linux-gcc-digest Monday, 21 June 1999 Volume 01 : Number 394
In this issue:
Re: C++ compiler cannot create executables
glibc 2.1.1
Re: glibc 2.1.1
gdb question?
glibc-2.0.122 problem.
glibc-2.0.112 problem.
glibc-2.0.122 problem.
Re: glibc-2.0.122 problem.
Re: glibc-2.0.122 problem.
Re: glibc-2.0.122 problem.
Re: gdb question?
Re: gdb question?
glibc-2.0 and glibc-2.1 on same systm
binutils 2.9.4.0.1 and Linux/ARM
binutils 2.9.4.0.1 is withdrawn.
Binutils 2.9.4.0.2 is released.
Binutils 2.9.4.0.3 is released.
Killing a process started with "popen()"
Re: Killing a process started with "popen()"
Re: Killing a process started with "popen()"
binutils 2.9.4.0.4 is released.
m68k-coff-as from binutils-2.9.4.0.4 coredumps
binutils 2.9.4.0.5 is released.
Computing Formula
Compilation & performances
Re: Compilation & performances
Re: Compilation & performances
Re: Computing Formula
Facing problem in compiling
Re: Computing Formula
Re: Computing Formula
Re: Computing Formula
Re: Facing problem in compiling
Re: Computing Formula
binutils 2.9.4.0.6 is released.
Re: Computing Formula
Re: Computing Formula
See the end of the digest for information on subscribing to the linux-gcc
or linux-gcc-digest mailing lists.
----------------------------------------------------------------------
From: "Martin v. Loewis" <[EMAIL PROTECTED]>
Date: Mon, 24 May 1999 21:12:25 +0200
Subject: Re: C++ compiler cannot create executables
> checking whether the C++ compiler (gcc ) works... no
[...]
> It surprised me because I can compile the kernel without any problems.
This is no surprise: A C++ compiler is not required for kernel
compilation. Please put the text
int main()
{}
into a file (say, hello.cc), and then compile it with
g++ -o hello hello.cc
if that gives you any error message, you probably need to install the
C++ compiler. If that works, you need to look at config.log to see
what the error was.
Regards,
Martin
------------------------------
From: Ulrich Drepper <[EMAIL PROTECTED]>
Date: 24 May 1999 21:22:42 -0700
Subject: glibc 2.1.1
Hi,
I've made glibc 2.1.1 available at
ftp://ftp.gnu.org/pub/gnu/glibc
ftp://sourceware.cygnus.com/pub/glibc
ftp://ftp.XX.kernel.org/pub/software/libs/glibc
(where XX is your country code). The new files are
glibc-2.1.1.tar.gz
glibc-2.1-2.1.1.diff.gz
glibc-2.1.1pre3-2.1.1.diff.gz (not on ftp.gnu.org)
glibc-linuxthreads-2.1.1.tar.gz
There is *no* new crypt add-on. I hope the site containing the crypt
add-on packages will soon have an appropriate symlink. For the time
being please simply use/download the 2.1 version of the add-on.
glibc 2.1.1 is a bug fix and maintenance release. No significant new
features were added and the few problems reported are either fixed or
documented. I consider glibc 2.1.1 very stable.
The main improvements from 2.1 beside bug fixes are:
- - lots of speedups (especially in the dynamic linker)
- - lots of SPARC and Linux/SPARC improvements
- - RPC/XDR cleanups
- - updates of many character set conversion functions
- - several memory leaks are fixed
I'd like to thank (in no specific order) Andreas Schwab, Roland
McGrath, Andreas Jaeger, Jakub Jelinek, Thorsten Kukuk, Richard
Henderson, Zack Weinberg, Wolfram Gloger, Xavier Leroy, H.J. Lu, Mark
Kettenis, Philip Blundell, and David Miller for their constant
help and support.
Consider showing them your appreciation when you meet them the next
time.
A few more words to those who actually want to compile everything
themselves:
- - do you really have to? It's not easy and the tools must really work.
We don't have the time to lead people through the process.
- - did you remember using the -E option for patch?
- - when something goes wrong, use a different compiler and or binutils.
Chances are very good meanwhile that something is wrong at your side.
- - did you read the FAQ?
- --
- ---------------. drepper at gnu.org ,-. 1325 Chesapeake Terrace
Ulrich Drepper \ ,-------------------' \ Sunnyvale, CA 94089 USA
Cygnus Solutions `--' drepper at cygnus.com `------------------------
------------------------------
From: "Mark H. Wood" <[EMAIL PROTECTED]>
Date: Fri, 28 May 1999 09:32:08 -0500 (EST)
Subject: Re: glibc 2.1.1
Thank you!
One nit: http://www.gnu.org/software/libc/libc.html *still* says that
"the current version is 2.0" and still points to the 2.0 announcement
(/software/libc/ANNOUNCE-2.0.html). This is kind of confusing.
Also, glibc-2.1.tar.gz never did reappear at
ftp://ftp.gnu.org/pub/gnu/glibc/ . There is *still* a glibc-2.1-README
from 11-Feb-1999 saying:
glibc-2.1 has been (temporarily) removed, until some
political issues are worked out.
- --
Mark H. Wood, Lead System Programmer [EMAIL PROTECTED]
Specializing in unusual perspectives for more than twenty years.
------------------------------
From: Fenglou Mao <[EMAIL PROTECTED]>
Date: Mon, 31 May 1999 10:19:42 -0700 (PDT)
Subject: gdb question?
Dear all,
When I want to print some variable, sometimes gdb always told me "no
such symbol", but I am sure it must have that symbol, and it have been
used, and it must be in current namespace.
I want to know if some else met this question and how to
resolve it.
Sincerely Yours,
FengLou Mao
*******************************
ADD:Mr. FengLou Mao
Peking University
BeiJing
P.R.China
Tel:86-10-62751490
Fax:86-10-62751725
------------------------------
From: Elias Athanasopoulos <[EMAIL PROTECTED]>
Date: Mon, 31 May 1999 21:37:57 +0300
Subject: glibc-2.0.122 problem.
Hi,
I tried to build glibc-2.0.122 and after about 2 hours of compiling
time everything seemed to be ok. I compiled glibc with 2.2.1 kernel
header files, although my system is running kernel 2.0.30, and installed
it in /usr/local.
Here is part of my configuration used to build glibc:
o egcs 1.1.1 (egc-2.91.60)
o binutils 2.9.1.0.25
o ld 1.9.9
o GNU Make 3.71.1
o libc 5.4.33
Now, when I'm trying to run the simpe foo.c program:
void main(void) {}
which is compiled as:
$ gcc foo.c -Wl,--dynamic-linker=/usr/lib/ld.linux.so.2 -o foo
I'm getting the following 'strange' msg:
$ ./foo
bash: ./foo: No such file or directory
I've also included /usr/local/lib in the LD_LIBRARY_PATH env.
variable.
Also, when I try the following:
bash# nm -u foo
__mon_start__
__ibc_start_main@@GLIBC_2.0
_p_hw
abort@@GLIBC_2.0
free@@GLIBC_2.0
malloc@@GLIBC_2.0
Any ideas?
If I'm asking in the wrong place I would appreciate u pointing me
to the appropriate one. TIA. :-)
Best Regards,
Elias
- --
Elias Athanasopoulos
------------------
Faculty Of Science
Department Of Physics
*University Of Athens*
------------------
--In a world without fences who needs GATES?
------------------------------
From: Elias Athanasopoulos <[EMAIL PROTECTED]>
Date: Mon, 31 May 1999 22:25:35 +0300
Subject: glibc-2.0.112 problem.
Err...I meant glibc-2.0.112. :-)
I apologise for any confusion....
- ---------- Forwarded Message ----------
Subject: glibc-2.0.122 problem.
Date: Mon, 31 May 1999 21:37:57 +0300
From: Elias Athanasopoulos <[EMAIL PROTECTED]>
Hi,
I tried to build glibc-2.0.122 and after about 2 hours of compiling
[snip]
- --
Elias Athanasopoulos
------------------
Faculty Of Science
Department Of Physics
*University Of Athens*
------------------
--In a world without fences who needs GATES?
