Ticket #2999 (closed defect: duplicate)

Opened 5 years ago

Last modified 14 months ago

Some packages don't respect the CC environment variable

Reported by: dfdeshom Owned by: mabshoff
Priority: critical Milestone: sage-duplicate/invalid/wontfix
Component: build Keywords:
Cc: mjo Work issues:
Report Upstream: N/A Reviewers: Michael Orlitzky, R. Andrew Ohana
Authors: Merged in:
Dependencies: Stopgaps:

Description (last modified by leif) (diff)

Packages which seem to not honor the CC environment variable (they use "gcc"):

flint-1.06.p2
atlas-3.8.1.p1
f2c-20070816.p0
symmetrica-2.0.p2
polybori-0.3.1.p1
rubiks-20070912.p5
zn_poly-0.8.p0
sage-3.0.rc1
gap-4.4.10.p7 // guava3.4
tachyon-0.98beta.p5
palp-1.1.p1

Still not fixed:

  • ratpoints-2.1.3.p2 (see #12682)
  • zn_poly-0.9.p5 (see #12433)

Attachments

pbori-custom_py.patch Download (394 bytes) - added by PolyBoRi 5 years ago.
CC/CXX patch

Change History

comment:1 Changed 5 years ago by PolyBoRi

For PolyBoRi? see the patch for the custom.py file. Feel free to add additional Variables there

Good Night,

Alexander

Changed 5 years ago by PolyBoRi

CC/CXX patch

comment:2 Changed 5 years ago by mabshoff

  • Status changed from new to assigned

comment:3 Changed 4 years ago by mvngu

Ticket #6437 has an updated spkg so that polybori can be built under Solaris.

comment:4 Changed 4 years ago by ddrake

There are other spkgs which also fail to respect CC: from  http://groups.google.com/group/sage-devel/msg/a9192a6b51a74d22 this thread, I see the following spkgs which are not listed above:

  • cliquer-1.2
  • symmetrica-2.0.p4
  • ratpoints-2.1.2.p2

comment:5 Changed 4 years ago by mvngu

Cliquer should respect the CC environment variable now. See ticket #6681.

comment:6 Changed 15 months ago by mjo

  • Cc mjo added
  • Description modified (diff)
  • Report Upstream set to N/A

The f2c package is fixed at #7027, so I've removed it from the list.

comment:7 Changed 15 months ago by mjo

  • Description modified (diff)

symmetrica fixed at #10719, so removing that too.

comment:8 Changed 15 months ago by mjo

  • Description modified (diff)

flint and zn_poly have their own tickets at #7024 and #12433 respectively, so there's no need to duplicate them here.

comment:9 follow-up: ↓ 18 Changed 15 months ago by mjo

  • Description modified (diff)

Polybori also respects $CC now, although I can't pin down the ticket where it went from doesn't-work-at-all to something else.

comment:10 Changed 15 months ago by mjo

  • Description modified (diff)

Tachyon should have been fixed by #9379 and #10681. Waiting on #7069 to confirm on Solaris.

comment:11 follow-up: ↓ 21 Changed 15 months ago by mjo

  • Description modified (diff)

And gap was fixed at #2575 and #4161...

comment:12 Changed 15 months ago by mjo

  • Description modified (diff)

Rubiks fixed at #7036.

comment:13 Changed 15 months ago by mjo

  • Description modified (diff)

Working on palp at #7071.

comment:14 Changed 15 months ago by vbraun

ATLAS supports CC since atlas-3.8.3.p18.

comment:15 Changed 15 months ago by mjo

  • Status changed from new to needs_review
  • Description modified (diff)
  • Milestone changed from sage-5.0 to sage-duplicate/invalid/wontfix

Ok, atlas was fixed in #10226 and we have a ticket for the sage library at #12443. I replaced the initial list for review.

comment:16 follow-up: ↓ 20 Changed 15 months ago by rohana

  • Status changed from needs_review to positive_review
  • Reviewers set to R. Andrew Ohana
  • Authors set to Michael Orlitzky

Yup, there are plenty of tickets regarding all of these packages -- some from me with the clang port, some from David Kirby with the Sun CC port.

comment:17 Changed 15 months ago by mjo

Thanks, I did the same thing with the $CXX ticket at #3000.

comment:18 in reply to: ↑ 9 Changed 15 months ago by AlexanderDreyer

Replying to mjo:

Polybori also respects $CC now, although I can't pin down the ticket where it went from doesn't-work-at-all to something else.

That was #6437 as mentioned above.

comment:19 Changed 15 months ago by jdemeyer

  • Status changed from positive_review to closed
  • Reviewers changed from R. Andrew Ohana to Michael Orlitzky, R. Andrew Ohana
  • Resolution set to duplicate
  • Authors Michael Orlitzky deleted

comment:20 in reply to: ↑ 16 ; follow-up: ↓ 22 Changed 14 months ago by leif

  • Description modified (diff)

Replying to rohana:

Yup, there are plenty of tickets regarding all of these packages -- some from me with the clang port, some from David Kirby with the Sun CC port.

AFAIK at least ratpoints doesn't [yet] have its own ticket; I would have left this ticket open as a meta-ticket until all issues (or spkgs) have really been fixed.

comment:21 in reply to: ↑ 11 Changed 14 months ago by leif

Replying to mjo:

And gap was fixed at #2575 and #4161...

Aha. I knew CC was "intentionally" unset in GAP's spkg-install for a long time (which was annoying anyway), but now I still get:

gcc version 4.6.3 (GCC) 
****************************************************
*WARNING*: Unsetting CC since that tends to break GAP building
*WARNING*: Unsetting CXX since that tends to break GAP building
checking build system type... x86_64-unknown-linux-gnu
checking host system type... x86_64-unknown-linux-gnu
checking target system type... x86_64-unknown-linux-gnu
checking for gcc... gcc
checking for C compiler default output file name... 
configure: error: C compiler cannot create executables
See `config.log' for more details.
Configuration of GAP failed.

real	0m0.793s
user	0m0.160s
sys	0m0.050s
************************************************************************
Error installing package gap-4.4.12.p6
************************************************************************

So if there's been an issue with CC and CXX set, it might have been fixed upstream (I believe so), but it isn't fixed in Sage.

[The problem here simply is that the "default" gcc, which is 4.4.3, doesn't understand some of the options I pass in CFLAGS. GCC 4.6.3, specified in CC, of course does understand them.]

comment:22 in reply to: ↑ 20 Changed 14 months ago by leif

  • Description modified (diff)

Replying to leif:

Replying to rohana:

Yup, there are plenty of tickets regarding all of these packages -- some from me with the clang port, some from David Kirby with the Sun CC port.

AFAIK at least ratpoints doesn't [yet] have its own ticket [...]

This (ratpoints) is now #12682 (needing review).

Note: See TracTickets for help on using tickets.