doesn't use GNU libc. So surely if you
>can get the libc version in RedCrap/Mandunny, you can get it for any
>distro, no?
VK> Well.. for RedHat/Mandrake, I just cheat and do 'rpm -q glibc' to get
VK> something a *bit* more specific than 'libc.so.6' (for starters, it's
still
VK> called libc.so.6 even under glibc 2.3.2 which has all the TLS stuff).
VK> A co-worker is going to feed me the Debian and Suse checks, and if it's
VK> none of those, I'll probably resort to 'ldd /bin/ls | grep' of some
sort.
What's wrong with 'libc.so.6 | head -1'? I don't know about other C
libs, but GNU libc spits out version info if you run it as a program.
--
|---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---|
| XEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<youngs(a)xemacs.org>---|
6:43 a.m.
>>>> "Valdis" == Valdis Kletnieks
<Valdis.Kletnieks(a)vt.edu> writes:
Valdis> On Wed, 30 Apr 2003 13:06:42 +1000, Steve Youngs
Valdis> <youngs(a)xemacs.org> said:
> What's wrong with 'libc.so.6 | head -1'? I don't
know about
> other C libs, but GNU libc spits out version info if you run it
> as a program.
Valdis> OOOH!.. Amazing. :)
... but worthless, on Red Hat. I think things are better now, but on
Sparc glibc nominally 5.0.9 contained code from 4.5.46 and code that
wouldn't be released until 6.0. And at one point Steve Baur said Red
Hat's x86 glibc RPM spec applied nearly 200 patches to the upstream
sources.
You really need the vendor revision information, too.
Hm ... I wonder if on Solaris we could find out whether the various
known Motif patches have been applied? :-)
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
5:36 p.m.
"Stephen J. Turnbull" <stephen(a)xemacs.org> wrote:
Hm ... I wonder if on Solaris we could find out whether the various
known Motif patches have been applied? :-)
Sure. Check the contents of /var/sadm/patch.
--
Jerry James
http://www.ittc.ku.edu/~james/
7664
days inactive
7669
days old
11 comments
4 participants
participants (4)
-
Jerry James -
Stephen J. Turnbull -
Steve Youngs -
Valdis Kl=?iso8859-4?Q?=BA?=tnieks