------------------------------
From: Elias Athanasopoulos <[EMAIL PROTECTED]>
Date: Mon, 31 May 1999 21:37:57 +0300
Subject: glibc-2.0.122 problem.
Hi,
I tried to build glibc-2.0.122 and after about 2 hours of compiling
time everything seemed to be ok. I compiled glibc with 2.2.1 kernel
header files, although my system is running kernel 2.0.30, and installed
it in /usr/local.
Here is part of my configuration used to build glibc:
o egcs 1.1.1 (egc-2.91.60)
o binutils 2.9.1.0.25
o ld 1.9.9
o GNU Make 3.71.1
o libc 5.4.33
Now, when I'm trying to run the simpe foo.c program:
void main(void) {}
which is compiled as:
$ gcc foo.c -Wl,--dynamic-linker=/usr/lib/ld.linux.so.2 -o foo
I'm getting the following 'strange' msg:
$ ./foo
bash: ./foo: No such file or directory
I've also included /usr/local/lib in the LD_LIBRARY_PATH env.
variable.
Also, when I try the following:
bash# nm -u foo
__mon_start__
__ibc_start_main@@GLIBC_2.0
_p_hw
abort@@GLIBC_2.0
free@@GLIBC_2.0
malloc@@GLIBC_2.0
Any ideas?
If I'm asking in the wrong place I would appreciate u pointing me
to the appropriate one. TIA. :-)
Best Regards,
Elias
- --
Elias Athanasopoulos
------------------
Faculty Of Science
Department Of Physics
*University Of Athens*
------------------
--In a world without fences who needs GATES?
------------------------------
From: [EMAIL PROTECTED]
Date: Mon, 31 May 1999 19:40:33 ric
Subject: Re: glibc-2.0.122 problem.
Dumb question: between installing libraries and trying to use them, did
you run ldconfig? Also, if you are going to run glibc from
/usr/local/lib, I think you want that in /etc/ld.so.conf.
/usr/lib/ld.linux.so.2 seems like an odd spelling and location. IIRC if
you install glibc in /usr/local/lib, ld-linux.so.2 goes there too,
though it's just a symbolic link to the real ld-2.0.112.so or so.
Well, in case this might help. :-)
Lawson
On Mon, 31 May 1999, Elias Athanasopoulos wrote:
>
> Hi,
>
> I tried to build glibc-2.0.122 and after about 2 hours of
compiling
> time everything seemed to be ok. I compiled glibc with 2.2.1
kernel
> header files, although my system is running kernel 2.0.30, and
installed
> it in /usr/local.
>
> Here is part of my configuration used to build glibc:
>
> o egcs 1.1.1 (egc-2.91.60)
> o binutils 2.9.1.0.25
> o ld 1.9.9
> o GNU Make 3.71.1
> o libc 5.4.33
>
> Now, when I'm trying to run the simpe foo.c program:
>
> void main(void) {}
>
> which is compiled as:
>
> $ gcc foo.c -Wl,--dynamic-linker=/usr/lib/ld.linux.so.2 -o foo
>
> I'm getting the following 'strange' msg:
>
> $ ./foo
> bash: ./foo: No such file or directory
>
> I've also included /usr/local/lib in the LD_LIBRARY_PATH env.
> variable.
>
> Also, when I try the following:
>
> bash# nm -u foo
> __mon_start__
> __ibc_start_main@@GLIBC_2.0
> _p_hw
> abort@@GLIBC_2.0
> free@@GLIBC_2.0
> malloc@@GLIBC_2.0
>
> Any ideas?
>
> If I'm asking in the wrong place I would appreciate u pointing
me
> to the appropriate one. TIA. :-)
>
> Best Regards,
> Elias
>
> --
>
> Elias Athanasopoulos
>
> ------------------
> Faculty Of Science
> Department Of Physics
>
> *University Of Athens*
> ------------------
>
> --In a world without fences who needs GATES?
>
___________________________________________________________________
You don't need to buy Internet access to use free Internet e-mail.
Get completely free e-mail from Juno at http://www.juno.com/getjuno.html
or call Juno at (800) 654-JUNO [654-5866]
------------------------------
From: Andreas Jaeger <[EMAIL PROTECTED]>
Date: 01 Jun 1999 07:08:56 +0200
Subject: Re: glibc-2.0.122 problem.
>>>>> Elias Athanasopoulos writes:
> Hi,
> I tried to build glibc-2.0.122 and after about 2 hours of compiling
> time everything seemed to be ok. I compiled glibc with 2.2.1 kernel
> header files, although my system is running kernel 2.0.30, and installed
> it in /usr/local.
> Here is part of my configuration used to build glibc:
> o egcs 1.1.1 (egc-2.91.60)
> o binutils 2.9.1.0.25
> o ld 1.9.9
> o GNU Make 3.71.1
> o libc 5.4.33
> Now, when I'm trying to run the simpe foo.c program:
> void main(void) {}
> which is compiled as:
> $ gcc foo.c -Wl,--dynamic-linker=/usr/lib/ld.linux.so.2 -o foo
That's the wrong path to ld-linux - you installed it in
/usr/local/lib.
> I'm getting the following 'strange' msg:
> $ ./foo
> bash: ./foo: No such file or directory
> I've also included /usr/local/lib in the LD_LIBRARY_PATH env.
> variable.
> Also, when I try the following:
> bash# nm -u foo
> __mon_start__
> __ibc_start_main@@GLIBC_2.0
> _p_hw
> abort@@GLIBC_2.0
> free@@GLIBC_2.0
> malloc@@GLIBC_2.0
That's ok.
> Any ideas?
> If I'm asking in the wrong place I would appreciate u pointing me
> to the appropriate one. TIA. :-)
You should really install glibc 2.1.1 and read the whole FAQ and
INSTALL documents first.
Andreas
- --
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
------------------------------
From: "Martin v. Loewis" <[EMAIL PROTECTED]>
Date: Tue, 1 Jun 1999 08:13:25 +0200
Subject: Re: glibc-2.0.122 problem.
> I tried to build glibc-2.0.122 and after about 2 hours of compiling
> time everything seemed to be ok.
[...]
> If I'm asking in the wrong place I would appreciate u pointing me
> to the appropriate one. TIA. :-)
You aren't asking in the wrong place. You are compiling the wrong
library. glibc 2.1.1 has just been released; please try this one.
Regards,
Martin
------------------------------
From: Scott McDermott <[EMAIL PROTECTED]>
Date: Tue, 1 Jun 1999 02:37:29 -0400
Subject: Re: gdb question?
Fenglou Mao on Mon 31/05 10:19 -0700:
> When I want to print some variable, sometimes gdb always told me "no
> such symbol", but I am sure it must have that symbol, and it have been
> used, and it must be in current namespace.
Sounds like you did not compile in debugging symbols.
- --
Scott
------------------------------
From: Mark Gray <[EMAIL PROTECTED]>
Date: 01 Jun 1999 09:25:26 -0400
Subject: Re: gdb question?
Fenglou Mao <[EMAIL PROTECTED]> writes:
>
> Dear all,
> When I want to print some variable, sometimes gdb always told me "no
> such symbol", but I am sure it must have that symbol, and it have been
> used, and it must be in current namespace.
> I want to know if some else met this question and how to
> resolve it.
When this happens to me, most often I find that what I thought was a
variable was actually a macro (whose symbol definition disappears in
the preprocessor.)
Hope this helps.
>
> Sincerely Yours,
>
> FengLou Mao
> *******************************
> ADD:Mr. FengLou Mao
> Peking University
> BeiJing
> P.R.China
> Tel:86-10-62751490
> Fax:86-10-62751725
>
>
>
------------------------------
From: Michael Lausch <[EMAIL PROTECTED]>
Date: Fri, 04 Jun 1999 12:53:24 +0200
Subject: glibc-2.0 and glibc-2.1 on same systm
I know it's possible to have libc5 and glibc one one system and i also
know how to do it. I've tried the sam things for glibc-2.1 but it
didn't work. The glibc-2.1 ld-linux isn't compatible to glibc-2.0.7
ld-linux and the other way round. Is there an easy way to have both
libcs on one machine without creating a chroot environemnt?
