Boost 1.32 patches for IBM z/Series S390 Linux
Attached are a couple of patches needed to compile boost 1.32 on S390 running Linux. Can someone apply these to the boost source? -- Jon Biggar Levanta jon@levanta.com --- boost/detail/limits.hpp Thu Feb 26 10:26:47 2004 +++ boost/detail/limits.hpp.new Sat Jan 29 13:56:49 2005 @@ -49,7 +49,7 @@ // The macros are not named appropriately. We don't care about integer // bit layout, but about floating-point NaN (etc.) bit patterns. -#if defined(__sparc) || defined(__sparc__) || defined(__powerpc__) || defined(__ppc__) || defined(__hppa) || defined(_MIPSEB) || defined(_POWER) +#if defined(__sparc) || defined(__sparc__) || defined(__powerpc__) || defined(__ppc__) || defined(__hppa) || defined(_MIPSEB) || defined(_POWER) || defined(__s390__) #define BOOST_BIG_ENDIAN #elif defined(__i386__) || defined(__alpha__) || defined(__ia64) || defined(__ia64__) #define BOOST_LITTLE_ENDIAN --- tools/build/v1/gcc-tools.jam Fri Jan 28 21:00:04 2005 +++ tools/build/v1/gcc-tools.jam.new Fri Jan 28 20:55:18 2005 @@ -297,6 +297,11 @@ flags gcc CFLAGS <address-model>32 : -mgp32 ; flags gcc CFLAGS <address-model>64 : -mgp64 -mlong64 ; } +#S390 +if $(OSPLAT) = 390 +{ + flags gcc CFLAGS <threading>multi : -D_REENTRANT ; +} # # define compiler names if not set:
Jonathan Biggar writes:
Attached are a couple of patches needed to compile boost 1.32 on S390 running Linux. Can someone apply these to the boost source?
Applied, thanks! Let us know if you are interested in testing Boost with your platform on a regular basis to make the next releases more IBM-friendly. -- Aleksey Gurtovoy MetaCommunications Engineering
Aleksey Gurtovoy wrote:
Jonathan Biggar writes:
Attached are a couple of patches needed to compile boost 1.32 on S390 running Linux. Can someone apply these to the boost source?
Applied, thanks!
Let us know if you are interested in testing Boost with your platform on a regular basis to make the next releases more IBM-friendly.
I'd love to, but in the development environment I've got, compiling boost 1.32 for S390 takes about 24 hours! But since we use boost internally in our products, we'll certainly provide feedback and bug reports as we find them. -- Jonathan Biggar jon@levanta.com
Jonathan Biggar writes:
Aleksey Gurtovoy wrote:
Let us know if you are interested in testing Boost with your platform on a regular basis to make the next releases more IBM-friendly.
I'd love to, but in the development environment I've got, compiling boost 1.32 for S390 takes about 24 hours!
If you can afford weekend runs, that would still be useful. -- Aleksey Gurtovoy MetaCommunications Engineering
At 02:27 PM 2/2/2005, Jonathan Biggar wrote:
Aleksey Gurtovoy wrote:
Jonathan Biggar writes:
Attached are a couple of patches needed to compile boost 1.32 on S390 running Linux. Can someone apply these to the boost source?
Applied, thanks!
Let us know if you are interested in testing Boost with your platform on a regular basis to make the next releases more IBM-friendly.
I'd love to, but in the development environment I've got, compiling boost 1.32 for S390 takes about 24 hours!
Could you comment on that, please? Is the machine and/or compiler slow? Are Boost libraries doing something that causes particularly slow compiles? --Beman
Beman Dawes wrote:
Let us know if you are interested in testing Boost with your platform on a regular basis to make the next releases more IBM-friendly.
I'd love to, but in the development environment I've got, compiling boost 1.32 for S390 takes about 24 hours!
Could you comment on that, please? Is the machine and/or compiler slow? Are Boost libraries doing something that causes particularly slow compiles?
