Opened 11 years ago

Closed 11 years ago

Last modified 11 years ago

#10092 closed defect (fixed)

export FC, F77, F90 and F95 from scipy_sandbox

Reported by: drkirkby Owned by: fbissey
Priority: major Milestone: sage-4.6
Component: build Keywords:
Cc: Merged in:
Authors: David Kirkby Reviewers: François Bissey
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: Stopgaps:

Status badges

Description

scipy_sandbox is picking up the wrong Fortran compiler (Sun compiler in /usr/bin) and ignoring the contents of the file $SAGE_LOCAL/bin/sage_fortran This adds patch exports all 4 of these Fortran related variables, so scipy_sandbox uses the right compiler.

For reasons unknown, following the update of Numpy (#9808), it became necessary to export these Fortran variables on both SciPy and scipy_sandbox.

I also cleaned up SPKG.txt, as it was very bare.

Attachments (2)

10092-Resolve-Fortran-issues.patch (3.1 KB) - added by drkirkby 11 years ago.
Export FC, F77, F90 and F95 variables so the right Fortran compiler is found.
10092-remove-Debian-stuff.patch (9.5 KB) - added by drkirkby 11 years ago.
Remove Debian related file in accordance with #5903

Download all attachments as: .zip

Change History (24)

Changed 11 years ago by drkirkby

Export FC, F77, F90 and F95 variables so the right Fortran compiler is found.

comment:1 Changed 11 years ago by drkirkby

  • Status changed from new to needs_review

This is now ready for review. A package can be found at

http://boxen.math.washington.edu/home/kirkby/patches/scipy_sandbox-20071020.p6.spkg

Dave

comment:2 Changed 11 years ago by fbissey

  • Status changed from needs_review to positive_review

It all look good to me. Bring it inline with the other packages and cannot possibly interfere with anything else. It's surprising it didn't need this earlier.

comment:3 Changed 11 years ago by fbissey

  • Status changed from positive_review to needs_work

Oh no! This needs clean up as per #5903 (removal of debian stuff).

comment:4 Changed 11 years ago by drkirkby

Give me 10 minutes and I'll remove the unneeded stuff.

dave

Changed 11 years ago by drkirkby

Remove Debian related file in accordance with #5903

comment:5 Changed 11 years ago by drkirkby

  • Status changed from needs_work to needs_review

comment:6 Changed 11 years ago by drkirkby

I've deleted the Debian related files.

comment:8 follow-up: Changed 11 years ago by fbissey

  • Owner changed from GeorgSWeber to fbissey
  • Reviewers set to François Bissey

patch looks fine, the spkg looks identical to the previous one. Unless I have a caching problem.

comment:9 in reply to: ↑ 8 Changed 11 years ago by drkirkby

Replying to fbissey:

patch looks fine, the spkg looks identical to the previous one. Unless I have a caching problem.

I think you do have a caching problem. I just downloaded it myself with:

wget http://boxen.math.washington.edu/home/kirkby/patches/scipy_sandbox-20071020.p6.spkg

and it looks fine. The Debian stuff is removed from the .spkg.

drkirkby@hawk:/tmp/2/scipy_sandbox-20071020.p6$ ls -l
total 40
drwxr-xr-x   4 drkirkby staff        906 Dec  9  2007 arpack
drwxr-xr-x   3 drkirkby staff       1051 Feb  1  2008 delaunay
drwxr-xr-x   2 drkirkby staff        748 Dec  8  2007 patches
-rwxr-xr-x   1 drkirkby staff        831 Oct  7 00:12 spkg-install
-rw-r--r--   1 drkirkby staff       2279 Oct  7 23:23 SPKG.txt
drkirkby@hawk:/tmp/2/scipy_sandbox-20071020.p6$ 

Dave

comment:10 Changed 11 years ago by drkirkby

The MD5 checksum of the package is e7df163ceb72d4f5aedbcab3ace90ebd

comment:11 follow-up: Changed 11 years ago by fbissey

  • Status changed from needs_review to positive_review

Just got it with wget instead of firefox and it looks fine now! Positive review it is then!

comment:12 in reply to: ↑ 11 Changed 11 years ago by drkirkby

Replying to fbissey:

Just got it with wget instead of firefox and it looks fine now! Positive review it is then!

Thank you. I've been caught out with that before too.

Dave

comment:13 Changed 11 years ago by mpatel

  • Merged in set to sage-4.6.alpha3
  • Resolution set to fixed
  • Status changed from positive_review to closed

comment:14 follow-up: Changed 11 years ago by drkirkby

I note this failed to build on fulvia the Solaris x86 system. The buildbot log shows it failed

sage: An error occurred while installing scipy_sandbox-20071020.p6

though I can't see what anything other than the fact it failed. We might need to open a new ticket for this, or revert this change. I can't possibly understand how it can cause a problem on any platform, but it appears to have done on Solaris x86, which is especially strange given I tested this on OpenSolaris x86.