- --
Michael Lausch
See my web page <http://www.lausch.at/> or query PGP key server for PGP key.
"Reality is that which, when you stop believing in it, doesn't go away".
-- Philip K. Dick
------------------------------
From: [EMAIL PROTECTED] (H.J. Lu)
Date: Sun, 6 Jun 1999 09:09:08 -0700 (PDT)
Subject: binutils 2.9.4.0.1 and Linux/ARM
Hi,
I forgot including some important Linux/ARM note on binutils 2.9.4.0.1.
If you use Linux/ARM, please make sure you read it first.
Sorry for that.
- --
H.J. Lu ([EMAIL PROTECTED])
- ---
This is the beta release of binutils 2.9.4.0.1 for Linux, which is
based on binutils 1999 0605 plus various changes.
Please report any bugs related to binutils 2.9.4.0.1 to [EMAIL PROTECTED]
For arm-linux targets, there are some important differences in behaviour
between these tools and binutils 2.9.1.0.x. The linker emulation name has
changed from elf32arm{26} to armelf_linux{26}. Also, the "-p" flag must be
passed with the linker when working with object files (or static libraries)
created using older versions of the assembler. If this flag is omitted the
linker will silently generate bad output when given old input files.
To get the correct behaviour from gcc, amend the *link section of your specs
file as follows:
*link:
%{h*} %{version:-v} %{b} %{Wl,*:%*} %{static:-Bstatic} %{shared:-shared
} %{symbolic:-Bsymbolic} %{rdynamic:-export-dynamic} %{!dynamic-linker:
- -dynamic-linker /lib/ld-linux.so.2} -X %{mbig-endian:-EB} %{mapcs-26:-m ar
melf_linux26} %{!mapcs-26:-m armelf_linux} -p
Changes from binutils 1999 0605:
1. Fix a -Bsymbolic bug for Linux/alpha.
2. Add ELF/i370.
3. Fix 8/16-bit relocations for i386.
4. Add --redefine-sym=old_form=new_form to objcopy.
5. Add "-j section" for objcopy.
6. Fix i386 disassembler for fwait.
7. Fix a Sparc asm bug.
8. Add Ada demangle support.
9. Fix MIPS/ELF bugs.
10. Add some vxworks suppport.
11. Fix a.out assembler.
The file list:
1. binutils-2.9.4.0.1.tar.gz. Source code.
2. binutils-2.9.4.0.1-1.src.rpm. Source RPM.
3. binutils-2.9.4.0.1-1.i386.rpm. Binary RPM for RedHat 6.0.
The primary ftp sites for the beta Linux binutils are:
1. ftp://ftp.varesearch.com/pub/support/hjl/binutils/beta
Thanks.
H.J. Lu
[EMAIL PROTECTED]
06/05/99
------------------------------
From: [EMAIL PROTECTED] (H.J. Lu)
Date: Sun, 6 Jun 1999 12:30:42 -0700 (PDT)
Subject: binutils 2.9.4.0.1 is withdrawn.
Hi,
binutils 2.9.4.0.1 has been withdrawn due to the i386 bug which
prevents me from building the Linux kernel. I got error messages:
ld -m elf_i386 -T /work/linux/src/kernel/linux/arch/i386/vmlinux.lds -e stext
arch/i386/kernel/head.o arch/i386/kernel/init_task.o init/main.o init/version.o
\
--start-group \
arch/i386/kernel/kernel.o arch/i386/mm/mm.o kernel/kernel.o mm/mm.o
fs/fs.o ipc/ipc.o \
fs/filesystems.a \
net/network.a \
drivers/block/block.a drivers/char/char.a drivers/misc/misc.a
drivers/net/net.a drivers/scsi/scsi.a drivers/cdrom/cdrom.a drivers/pci/pci.a
drivers/pnp/pnp.a drivers/video/video.a \
/work/linux/src/kernel/linux/arch/i386/lib/lib.a
/work/linux/src/kernel/linux/lib/lib.a
/work/linux/src/kernel/linux/arch/i386/lib/lib.a \
--end-group \
-o vmlinux
arch/i386/kernel/head.o: In function `_stext':
arch/i386/kernel/head.o(.text+0x18): relocation truncated to fit: R_386_32
boot_cpu_data
arch/i386/kernel/init_task.o(.data.init_task+0x2fc): relocation truncated to
fit: R_386_32 swapper_pg_dir
arch/i386/kernel/kernel.o: In function `machine_restart':
arch/i386/kernel/kernel.o(.text+0x178a): relocation truncated to fit: R_386_32
swapper_pg_dir
arch/i386/kernel/kernel.o: In function `trap_init_f00f_bug':
arch/i386/kernel/kernel.o(.text.init+0x844): relocation truncated to fit:
R_386_32 idt_table
arch/i386/kernel/kernel.o: In function `gdt_48':
arch/i386/kernel/kernel.o(.data+0x1e13): relocation truncated to fit: R_386_32
gdt_table
arch/i386/mm/mm.o: In function `__bad_page':
arch/i386/mm/mm.o(.text+0x285): relocation truncated to fit: R_386_32
empty_bad_page
kernel/kernel.o: In function `exit_mm':
kernel/kernel.o(.text+0x8b07): relocation truncated to fit: R_386_32
swapper_pg_dir
kernel/kernel.o: In function `do_exit':
kernel/kernel.o(.text+0x8e79): relocation truncated to fit: R_386_32
swapper_pg_dir
mm/mm.o: In function `zeromap_page_range':
mm/mm.o(.text+0x609): relocation truncated to fit: R_386_32 empty_zero_page
mm/mm.o: In function `do_anonymous_page':
mm/mm.o(.text+0x1187): relocation truncated to fit: R_386_32 empty_zero_page
I am invesigating it now. There is no time table for the next release.
- --
H.J. Lu ([EMAIL PROTECTED])
------------------------------
From: [EMAIL PROTECTED] (H.J. Lu)
Date: Sun, 6 Jun 1999 17:57:42 -0700 (PDT)
Subject: Binutils 2.9.4.0.2 is released.
Hi,
binutils 2.9.4.0.2 seems ok. I can build Linux kernel and gcc with
it now.
- --
H.J. Lu ([EMAIL PROTECTED])
- ---
This is the beta release of binutils 2.9.4.0.2 for Linux, which is
based on binutils 1999 0606 plus various changes.
Please report any bugs related to binutils 2.9.4.0.2 to [EMAIL PROTECTED]
For arm-linux targets, there are some important differences in behaviour
between these tools and binutils 2.9.1.0.x. The linker emulation name has
changed from elf32arm{26} to armelf_linux{26}. Also, the "-p" flag must be
passed with the linker when working with object files (or static libraries)
created using older versions of the assembler. If this flag is omitted the
linker will silently generate bad output when given old input files.
To get the correct behaviour from gcc, amend the *link section of your specs
file as follows:
*link:
%{h*} %{version:-v} %{b} %{Wl,*:%*} %{static:-Bstatic} %{shared:-shared
} %{symbolic:-Bsymbolic} %{rdynamic:-export-dynamic} %{!dynamic-linker:
- -dynamic-linker /lib/ld-linux.so.2} -X %{mbig-endian:-EB} %{mapcs-26:-m ar
melf_linux26} %{!mapcs-26:-m armelf_linux} -p
Changes from binutils 2.9.4.0.1:
1. Update from binutils 1999 0606.
2. Restore relocation overflow checking in binutils 2.9.1.0.25 so that
Linux kernel can build.
3. Fix i370 for the new gas.
Changes from binutils 1999 0605:
1. Fix a -Bsymbolic bug for Linux/alpha.
2. Add ELF/i370.
3. Fix 8/16-bit relocations for i386.
4. Add --redefine-sym=old_form=new_form to objcopy.
5. Add "-j section" for objcopy.
6. Fix i386 disassembler for fwait.
7. Fix a Sparc asm bug.
8. Add Ada demangle support.
9. Fix MIPS/ELF bugs.
10. Add some vxworks suppport.
11. Fix a.out assembler.
The file list:
1. binutils-2.9.4.0.2.tar.gz. Source code.
