#10303: clean up sage-check-64 and use of SAGE64
--------------------------------------------+-------------------------------
Reporter: jhpalmieri | Owner: GeorgSWeber
Type: defect | Status: positive_review
Priority: minor | Milestone: sage-4.6.1
Component: build | Keywords: 64
Author: John Palmieri | Upstream: N/A
Reviewer: David Kirkby, Leif Leonhardy | Merged:
Work_issues: |
--------------------------------------------+-------------------------------
Comment(by leif):
Replying to [comment:34 jhpalmieri]:
> Replying to [comment:33 leif]:
> > Well, the status quo is (or was) to '''do nothing on any operating
system other than MacOS X or SunOS/Solaris'''. (I.e., `SAGE64` '''was'''
effectively ignored w.r.t. `sage-env` / `sage-check-64`. The assumption in
`spkg-install`s is that a non-empty `SAGE64` ''implies'' MacOS X or SunOS;
we previously had tests like `if [ "$SAGE64" = yes -a "$UNAME" = Darwin
]`, until Dave began to port to Solaris, first ''adding'' a test for
SunOS, then dropping the distinction on `$UNAME`, which as a side-effect
then slightly changed the meaning or possible interpretations of the
`spkg-install` files regarded in isolation. Of course removing the OS
dependency '''there''' makes sense, as long as the overall meaning is
kept.)
> >
> > So this patch introduces quite different behavoir on the other
platforms rather than just removing stupid messages (and fixing the
previous ill and incomplete logic).
>
> "Quite different" is not accurate. The only functional difference was
that previously on non OS X or Solaris systems, if SAGE64=yes, then the
file sage-64.txt would not be created or read. Now it is.
And exactly ''that'' is the "problem", besides interpreting `SAGE64` not
set as "use previous setting", which wasn't recorded on e.g. Linux.
If a user now erroneously sets `SAGE64=yes` once, potential problems
persist, where previously doing `unset SAGE64` (or simply leaving /
changing the shell) was sufficient. This has to be documented.
(Though it's [at least meanwhile] simply wrong what some spkgs still do
when `SAGE64` is set [to "yes"], namely e.g. also modifying flags
unrelated to building 32- or 64-bit. We've already removed lots of wrong
''Building a 32-bit version of ...'' messages.)
IMHO setting `SAGE64` (to "yes" or anything else then "no", and perhaps in
the future "ignore") on e.g. Linux should raise an error, nothing else,
since the behavior is - at least currently - more or less "undefined" in
that case.
> Some extra messages are also printed.
I won't mind that.
> (Yes, previously sage-check-64 ignored SAGE64 on Linux boxes, but many
spkg-install files do not.)
Because `SAGE64` is not expected to be set on platforms not defaulting to
32-bit builds; this should be catched in a central place, where also new
OSs would and easily could be added in the first place.
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/10303#comment:36>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en.
