FS#22639 - [sage-mathematics] float numbers support broken?

Attached to Project: Community Packages
Opened by Fabio Zanini (iosonofabio) - Friday, 28 January 2011, 12:38 GMT
Last edited by Thomas Dziedzic (tomd123) - Tuesday, 01 February 2011, 04:21 GMT
Task Type Bug Report
Category Packages
Status Closed
Assigned To Thomas Dziedzic (tomd123)
Architecture x86_64
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:
I installed sage-mathematics from [community]. Everything seems to work except for a very important thing: float numbers! By typing e.g. 0.6 in the sage cli and pressing enter, sage exits with an error report. Please note that internal global variables such as pi do work correctly.

Additional info:
* sage-mathematics 4.6.1-1

Steps to reproduce:
* Install sage-mathematics from [community]
* Open it in a terminal emilator with 'sage'
* Type a float number, e.g. 0.6
* Press ENTER
* Sage exits unexpectedly.
This task depends upon

Closed by  Thomas Dziedzic (tomd123)
Tuesday, 01 February 2011, 04:21 GMT
Reason for closing:  Fixed
Additional comments about closing:  pkgrel 2 should fix this
Comment by Thomas Dziedzic (tomd123) - Friday, 28 January 2011, 14:51 GMT
Could you please post the full error report?
Comment by Fabio Zanini (iosonofabio) - Friday, 28 January 2011, 15:35 GMT
My system is in italian language, in which it sounds as follows:

/opt/sage/local/bin/sage-sage: line 300: 17108 Istruzione non consentita sage-ipython "$@" -i

I attempt a translation. ITA: "Istruzione non consentita" --> ENG:"Instruction not allowed". I had a look at that line, but could not figure out what the problem is.

N.B.: the 17108 number is not fixed, it changes every time... it turns out to be the PID of the "python /opt/sage/local/bin/sage-ipython -i" process. Perhaps we should use python2 here.

P.S.: as mentioned in the wiki - thanks minimizer - the problem does not arise if I compile sage-mathematics from ABS [community], and then install it via pacman -U. This is still annoying, since the compilation time is ~5 hours!
Comment by Thomas Dziedzic (tomd123) - Saturday, 29 January 2011, 02:50 GMT
Hey fabio, can you tell me what architecture you're running on?
Comment by Fabio Zanini (iosonofabio) - Saturday, 29 January 2011, 08:45 GMT
I'm under x86_64, if that's what you mean. More precisely:

$ uname -a
Linux X200 2.6.36-ARCH #1 SMP PREEMPT Mon Jan 24 18:41:03 CET 2011 x86_64 Intel(R) Core(TM)2 Duo CPU L9400 @ 1.86GHz GenuineIntel GNU/Linux
Comment by Thomas Dziedzic (tomd123) - Saturday, 29 January 2011, 16:46 GMT
Have you experienced this with 4.6.0 ?
Comment by Fabio Zanini (iosonofabio) - Saturday, 29 January 2011, 17:26 GMT
Unfortunately, I can't comment on this point. I first installed sage in the version 4.6.1.

In any case, I find interesting that it works correctly if compiled locally with 'makepkg -s', as described in the wiki. Perhaps it is only some dependency missing in the compilation?
Comment by Thomas Dziedzic (tomd123) - Monday, 31 January 2011, 19:59 GMT
ok, I believe this is an upstream bug because it seems that during the build, one of the packages doesn't honor SAGE_FAT_BINARY=1
I will try to work with upstream to sort this out.
Comment by Thomas Dziedzic (tomd123) - Monday, 31 January 2011, 22:07 GMT
I think I have figured this out...
The following is needed to be changed from
export SAGE_FAT_BINARY=1
to
export SAGE_FAT_BINARY='yes'
I'm rebuilding packages at this moment to test.
Comment by Thomas Dziedzic (tomd123) - Monday, 31 January 2011, 23:45 GMT
ok, this has fixed the issue, I will upload the new packages tonight!

Loading...