2. binutils-2.9.4.0.1-2.9.4.0.2.diff.gz. Patch against the previous
beta source code.
3. binutils-2.9.1.0.25-2.9.4.0.2.diff.gz. Patch against binutils
2.9.1.0.25.
4. binutils-2.9.4.0.2-1.src.rpm. Source RPM.
5. binutils-2.9.4.0.2-1.i386.rpm. Binary RPM for RedHat 6.0.
The primary ftp sites for the beta Linux binutils are:
1. ftp://ftp.varesearch.com/pub/support/hjl/binutils/beta
Thanks.
H.J. Lu
[EMAIL PROTECTED]
06/06/99
------------------------------
From: [EMAIL PROTECTED] (H.J. Lu)
Date: Mon, 7 Jun 1999 15:31:08 -0700 (PDT)
Subject: Binutils 2.9.4.0.3 is released.
>
> Hi,
>
> there are some problems with 2.9.4.0.1 on PPC. First, I can't compile
> gcc-2.95pre anymore, cause your version contains the same bug as
> current cvs-binutils. This bug is very new, it doesn't happen with a 990601 cvs
> eabi
> eabi.S: Assembler messages:
> eabi.S:45: Internal error, aborting at symbols.c line 1625 in copy_symbol_attributes
> Please report this bug.
Please try binutils 2.9.4.0.3 and let me know what you get.
Thanks.
H.J.
- ----
This is the beta release of binutils 2.9.4.0.3 for Linux, which is
based on binutils 1999 0607 plus various changes.
Please report any bugs related to binutils 2.9.4.0.3 to [EMAIL PROTECTED]
For arm-linux targets, there are some important differences in behaviour
between these tools and binutils 2.9.1.0.x. The linker emulation name has
changed from elf32arm{26} to armelf_linux{26}. Also, the "-p" flag must be
passed with the linker when working with object files (or static libraries)
created using older versions of the assembler. If this flag is omitted the
linker will silently generate bad output when given old input files.
To get the correct behaviour from gcc, amend the *link section of your specs
file as follows:
*link:
%{h*} %{version:-v} %{b} %{Wl,*:%*} %{static:-Bstatic} %{shared:-shared
} %{symbolic:-Bsymbolic} %{rdynamic:-export-dynamic} %{!dynamic-linker:
- -dynamic-linker /lib/ld-linux.so.2} -X %{mbig-endian:-EB} %{mapcs-26:-m ar
melf_linux26} %{!mapcs-26:-m armelf_linux} -p
Changes from binutils 2.9.4.0.2:
1. Update from binutils 1999 0607.
2. Remove my Sparc hacks.
3. Fix local symbol copy.
Changes from binutils 2.9.4.0.1:
1. Update from binutils 1999 0606.
2. Restore relocation overflow checking in binutils 2.9.1.0.25 so that
Linux kernel can build.
3. Fix i370 for the new gas.
Changes from binutils 1999 0605:
1. Fix a -Bsymbolic bug for Linux/alpha.
2. Add ELF/i370.
3. Fix 8/16-bit relocations for i386.
4. Add --redefine-sym=old_form=new_form to objcopy.
5. Add "-j section" for objcopy.
6. Fix i386 disassembler for fwait.
7. Fix a Sparc asm bug.
8. Add Ada demangle support.
9. Fix MIPS/ELF bugs.
10. Add some vxworks suppport.
11. Fix a.out assembler.
The file list:
1. binutils-2.9.4.0.3.tar.gz. Source code.
2. binutils-2.9.4.0.2-2.9.4.0.3.diff.gz. Patch against the previous
beta source code.
3. binutils-2.9.4.0.3-1.src.rpm. Source RPM.
4. binutils-2.9.4.0.3-1.i386.rpm. Binary RPM for RedHat 6.0.
There are also bzip2 versions of tar and diff files.
The primary ftp sites for the beta Linux binutils are:
1. ftp://ftp.varesearch.com/pub/support/hjl/binutils/beta
Thanks.
H.J. Lu
[EMAIL PROTECTED]
06/07/99
------------------------------
From: holotko <[EMAIL PROTECTED]>
Date: Tue, 08 Jun 1999 17:32:42 +0000
Subject: Killing a process started with "popen()"
This is a multi-part message in MIME format.
- --------------D01B7316A8A6B566742C366C
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
I have an application here that launches a process via the "popen()"
command.
For anyone unfamiliar popen() works as follows:
FILE *fp; /* a FILE pointer */
...
fp = popen(command, "r");
Where "command" is any common Unix command that can be run from the
shell written as it would be if run from the command line. For
example, "command" could be:
char *command = "ls -l";
or anything that could be run in that manner. The "r" is the mode
which in this case specifies that the pipe will be used for reading.
What "popen()" essentially does is it opens a pipe, forks a new
(child) process and launches the command specified by "command". The
output of the command is then sent down the pipe and can be read in
the parent via the file pointer "fp" using something like "fread()".
Now here's the main question. The processed launched via "popen()"
from the parent process runs over a period of time. Is there a simple
way to kill the process that was started via "popen()" from the parent
process? I considered the possibility of sending a signal from the
parent to the process launched by popen() using the "kill()" function
but "kill()" requires knowledge of the pid of the process to which I
want the signal to be sent. Unfortunately "popen()" returns a file
pointer but nowhere does it return the pid of the process that was
started.
Thus the question boils down to, is there a simple way for the parent
process to get the pid of the process that it stared via "popen()" so
that it could thereby send signals to it? and/or is there a simple
way in which the parent can kill such a process? Perhaps there is but
I am just overlooking it.
or...
Would it make more sense to change the whole manner in which the
process is being started, for example, instead of using "popen()",
should I manually create a pipe via pipe(), fork a new process via
"fork()" and then replacing the child process with the desired command
using "exec()" and simply redirect the file descriptors so that
standard output from the desired command is sent down the pipe and can
thereby be read via a low level "read()"??
Any suggestions would be appreciated.
Thank You.
/John <[EMAIL PROTECTED]>
- --
email: [EMAIL PROTECTED]
Local mailserver <landreau.ruffe.edu> , remote <ns.computer.net>
- --------
Linx:
http://www.computer.net/
http://einval.vol.8m.com/
Yeah!!! Look at him run that ball down!!!
- --------------D01B7316A8A6B566742C366C
Content-Type: text/x-vcard; charset=us-ascii;
name="micros50.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for holotko
Content-Disposition: attachment;
filename="micros50.vcf"
begin:vcard
n:Holotko;John
x-mozilla-html:FALSE
org:MicroService Co.
adr:;;;;;;
version:2.1
email;internet:[EMAIL PROTECTED]
title:Programmer
note:Phone: (914)-779-3966
x-mozilla-cpt:;5088
fn:John Holotko
end:vcard
- --------------D01B7316A8A6B566742C366C--
------------------------------
From: [EMAIL PROTECTED] (Aaron M. Ucko)
Date: 08 Jun 1999 19:36:15 -0400
Subject: Re: Killing a process started with "popen()"
As man popen would tell you:
The pclose function waits for the associated process to
terminate and returns the exit status of the command as
returned by wait4.
- --
Aaron M. Ucko, KB1CJC <[EMAIL PROTECTED]> (finger [EMAIL PROTECTED])
------------------------------
From: holotko <[EMAIL PROTECTED]>
Date: Wed, 09 Jun 1999 17:47:20 +0000
Subject: Re: Killing a process started with "popen()"
This is a multi-part message in MIME format.
- --------------838790DA3DCED9B544AFF22B
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
"Aaron M. Ucko" wrote:
>
> As man popen would tell you:
>
> The pclose function waits for the associated process to
> terminate and returns the exit status of the command as
> returned by wait4.
>
> --
> Aaron M. Ucko, KB1CJC <[EMAIL PROTECTED]> (finger [EMAIL PROTECTED])
Thanks, but this offers me little help with what I am actually trying
to do. Nonetheless it did get me looking at the whole problem and I
decided that using "popen()" is not the best way to start the desired
process. Thus I am going to recode that part of the application using
pipes redirection of file descriptors, forking and replacing the child
with the desired process, reading it's output from stdout which will
be redirected into a pipe. All in all this will give me better control
of the whole thing, it will improve reliability and enable better
control over passing data and signals between the 2 processes and
eliminate some potential known bugs that are often associated with
"popen()".