Dave

comment:15 in reply to: ↑ 14 Changed 11 years ago by fbissey

Replying to drkirkby:

I note this failed to build on fulvia the Solaris x86 system. The buildbot log shows it failed

sage: An error occurred while installing scipy_sandbox-20071020.p6

though I can't see what anything other than the fact it failed. We might need to open a new ticket for this, or revert this change. I can't possibly understand how it can cause a problem on any platform, but it appears to have done on Solaris x86, which is especially strange given I tested this on OpenSolaris x86.

That's indeed very strange. Any chance that we will have access to a build log? CC me on any ticket you open about this please.

Francois

comment:16 Changed 11 years ago by mpatel

Here's the error message:

./spkg-install: FC=/home/buildbot/build/sage/fulvia-1/fulvia_full/build/sage-4.6.alpha3/local/bin/sage_fortran: is not an identifier

I think fulvia's /bin/sh does not support the compact export FOO=bar. Maybe we should just make spkg-install a bash script?

David or François, could one of you please open a 4.6 blocker to fix this?

By the way, adding links to individual spkg logs is already on the Sage Buildbot TODO list.

comment:17 Changed 11 years ago by drkirkby

You beat me to it by 30 seconds. Here is what I had to say

The main buildbot page is here:

http://build.sagemath.org/sage/

The 'Waterfall display' seems the most interesting looking to me:

http://build.sagemath.org/sage/waterfall

  • fulvia is the Solaris x86 machine at Skynet
  • hawk (on far right) is my own personal OpenSolaris machine

I was not initially aware one could get the full log, but it is available. You have to click on the build stdio for the machine.

http://build.sagemath.org/sage/builders/fulvia%20full/builds/6/steps/shell_1/logs/stdio

The relevant bit is:

scipy_sandbox-20071020.p6/.hgignore
Finished extraction
****************************************************
Host system
uname -a:
SunOS fulvia 5.10 Generic_127128-11 i86pc i386 i86pc
****************************************************
****************************************************
CC Version
gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/local/gcc-4.5.1/x86_64-SunOS-core2-sun-ld/libexec/gcc/i386-pc-solaris2.10/4.5.1/lto-wrapper
Target: i386-pc-solaris2.10
Configured with: /usr/local/gcc-4.5.1/src/gcc-4.5.1/configure --enable-languages=c,c++,fortran --with-gnu-as --with-as=/usr/local/binutils-2.20.1/x86_64-SunOS-core2-gcc-4.4.3/bin/as --with-ld=/usr/ccs/bin/ld --with-gmp=/usr/local/gmp-5.0.1/x86_64-SunOS-core2-gcc-4.5.0-abi32 --with-mpfr=/usr/local/mpfr-3.0.0/x86_64-SunOS-core2-gmp-5.0.1-gcc-4.5.0-abi32 --with-mpc=/usr/local/mpc-0.8.2/x86_64-SunOS-core2-mpfr-3.0.0-gmp-5.0.1-gcc-4.5.0-abi32 --prefix=/usr/local/gcc-4.5.1/x86_64-SunOS-core2-sun-ld
Thread model: posix
gcc version 4.5.1 (GCC) 
****************************************************
./spkg-install: FC=/home/buildbot/build/sage/fulvia-1/fulvia_full/build/sage-4.6.alpha3/local/bin/sage_fortran: is not an identifier

real    0m0.017s
user    0m0.001s
sys     0m0.005s
sage: An error occurred while installing scipy_sandbox-20071020.p6

I can now see what is wrong:

export FC="${SAGE_LOCAL}/bin/sage_fortran"
export F77="${SAGE_LOCAL}/bin/sage_fortran"
export F90="${SAGE_LOCAL}/bin/sage_fortran"
export F95="${SAGE_LOCAL}/bin/sage_fortran"

is a bashism - using export on the same line. But the spkg-install script starts with:

#!/bin/sh

The Solaris /bin/sh will not accept that syntax. (On OpenSolaris, a newer shell is used, which is why it worked on my machine).

I've opened #10106 to address this. One could get around it by changing the shell used in the spkg-install, but personally I prefer to write code in a more portable fashion, so I'll change that later today. I've very tired and must get some sleep, so it will have to wait for 8 hours or so, unless someone else wants to beat me to it.

Dave

comment:18 Changed 11 years ago by drkirkby

I've updated #10106 to a blocker, but I have not slept yet and it is 5:25 am here!! I need some sleep.

Dave

comment:19 Changed 11 years ago by drkirkby

I'll do this now, perhaps someone can review it in 15 minutes or so.

comment:20 Changed 11 years ago by drkirkby

#10106 iss now ready for review - I have check it on fulvia to be sure it worked.

Dave

comment:21 Changed 11 years ago by mpatel

  • Merged in sage-4.6.alpha3 deleted

It worked for me.

comment:22 Changed 11 years ago by mpatel

  • Milestone changed from sage-4.6.1 to sage-4.6
Note: See TracTickets for help on using tickets.