Opened 11 years ago

Closed 11 years ago

#9435 closed defect (worksforme)

sage-4.5.alpha3 fails on bsd.math

Reported by: rlm Owned by: tbd
Priority: blocker Milestone: sage-4.5.1
Component: packages: standard Keywords:
Cc: Merged in:
Authors: Reviewers:
Report Upstream: N/A Work issues:
Branch: Commit:
Dependencies: Stopgaps:

Status badges


... and I haven't yet tried this version on any other OS X platforms. Reported by William Stein:

This fails completely to build on (OS X 10.6).

/i686-apple-darwin10/4.2.1/../../.. -lSystem
checking for dummy main to link with Fortran 77 libraries... none
checking for Fortran 77 name-mangling scheme... unknown
configure: WARNING: unknown Fortran name-mangling scheme
checking whether sage_fortran appends underscores to external names... unknown
configure: error: cannot use Fortran
Error configuring R.

real    1m10.765s
user    0m10.200s
sys     0m16.450s
sage: An error occurred while installing r-2.10.1.p2

Change History (7)

comment:1 Changed 11 years ago by mpatel

Sage 4.5.alpha4 builds for me on bsd.math with env MAKE="make -j4" make build. I did not set any SAGE_* variables. The long doctests pass.

By the way, Maxima is built after the Sage library:

$ ls -ltr spkg/installed | tail -15
8 -rw-r--r--  1 mpatel  staff  287 Jul  6 23:11 sage-4.5.alpha4
8 -rw-r--r--  1 mpatel  staff  285 Jul  6 23:12 gap-4.4.12.p4
8 -rw-r--r--  1 mpatel  staff  287 Jul  6 23:14 gfan-0.4plus.p1
8 -rw-r--r--  1 mpatel  staff  294 Jul  6 23:14 lcalc-20100428-1.23.p0
8 -rw-r--r--  1 mpatel  staff  288 Jul  6 23:24 maxima-5.20.1.p1
8 -rw-r--r--  1 mpatel  staff  285 Jul  6 23:27 moin-1.9.1.p1
8 -rw-r--r--  1 mpatel  staff  283 Jul  6 23:27 palp-1.1.p3
8 -rw-r--r--  1 mpatel  staff  285 Jul  6 23:28 sagetex-2.2.5
8 -rw-r--r--  1 mpatel  staff  293 Jul  6 23:28 polytopes_db-20100210
8 -rw-r--r--  1 mpatel  staff  284 Jul  6 23:28 pil-1.1.6.p2
8 -rw-r--r--  1 mpatel  staff  284 Jul  6 23:38 scipy-0.7.p5
8 -rw-r--r--  1 mpatel  staff  297 Jul  6 23:38 scipy_sandbox-20071020.p5
8 -rw-r--r--  1 mpatel  staff  289 Jul  6 23:39 sympow-1.018.1.p7
8 -rw-r--r--  1 mpatel  staff  292 Jul  6 23:39 tachyon-0.98beta.p11
8 -rw-r--r--  1 mpatel  staff  286 Jul  6 23:39 weave-0.4.9.p0

But I don't know if this is significant.

comment:2 Changed 11 years ago by rlm

Fortran and R were both updated in sage-4.5.alpha0. Can someone try replacing the R spkg with the old one, to see if it is that or the new fortran spkg which is causing this?

comment:3 follow-up: Changed 11 years ago by drkirkby

#9464 might (probably is) the cause of this.

comment:4 in reply to: ↑ 3 Changed 11 years ago by drkirkby

Replying to drkirkby:

#9464 might (probably is) the cause of this.

John Palmieri has rather dashed that hope, so the reasons for this failure are almost certainly not related to #9464.

comment:5 Changed 11 years ago by drkirkby

Is William's problem reproducible by others? I just took:

  • host
  • sage-4.5.alpha4
  • An updated R, which disables the ICU library when SAGE_FAT_BINARY is set to "yes". This is #9396 (But I did not set SAGE_FAT_BINARY)
  • The updated deps at #9464 which makes 'deps' look a bit cleaner.

Set typed the following.

$ export MAKE="make -j6"
$ make

R builds OK for me on bsd.math. Then Sage completes the build.

dumping object inventory... done
build succeeded.
Build finished.  The built documents can be found in /Users/kirkby/sage-4.5.alpha4/devel/sage/doc/output/html/fr/tutorial
[kirkby@bsd sage-4.5.alpha4]$ make ptestlong
cd spkg && ./install all 2>&1 | tee -a ../install.log
make[1]: Nothing to be done for `all'.

real	0m0.099s
user	0m0.004s
sys	0m0.007s
To install gap, gp, singular, etc., scripts
in a standard bin directory, start sage and
type something like
   sage: install_scripts('/usr/local/bin')
at the Sage command prompt.

To build the documentation, run
   make doc

Sage build/upgrade complete!

I do not believe the updated deps or R will make any difference.

  • The updated R only disables the ICU library if SAGE_FAT_BINARY is set to "yes", but I did not set it to yes, so it would have had no effect.
  • The updated 'deps' makes the dependency of R on Fortran clearer, but it is implied via a long chain rule, so it should not be necessary.

Anyway, it works for me, as it does for Mitesh.


comment:6 Changed 11 years ago by rlm

  • Status changed from new to needs_info

comment:7 Changed 11 years ago by rlm

  • Resolution set to worksforme
  • Status changed from needs_info to closed

Fixed by adding Python to the deps for Fortran, change to deps file is on ticket #9368.

Note: See TracTickets for help on using tickets.