Anyway's thanks for the help.
John <[EMAIL PROTECTED]>
- --
email: [EMAIL PROTECTED]
Local mailserver <landreau.ruffe.edu> , remote <ns.computer.net>
- --------
Linx:
http://www.computer.net/
http://einval.vol.8m.com/
Great moments come to every man... And so do miserable moments too...
- --------------838790DA3DCED9B544AFF22B
Content-Type: text/x-vcard; charset=us-ascii;
name="micros50.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for holotko
Content-Disposition: attachment;
filename="micros50.vcf"
begin:vcard
n:Holotko;John
x-mozilla-html:FALSE
org:MicroService Co.
adr:;;;;;;
version:2.1
email;internet:[EMAIL PROTECTED]
title:Programmer
note:Phone: (914)-779-3966
x-mozilla-cpt:;5088
fn:John Holotko
end:vcard
- --------------838790DA3DCED9B544AFF22B--
------------------------------
From: [EMAIL PROTECTED] (H.J. Lu)
Date: Fri, 11 Jun 1999 07:53:12 -0700 (PDT)
Subject: binutils 2.9.4.0.4 is released.
Sparc32 bug should be fixed in this release.
- --
H.J. Lu ([EMAIL PROTECTED])
- ---
This is the beta release of binutils 2.9.4.0.4 for Linux, which is
based on binutils 1999 0611 plus various changes. It is purely for
Linux, although it has been tested on Solaris/Sparc and Solaris/x86
from time to time.
Please report any bugs related to binutils 2.9.4.0.4 to [EMAIL PROTECTED]
For arm-linux targets, there are some important differences in behaviour
between these tools and binutils 2.9.1.0.x. The linker emulation name has
changed from elf32arm{26} to armelf_linux{26}. Also, the "-p" flag must be
passed with the linker when working with object files (or static libraries)
created using older versions of the assembler. If this flag is omitted the
linker will silently generate bad output when given old input files.
To get the correct behaviour from gcc, amend the *link section of your specs
file as follows:
*link:
%{h*} %{version:-v} %{b} %{Wl,*:%*} %{static:-Bstatic} %{shared:-shared
} %{symbolic:-Bsymbolic} %{rdynamic:-export-dynamic} %{!dynamic-linker:
- -dynamic-linker /lib/ld-linux.so.2} -X %{mbig-endian:-EB} %{mapcs-26:-m ar
melf_linux26} %{!mapcs-26:-m armelf_linux} -p
Changes from binutils 2.9.4.0.3:
1. Update from binutils 1999 0611.
2. Remove my ELF/Alpha bfd changes.
3. Use the local symbol copy fix in binutils 1999 0611.
Changes from binutils 2.9.4.0.2:
1. Update from binutils 1999 0607.
2. Remove my Sparc hacks.
3. Fix local symbol copy.
Changes from binutils 2.9.4.0.1:
1. Update from binutils 1999 0606.
2. Restore relocation overflow checking in binutils 2.9.1.0.25 so that
Linux kernel can build.
3. Fix i370 for the new gas.
Changes from binutils 1999 0605:
1. Fix a -Bsymbolic bug for Linux/alpha.
2. Add ELF/i370.
3. Fix 8/16-bit relocations for i386.
4. Add --redefine-sym=old_form=new_form to objcopy.
5. Add "-j section" for objcopy.
6. Fix i386 disassembler for fwait.
7. Fix a Sparc asm bug.
8. Add Ada demangle support.
9. Fix MIPS/ELF bugs.
10. Add some vxworks suppport.
11. Fix a.out assembler.
The file list:
1. binutils-2.9.4.0.4.tar.gz. Source code.
2. binutils-2.9.4.0.3-2.9.4.0.4.diff.gz. Patch against the previous
beta source code.
3. binutils-2.9.4.0.4-1.src.rpm. Source RPM.
4. binutils-2.9.4.0.4-1.i386.rpm. Binary RPM for RedHat 6.0.
There are also bzip2 versions of tar and diff files.
The primary ftp sites for the beta Linux binutils are:
1. ftp://ftp.varesearch.com/pub/support/hjl/binutils/beta
Thanks.
H.J. Lu
[EMAIL PROTECTED]
06/11/99
------------------------------
From: "Aaron J. Grier" <[EMAIL PROTECTED]>
Date: Fri, 11 Jun 1999 16:40:32 -0700
Subject: m68k-coff-as from binutils-2.9.4.0.4 coredumps
I've been trying for at least two weeks to make an m68k-coff cross
assembler under linux, but have been getting this:
[aaron@aaron2 src]$ as -v --bitwise-or -k --register-prefix-optional -ovideo.o video.s
GNU assembler version 2.9.1 (m68k-coff)
Segmentation fault (core dumped)
[aaron@aaron2 src]$ gdb `type -path as`
GNU gdb 4.17.0.11 with Linux support
[...]
Program terminated with signal 11, Segmentation fault.
[...]
#0 0x8062658 in do_linenos_for (abfd=0x8107558, h=0xbffff864,
file_cursor=0xbffff860) at ../../gas/config/obj-coff.c:3072
3072 line_ptr->line.l_addr.l_symndx =
(gdb) bt
#0 0x8062658 in do_linenos_for (abfd=0x8107558, h=0xbffff864,
file_cursor=0xbffff860) at ../../gas/config/obj-coff.c:3072
#1 0x8062a18 in write_object_file () at
../../gas/config/obj-coff.c:3273
#2 0x804b3e9 in main (argc=2, argv=0x80b0e68) at ../../gas/as.c:812
#3 0x4006fcb3 in __libc_start_main (main=0x804b2c0 <main>, argc=7,
argv=0xbffffa94, init=0x80490d4 <_init>, fini=0x806632c <_fini>,
rtld_fini=0x4000a350 <_dl_fini>, stack_end=0xbffffa8c)
at ../sysdeps/generic/libc-start.c:78
so when I saw that binutils was actively being hacked on again, I
figured I'd give it a try, but gas coredumps yet again, but in a
different place. (This is with binutils-2.9.4.0.4)
[...]
Program terminated with signal 11, Segmentation fault.
[...]
#0 0x804f9c8 in hash_lookup (table=0x0, key=0x80f0dd3 "EPROM_BASE",
plist=0x0, phash=0x0) at ../../gas/hash.c:164
164 index = hash % table->size;
(gdb) bt
#0 0x804f9c8 in hash_lookup (table=0x0, key=0x80f0dd3 "EPROM_BASE",
plist=0x0, phash=0x0) at ../../gas/hash.c:164
#1 0x804fbd6 in hash_find (table=0x0, key=0x80f0dd3 "EPROM_BASE")
at ../../gas/hash.c:315
#2 0x8058f0b in symbol_find_base (name=0x80f0dd3 "EPROM_BASE",
strip_underscore=0) at ../../gas/symbols.c:656
#3 0x8058e75 in symbol_find (name=0x80f0dd3 "EPROM_BASE")
at ../../gas/symbols.c:613
#4 0x8058e1e in symbol_find_or_make (name=0x80f0dd3 "EPROM_BASE")
at ../../gas/symbols.c:556
#5 0x8057596 in equals (sym_name=0x80f0dd3 "EPROM_BASE", reassign=1)
at ../../gas/read.c:4844
#6 0x8053d5a in read_a_source_file (name=0xbffff9c8 "video.s")
at ../../gas/read.c:732
#7 0x804bbe1 in perform_an_assembly_pass (argc=2, argv=0x80b6250)
at ../../gas/as.c:993
#8 0x804b993 in main (argc=2, argv=0x80b6250) at ../../gas/as.c:810
#9 0x40072cb3 in __libc_start_main (main=0x804b880 <main>, argc=8,
argv=0xbffff834, init=0x804954c <_init>, fini=0x80699bc <_fini>,
rtld_fini=0x4000a350 <_dl_fini>, stack_end=0xbffff82c)
at ../sysdeps/generic/libc-start.c:78
All binutils have been configured with:
../configure --target=m68k-coff --enable-shared \
--prefix=/usr/local/cross --exec-prefix=/usr/local/cross \
--enable-targets=m68k-ieee
It's strange because if there are errors in the source, the assembler
dutifully reports them without segfaulting.