We don't have a S390 machine in house, so our builds are done using hercules, which is a S390 emulator. We also do cross-compiling, but getting Redhat rpmbuild to use a cross-compiler is a chore we haven't found worth doing so far. -- Jon Biggar Levanta jon@levanta.com
At 12:23 PM 2/4/2005, Jonathan Biggar wrote:
Beman Dawes wrote:
Let us know if you are interested in testing Boost with your platform on a regular basis to make the next releases more IBM-friendly.
I'd love to, but in the development environment I've got, compiling boost 1.32 for S390 takes about 24 hours!
Could you comment on that, please? Is the machine and/or compiler slow?
Are Boost libraries doing something that causes particularly slow compiles?
We don't have a S390 machine in house, so our builds are done using hercules, which is a S390 emulator.
Ah! An emulator makes the 24 hour builds much more understandable. Thanks for sending in patches for this platform. Every additional platform Boost code runs on is a win for programmers using that platform, and improves general confidence in the portability of Boost code. --Beman
On 2/4/05 4:03 PM, "Beman Dawes" <bdawes@acm.org> wrote:
At 12:23 PM 2/4/2005, Jonathan Biggar wrote: [SNIP]
We don't have a S390 machine in house, so our builds are done using hercules, which is a S390 emulator.
Ah! An emulator makes the 24 hour builds much more understandable. [TRUNCATE]
I wonder if we can find anyone with a real S390 machine?... -- Daryle Walker Mac, Internet, and Video Game Junkie darylew AT hotmail DOT com
Daryle Walker wrote: [...]
I wonder if we can find anyone with a real S390 machine?...
Try http://www.ibm.com/servers/enable/site/testdrive/zseries/ regards, alexander.
Daryle Walker wrote:
I wonder if we can find anyone with a real S390 machine?...
definitely :) We would like to use boost on a zOS mainframe; I could do any support for it and make the necessary adaptions to boost. I've just no idea at the moment how to set up the regression tests! Stefan
At 12:33 PM 2/7/2005, Stefan Slapeta wrote:
Daryle Walker wrote:
I wonder if we can find anyone with a real S390 machine?...
definitely :)
We would like to use boost on a zOS mainframe; I could do any support for it and make the necessary adaptions to boost. I've just no idea at the moment how to set up the regression tests!
Have you looked in the "tools", and particularly, "tools/regression" subdirectories? Start by reading boost-root/tools/regression/xsl_reports/runner/instructions.html --Beman
Beman Dawes wrote:
Have you looked in the "tools", and particularly, "tools/regression" subdirectories? Start by reading boost-root/tools/regression/xsl_reports/runner/instructions.html
As a regression tester I'm very familiar with that :) However, I don't know yet how to configure the whole stuff on a zOS mainframe environment (there's no automatic download/upload, there's a completely different file system etc.) Stefan
Stefan Slapeta wrote: [...]
As a regression tester I'm very familiar with that :) However, I don't know yet how to configure the whole stuff on a zOS mainframe environment (there's no automatic download/upload, there's a completely different file system etc.)
Use z/OS UNIX System Services. http://www.ibm.com/servers/eserver/zseries/zos/unix/ http://www.ibm.com/servers/eserver/zseries/zos/unix/compile/compiler.html regards, alexander.
Alexander Terekhov wrote:
Use z/OS UNIX System Services.
http://www.ibm.com/servers/eserver/zseries/zos/unix/ http://www.ibm.com/servers/eserver/zseries/zos/unix/compile/compiler.html
Do you also know the settings for bjam? Maybe I'll have to ask the boost.build list. Stefan
Beman Dawes wrote:
Have you looked in the "tools", and particularly, "tools/regression" subdirectories? Start by reading boost-root/tools/regression/xsl_reports/runner/instructions.html
The first major problem is that there is only a python 2.2 port for zOS available (http://tinyurl.com/6rfbz). regression.py requires 2.3! Stefan
participants (6)
-
Aleksey Gurtovoy -
Alexander Terekhov -
Beman Dawes -
Daryle Walker -
Jonathan Biggar -
Stefan Slapeta