- ----
Aaron J. Grier -- Frye Electronics -- [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (H.J. Lu)
Date: Sat, 12 Jun 1999 12:28:20 -0700 (PDT)
Subject: binutils 2.9.4.0.5 is released.
This is the beta release of binutils 2.9.4.0.5 for Linux, which is
based on binutils 1999 0612 plus various changes. It is purely for
Linux, although it has been tested on Solaris/Sparc and Solaris/x86
from time to time.
Please report any bugs related to binutils 2.9.4.0.5 to [EMAIL PROTECTED]
For arm-linux targets, there are some important differences in behaviour
between these tools and binutils 2.9.1.0.x. The linker emulation name has
changed from elf32arm{26} to armelf_linux{26}. Also, the "-p" flag must be
passed with the linker when working with object files (or static libraries)
created using older versions of the assembler. If this flag is omitted the
linker will silently generate bad output when given old input files.
To get the correct behaviour from gcc, amend the *link section of your specs
file as follows:
*link:
%{h*} %{version:-v} %{b} %{Wl,*:%*} %{static:-Bstatic} %{shared:-shared
} %{symbolic:-Bsymbolic} %{rdynamic:-export-dynamic} %{!dynamic-linker:
- -dynamic-linker /lib/ld-linux.so.2} -X %{mbig-endian:-EB} %{mapcs-26:-m ar
melf_linux26} %{!mapcs-26:-m armelf_linux} -p
Changes from binutils 2.9.4.0.4:
1. Update from binutils 1999 0612.
2. Remove various temporary fixes of mine since those bugs are fixed
now.
Changes from binutils 2.9.4.0.3:
1. Update from binutils 1999 0611.
2. Remove my ELF/Alpha bfd changes.
3. Use the local symbol copy fix in binutils 1999 0611.
Changes from binutils 2.9.4.0.2:
1. Update from binutils 1999 0607.
2. Remove my Sparc hacks.
3. Fix local symbol copy.
Changes from binutils 2.9.4.0.1:
1. Update from binutils 1999 0606.
2. Restore relocation overflow checking in binutils 2.9.1.0.25 so that
Linux kernel can build.
3. Fix i370 for the new gas.
Changes from binutils 1999 0605:
1. Fix a -Bsymbolic bug for Linux/alpha.
2. Add ELF/i370.
3. Fix 8/16-bit relocations for i386.
4. Add --redefine-sym=old_form=new_form to objcopy.
5. Add "-j section" for objcopy.
6. Fix i386 disassembler for fwait.
7. Fix a Sparc asm bug.
8. Add Ada demangle support.
9. Fix MIPS/ELF bugs.
10. Add some vxworks suppport.
11. Fix a.out assembler.
The file list:
1. binutils-2.9.4.0.5.tar.gz. Source code.
2. binutils-2.9.4.0.4-2.9.4.0.5.diff.gz. Patch against the previous
beta source code.
3. binutils-2.9.4.0.5-1.src.rpm. Source RPM.
4. binutils-2.9.4.0.5-1.i386.rpm. Binary RPM for RedHat 6.0.
There are also bzip2 versions of tar and diff files.
The primary ftp sites for the beta Linux binutils are:
1. ftp://ftp.varesearch.com/pub/support/hjl/binutils/beta
Thanks.
H.J. Lu
[EMAIL PROTECTED]
06/12/99
------------------------------
From: Rafael Marco de Lucas <[EMAIL PROTECTED]>
Date: Thu, 17 Jun 1999 13:35:00 +0200
Subject: Computing Formula
Hello,
i am writing a program that has as input a formula text (i.e.
X+log(Y)*(W-Z))
and i want to compute this formula many times for several values of X,Y,W,Z,
how can i do it ? i need it to run quite fast around 300.000times/sec,
i tried to write a little interpreter of the formula and i stored a sequence
of
functions and arguments and temporal variables used.. so that it works but i
guess
it should be easier,
i have in mind other posibility that is to write the formula in a file and
compile
it and relink,.. but i dont know well how to do it and maybe it is even more
copmlex
and i have to do exactly the same but using conditions (i.e. W+Z>4.and.Z<1
sorry, the
users will be fortran fans)
is there any library to do this ??
can somebody help me or send me any sugesstion please ?
thanks in advance, Cheers
Rafa
pd. also thanks to the people that send me answers about my last off-topic
question
about the ramdisk
------------------------------
From: Nicolas Weeger <[EMAIL PROTECTED]>
Date: Thu, 17 Jun 1999 20:41:24 +0200
Subject: Compilation & performances
Hi everybody
I recently installed an out-of-the box RedHat 6.0.
Since i have a Pentium II, i assume i would gain performance by
compiling with i686 optimizations, and maybe by recompiling kernel, KDE
and so on.......
So my question is : how do i do this ?
Do i need to recompile egcs ? glibs ? in which order ? with which flags
?
and after i compiled the libraries, do i need to change for instance
/etc/ld.so.conf ?
Could somebody point me at some documents ? Or simply answer on the
mailing-list ;-)
I tried to compile both egcs & then glibc-2.1.1, but when i installed
(in /usr/local/lib), and tried to update my /etc/ld.so.conf (adding
/usr/local/lib & then running ldconfig), i just ended crashing my system
(ie whenever i tried to launch a process, i go a message like 'can't
find ld-linux.so.2', and of course nothing did work at all...........)
Since i don't really enjoy booting from rescue disk & moving the
/usr/local/lib dir to something else, i would appreciate some help ;-)
TIA for your answers
Nicolas Weeger
------------------------------
From: "Martin v. Loewis" <[EMAIL PROTECTED]>
Date: Thu, 17 Jun 1999 21:14:22 +0200
Subject: Re: Compilation & performances
> I recently installed an out-of-the box RedHat 6.0. Since i have a
> Pentium II, i assume i would gain performance by compiling with i686
> optimizations, and maybe by recompiling kernel, KDE and so on.......
I think it is questionable whether you'll see any significant
speed-up. The only advantage of doing this: You'll learn a lot about
Linux :-)
> So my question is : how do i do this ?
> Do i need to recompile egcs ? glibs ? in which order ?
You can recompile any of them, in any order. Problems arise if you
compile and install a *different* version from the one you are
currently using.
If you are interested in installing the same version, I'd recommend to
use source RPMs. Please see the RPM documentation on how to rebuild
packages from source RPMs.
You don't need to recompile egcs - unless you want egcs itself to run
faster, or unless you want to upgrade to a more recent version. Please
see the egcs documentation for installation instructions.
You can also configure egcs to use a different architecture by default
- - in which case you don't need to pass additional flags in the later
recompilations.
> I tried to compile both egcs & then glibc-2.1.1, but when i installed
> (in /usr/local/lib), and tried to update my /etc/ld.so.conf (adding
> /usr/local/lib & then running ldconfig), i just ended crashing my system
> (ie whenever i tried to launch a process, i go a message like 'can't
> find ld-linux.so.2', and of course nothing did work at all...........)
In order to re-install glibc, it is not sufficient to just recompile
it: you need to install it. Please read the glibc documentation on how
to do this. Updating from 2.0 to 2.1 is not trivial, so read the
documentation carefully.
> Since i don't really enjoy booting from rescue disk & moving the
> /usr/local/lib dir to something else, i would appreciate some help ;-)
First read all documentation, then start compiling. Reading everything
might well take a day or so.
Regards,
Martin
------------------------------
From: Andreas Jaeger <[EMAIL PROTECTED]>
Date: 17 Jun 1999 21:11:38 +0200
Subject: Re: Compilation & performances
>>>>> Nicolas Weeger writes:
> Hi everybody
> I recently installed an out-of-the box RedHat 6.0.
> Since i have a Pentium II, i assume i would gain performance by
> compiling with i686 optimizations, and maybe by recompiling kernel, KDE
> and so on.......
> So my question is : how do i do this ?
> Do i need to recompile egcs ? glibs ? in which order ? with which flags
> ?
> and after i compiled the libraries, do i need to change for instance
> /etc/ld.so.conf ?
Order doesn't matter, just compile those applications you think could
benefit from it.
> Could somebody point me at some documents ? Or simply answer on the
> mailing-list ;-)
> I tried to compile both egcs & then glibc-2.1.1, but when i installed
> (in /usr/local/lib), and tried to update my /etc/ld.so.conf (adding
> /usr/local/lib & then running ldconfig), i just ended crashing my system
> (ie whenever i tried to launch a process, i go a message like 'can't
> find ld-linux.so.2', and of course nothing did work at all...........)
Read the glibc FAQ first, especially question 3.18, then README and
INSTALL files. You really should install with --prefix=/usr.
> Since i don't really enjoy booting from rescue disk & moving the
> /usr/local/lib dir to something else, i would appreciate some help ;-)
Andreas
- --
Andreas Jaeger [EMAIL PROTECTED] [EMAIL PROTECTED]
for pgp-key finger [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Date: Thu, 17 Jun 1999 23:54:23 ric
Subject: Re: Computing Formula
On Thu, 17 Jun 1999, Rafael Marco de Lucas wrote:
>
> Hello,
> i am writing a program that has as input a formula text (i.e.
> X+log(Y)*(W-Z))
> and i want to compute this formula many times for several values of
X,Y,W,Z,
> how can i do it ? i need it to run quite fast around
300.000times/sec,
> i tried to write a little interpreter of the formula and i stored a
sequence
> of
> functions and arguments and temporal variables used.. so that it works
but i
> guess
> it should be easier,
> i have in mind other posibility that is to write the formula in a
file and
> compile
> it and relink,.. but i dont know well how to do it and maybe it is
even more
> copmlex
> and i have to do exactly the same but using conditions (i.e.
W+Z>4.and.Z<1
> sorry, the
> users will be fortran fans)
What's wrong with fortran? :-)
If it's a fortran formula, why not put a fortran wrapper on it and feed
it to the fortran compiler? Then you can link your code with it and
call it with what values you like. Wrap it with something like:
implicit integer (a-z)
function foo(w,x,y,z)
foo=<formula>
return
end
My fortran might be a little rusty, but I think that should work.
> is there any library to do this ??
> can somebody help me or send me any sugesstion please ?
>
> thanks in advance, Cheers
>
> Rafa
>
> pd. also thanks to the people that send me answers about my last
off-topic
> question
> about the ramdisk
>
>
Lawson
>< Microsoft free environment
This mail client runs on Wine. Your mileage may vary.
___________________________________________________________________
Get the Internet just the way you want it.
Free software, free e-mail, and free Internet access for a month!
Try Juno Web: http://dl.www.juno.com/dynoget/tagj.
------------------------------
From: Praveen Sinha <[EMAIL PROTECTED]>
Date: Fri, 18 Jun 1999 13:08:44 +0530
Subject: Facing problem in compiling
Hi,
I am using red hat linux 5.2
My GCC version is 2.7.2.3
I am stuck at compiling c++ program .I wrote a simple
program to display a string .
when I tried to compile it a file called cc1plus is not
found . What is this file ? I am using iostream.h file which it
could not find.This file does exist in header file list.
Then I downloaded gccfrom one of GNU ftp sites. I downloaded
gcc-2.1-2.2.1.diff.gz
This is a zip file . How should I install this version ,? Please let
me know the step by step procedure of installation
I am a novice to linux OS.
Thanking You
Praveen
------------------------------
From: Rafael Marco de Lucas <[EMAIL PROTECTED]>
Date: Fri, 18 Jun 1999 12:02:46 +0200
Subject: Re: Computing Formula
Hello, thanks for your answer Lawson,
>If it's a fortran formula, why not put a fortran wrapper on it and feed
>it to the fortran compiler? Then you can link your code with it and
>call it with what values you like.
but each time the program runs with a different formula i will have to
compile again the code, and i will have to build a new executable file
and run it, isn't it? i would prefer to run a single executable code
and not to jump from one to other (ie. using atexit from stdlib.h)
is it possible to compile/link just a function and call it from the initial
program (i mean, to add a function to the program while it is running in a way
that i do not need to exit from the initial program ) ?
Cheers,
Rafa
------------------------------
From: "Mark H. Wood" <[EMAIL PROTECTED]>
Date: Fri, 18 Jun 1999 09:15:50 -0500 (EST)
Subject: Re: Computing Formula
On Fri, 18 Jun 1999, Rafael Marco de Lucas wrote:
> but each time the program runs with a different formula i will have to
> compile again the code, and i will have to build a new executable file
> and run it, isn't it? i would prefer to run a single executable code
> and not to jump from one to other (ie. using atexit from stdlib.h)
> is it possible to compile/link just a function and call it from the initial
> program (i mean, to add a function to the program while it is running in a way
> that i do not need to exit from the initial program ) ?
That makes your original question clearer. Yes, it is possible. You need
to fork a command that will build the function as a shared library (even
though in this case your library has only one member!) and then load it
explicitly rather than linking your main program with it. The explicit
dynamic loading functions are in libdl (see `info ld.so libdl` for
details and examples).
You're going to learn a lot about dynamic linking along the way.
Unfortunately you can't learn much more of it from me, because I haven't
done this myself -- not on Unix, anyway. But I know enough about the
facilities to believe that what you want can be done.
- --
Mark H. Wood, Lead System Programmer [EMAIL PROTECTED]
Specializing in unusual perspectives for more than twenty years.
------------------------------
From: Kurt Wall <[EMAIL PROTECTED]>
Date: Fri, 18 Jun 1999 10:16:47 -0600
Subject: Re: Computing Formula
You *may* be able to use the dl interface to do this. Have a look at the
man page, dlopen(3).
Kurt
Also sprach Rafael Marco de Lucas:
>
> Hello, thanks for your answer Lawson,
> >If it's a fortran formula, why not put a fortran wrapper on it and feed
> >it to the fortran compiler? Then you can link your code with it and
> >call it with what values you like.
> but each time the program runs with a different formula i will have to
> compile again the code, and i will have to build a new executable file
> and run it, isn't it? i would prefer to run a single executable code
> and not to jump from one to other (ie. using atexit from stdlib.h)
> is it possible to compile/link just a function and call it from the initial
> program (i mean, to add a function to the program while it is running in a way
> that i do not need to exit from the initial program ) ?
Kurt
- --
Boy! Eucalyptus!
------------------------------
From: [EMAIL PROTECTED]
Date: Fri, 18 Jun 1999 12:46:11 ric
Subject: Re: Facing problem in compiling
On Fri, 18 Jun 1999, Praveen Sinha wrote:
> Hi,
> I am using red hat linux 5.2
> My GCC version is 2.7.2.3
> I am stuck at compiling c++ program .I wrote a simple
> program to display a string .
> when I tried to compile it a file called cc1plus is not
> found . What is this file ? I am using iostream.h file which it
> could not find.This file does exist in header file list.
But does it exist in /usr/include? That is where the compiler looks for
it. To compile programs on a RedHat system you need the -devel rpms for
each of the libraries the program uses. In the case of c++, that is
glibc-devel and libg++-devel, as well as glibc- and libg++. These are
all included in the standard RedHat distro, but you led the install to
believe you didn't want some of them. A compiler is useless without
libraries. You can use rpm -q to see if you have a package, or
rpm -qa |grep <package name> or rpm -qa |less.
>
> Then I downloaded gccfrom one of GNU ftp sites. I downloaded
> gcc-2.1-2.2.1.diff.gz
> This is a zip file . How should I install this version ,? Please let
> me know the step by step procedure of installation
This is not a version of gcc at all. If you have the source of gcc-2.1
installed, you can use this file to update ("patch") it to gcc-2.2.1.
Both these versions of gcc are thoroughly obsolete. There is nothing
wrong with the gcc you have. It just needs libraries and header files
to work with.
>
> I am a novice to linux OS.
>
> Thanking You
> Praveen
Lawson
>< Microsoft free environment
This mail client runs on Wine. Your mileage may vary.
___________________________________________________________________
Get the Internet just the way you want it.
Free software, free e-mail, and free Internet access for a month!
Try Juno Web: http://dl.www.juno.com/dynoget/tagj.
------------------------------
From: [EMAIL PROTECTED]
Date: Fri, 18 Jun 1999 14:32:33 ric
Subject: Re: Computing Formula
I'd be inclined to cobble the whole thing together into a shell script,
compiles, links, and all, but if you want to you can certainly just fork
and exec the fortran compile/link and load it with dlopen, amd it will
run faster that way. Info libc is sometimes more helpful than the man
pages.
Lawson
>< Microsoft free environment
This mail client runs on Wine. Your mileage may vary.
On Fri, 18 Jun 1999, Kurt Wall wrote:
> You *may* be able to use the dl interface to do this. Have a look at
the
> man page, dlopen(3).
>
> Kurt
>
> Also sprach Rafael Marco de Lucas:
> >
> > Hello, thanks for your answer Lawson,
> > >If it's a fortran formula, why not put a fortran wrapper on it and
feed
> > >it to the fortran compiler? Then you can link your code with it and
> > >call it with what values you like.
> > but each time the program runs with a different formula i will have
to
> > compile again the code, and i will have to build a new executable
file
> > and run it, isn't it? i would prefer to run a single executable
code
> > and not to jump from one to other (ie. using atexit from stdlib.h)
> > is it possible to compile/link just a function and call it from
the initial
> > program (i mean, to add a function to the program while it is
running in a
> way
> > that i do not need to exit from the initial program ) ?
>
>
> Kurt
> --
> Boy! Eucalyptus!
>
___________________________________________________________________
Get the Internet just the way you want it.
Free software, free e-mail, and free Internet access for a month!
Try Juno Web: http://dl.www.juno.com/dynoget/tagj.
------------------------------
From: [EMAIL PROTECTED] (H.J. Lu)
Date: Sun, 20 Jun 1999 17:43:00 -0700 (PDT)
Subject: binutils 2.9.4.0.6 is released.
Thanks to Ian, more and more changes in binutils 2.9.1.0.25 are finding
their ways into the official binutils.
Thanks.
- --
H.J. Lu ([EMAIL PROTECTED])
- ---
This is the beta release of binutils 2.9.4.0.6 for Linux, which is
based on binutils 1999 0620 plus various changes. It is purely for
Linux, although it has been tested on Solaris/Sparc and Solaris/x86
from time to time.
Please report any bugs related to binutils 2.9.4.0.6 to [EMAIL PROTECTED]
For arm-linux targets, there are some important differences in behaviour
between these tools and binutils 2.9.1.0.x. The linker emulation name has
changed from elf32arm{26} to armelf_linux{26}. Also, the "-p" flag must be
passed with the linker when working with object files (or static libraries)
created using older versions of the assembler. If this flag is omitted the
linker will silently generate bad output when given old input files.
To get the correct behaviour from gcc, amend the *link section of your specs
file as follows:
*link:
%{h*} %{version:-v} %{b} %{Wl,*:%*} %{static:-Bstatic} %{shared:-shared
} %{symbolic:-Bsymbolic} %{rdynamic:-export-dynamic} %{!dynamic-linker:
- -dynamic-linker /lib/ld-linux.so.2} -X %{mbig-endian:-EB} %{mapcs-26:-m ar
melf_linux26} %{!mapcs-26:-m armelf_linux} -p
Changes from binutils 2.9.4.0.5:
1. Update from binutils 1999 0620.
2. Remove my fwait fix and use the one in cvs.
3. Use "--only-section=section" instead of ""--extract-section=section"
for objcopy.
Changes from binutils 2.9.4.0.4:
1. Update from binutils 1999 0612.
2. Remove various temporary fixes of mine since those bugs are fixed
now.
Changes from binutils 2.9.4.0.3:
1. Update from binutils 1999 0611.
2. Remove my ELF/Alpha bfd changes.
3. Use the local symbol copy fix in binutils 1999 0611.
Changes from binutils 2.9.4.0.2:
1. Update from binutils 1999 0607.
2. Remove my Sparc hacks.
3. Fix local symbol copy.
Changes from binutils 2.9.4.0.1:
1. Update from binutils 1999 0606.
2. Restore relocation overflow checking in binutils 2.9.1.0.25 so that
Linux kernel can build.
3. Fix i370 for the new gas.
Changes from binutils 1999 0605:
1. Fix a -Bsymbolic bug for Linux/alpha.
2. Add ELF/i370.
3. Fix 8/16-bit relocations for i386.
4. Add --redefine-sym=old_form=new_form to objcopy.
5. Add "-j section" for objcopy.
6. Fix i386 disassembler for fwait.
7. Fix a Sparc asm bug.
8. Add Ada demangle support.
9. Fix MIPS/ELF bugs.
10. Add some vxworks suppport.
11. Fix a.out assembler.
The file list:
1. binutils-2.9.4.0.6.tar.gz. Source code.
2. binutils-2.9.4.0.5-2.9.4.0.6.diff.gz. Patch against the previous
beta source code.
3. binutils-2.9.4.0.6-1.src.rpm. Source RPM.
4. binutils-2.9.4.0.6-1.i386.rpm. Binary RPM for RedHat 6.0.
There are also bzip2 versions of tar and diff files.
The primary ftp sites for the beta Linux binutils are:
1. ftp://ftp.varesearch.com/pub/support/hjl/binutils/beta
Thanks.
H.J. Lu
[EMAIL PROTECTED]
06/20/99
------------------------------
From: Rafael Marco de Lucas <[EMAIL PROTECTED]>
Date: Mon, 21 Jun 1999 17:49:08 +0200
Subject: Re: Computing Formula
Hello, thank you very much for your mails, it was really helpfull,
>You're going to learn a lot about dynamic linking along the way.
.. yes, i was reading a little about it, now what i need is to compile
some C and fortran functions to create the .so files,
can somebody tell me how to compile them ? i had a look to the man but
i didnt found it, (i am working on linux and digital-unix)
is there any faqs or little manual about dynamic linking ?
Cheers,
Rafa
------------------------------
From: [EMAIL PROTECTED]
Date: Mon, 21 Jun 1999 18:33:16 ric
Subject: Re: Computing Formula
On Mon, 21 Jun 1999, Rafael Marco de Lucas wrote:
>
> Hello, thank you very much for your mails, it was really helpfull,
> >You're going to learn a lot about dynamic linking along the way.
> .. yes, i was reading a little about it, now what i need is to compile
> some C and fortran functions to create the .so files,
> can somebody tell me how to compile them ? i had a look to the man
but
> i didnt found it, (i am working on linux and digital-unix)
If you have g77, it _should_ behave like gcc: to make a shared library
you just need -shared..
g77 -o foo.so -shared foo.f
or so. If you don't, perhaps there is one at
metalab.unc.edu/pub/Linux/devel/lang/fortran
fort77 looks like it would also work with -shared, is smaller, and does
not interfere with gcc. But it is a Perl script. About half of the
linux system commands are scripts of one sort or another. :-)
> is there any faqs or little manual about dynamic linking ?
> Cheers,
> Rafa
>
All I know about dynamic linking I learned from info -f ld.so and man
dlopen, and the school of hard knocks. :-) The knocks aren't terribly
hard.
>
Lawson
>< Microsoft free environment
This mail client runs on Wine. Your mileage may vary.
___________________________________________________________________
Get the Internet just the way you want it.
Free software, free e-mail, and free Internet access for a month!
Try Juno Web: http://dl.www.juno.com/dynoget/tagj.
------------------------------
End of linux-gcc-digest V1 #394
*******************************
To subscribe to linux-gcc-digest, send the command:
subscribe linux-gcc-digest
in the body of a message to "[EMAIL PROTECTED]". If you want
to subscribe something other than the account the mail is coming from,
such as a local redistribution list, then append that address to the
"subscribe" command; for example, to subscribe "local-linux-gcc":
subscribe linux-gcc-digest [EMAIL PROTECTED]
A non-digest (direct mail) version of this list is also available; to
subscribe to that instead, replace all instances of "linux-gcc-digest"
in the commands above with "linux-gcc".
