Upgrade lrcalc to 1.1.7

[tscrim]

Using the upstream based off of here: http://www.math.rutgers.edu/~asbuch/lrcalc/ but use the one in the attachments as the root directory is lrcalc-sage-1.1.7.

This should (help) fix #14625.

Upstream: Reported upstream. No feedback yet.

CC: @sagetrac-sage-combinat @nthiery

Component: packages: standard

Keywords: lrcalc

Author: Travis Scrimshaw

Branch: b124959

Reviewer: François Bissey

Issue created by migration from https://trac.sagemath.org/ticket/16560

[tscrim]

Attachment: lrcalc-1.1.7.tar.gz

[tscrim]

Branch: public/packages/lrcalc_1_1_7-16560

[tscrim]

comment:1

Running tests now...

[tscrim]

comment:2

All (long) tests pass for me.

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

190501c	Changed version number.
69d3573	Update SPKG.txt and checksums.ini.
7134ee1	Changed checksum.

[sagetrac-git]

Commit: 7134ee1

[kiwifb]

comment:4

So the only difference between the two upstream tarballs is that one of them is properly autool=ed and the other is not?

[kiwifb]

Reviewer: François Bissey

[kiwifb]

comment:5

This version has otherwise been in use in sage-on-gentoo since December without any reported issue so I am putting that positive review.

[tscrim]

comment:6

Replying to @kiwifb:

So the only difference between the two upstream tarballs is that one of them is properly autool=ed and the other is not?

Nicolas would be the best one to answer that since he helped create it.

Thanks for doing the review.

[nthiery]

comment:7

Replying to @tscrim:

Replying to @kiwifb:

So the only difference between the two upstream tarballs is that one of them is properly autool=ed and the other is not?

Nicolas would be the best one to answer that since he helped create it.

As far as I remember, yes. And this one is an official release by Anders rather than a tarball produced by myself.

[vbraun]

comment:8

Fails on OSX

libtool: link: gcc -dynamiclib -Wl,-undefined -Wl,dynamic_lookup -o .libs/liblrcalc.1.dylib   -Wl,-force_load,mathlib/.libs/libmath.a -Wl,-force_load,lrcoef/.libs/libsymfcn.a -Wl,-force_load,schubert/.libs/libschub.a   -O2   -install_name  /Users/buildslave-sage/slave/sage_git/build/local/lib/liblrcalc.1.dylib -compatibility_version 2 -current_version 2.0 -Wl,-single_module
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    mathlib/.libs/libmath.a(hashtab.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    mathlib/.libs/libmath.a(list.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    mathlib/.libs/libmath.a(set.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    mathlib/.libs/libmath.a(vectarg.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    mathlib/.libs/libmath.a(vector.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    lrcoef/.libs/libsymfcn.a(symfcn.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    lrcoef/.libs/libsymfcn.a(maple.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    schubert/.libs/libschub.a(schublib.o)
duplicate symbol _lrcalc_panic_frame in:
    mathlib/.libs/libmath.a(alloc.o)
    schubert/.libs/libschub.a(lincomb.o)
ld: 9 duplicate symbols for architecture x86_64
collect2: error: ld returned 1 exit status
make[5]: *** [liblrcalc.la] Error 1
make[5]: Target `all-am' not remade because of errors.
make[4]: *** [all-recursive] Error 1
make[3]: *** [all] Error 2
Error building lrcalc.

Full log: http://build.sagemath.org/sage/builders/%20%20fast%20Volker%20MiniMac%20%28OSX%2010.9%20x86_64%29%20incremental/builds/279/steps/compile/logs/stdio

[kiwifb]

comment:9

Hum.... it seems it cannot cop with lrcalc_panic_frame being defined in mathlib/alloc.h

#include <setjmp.h>

/*  Programs using the lrcalc library should set lrcalc_panic_frame
 *  with setjmp(lrcalc_panic_frame).  The lrcalc library will call
 *  longjmp(lrcalc_panic_frame, 1) if an "out of memory" event occurs.
 */
jmp_buf lrcalc_panic_frame;

All the files for which the linker is complaining about are inheriting the declaration through multiple include statement but are not using it. I guess the proper course of action, from a correctness point of view, would be to create a specific header for it and include it only where needed. Looking for other options....

[kiwifb]

comment:10

@ nthiery: in configure.ac, oops:

AC_PREREQ([2.67])
AC_INIT([lrcalc],[1.1.6],[asbuch at math rutgers edu])

Not hurtful of course.

[kiwifb]

Attachment: lrcalc-1.1.7-jump.patch.gz

solve the lrcalc_panic_frame definition problem

[kiwifb]

comment:11

OK the attached patch solve the problem on OS X, it should work on linux too but I haven't tested it there so far. It is a bit invasive and should be submitted upstream for consideration.

[tscrim]

comment:12

I will test it on my Ubuntu machine later tonight (if no one beats me to it).

[kiwifb]

comment:13

Well I have now checked it builds and installs in sage-on-gentoo on linux and OS X but haven't run tests.

[sagetrac-git]

Changed commit from 7134ee1 to b124959

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

b124959

Added patch file to lrcalc.

[tscrim]

comment:15

Git-a-fied and the spkg and sage checks passed for me.

[tscrim]

comment:16

Thanks for figuring out the OSX François (I couldn't do anything since I don't have access to one).

[kiwifb]

Upstream: Reported upstream. No feedback yet.

[kiwifb]

comment:17

Well it is all a bit sloppy in my opinion. You have this variable that I put in separate header that needs to be global in the programs using it. Problems is it was put everywhere including the library parts. The gnu linker used in linux is infintely tolerant to some stuff and possibly has a way of dealing with it. But in this case I think the OS X linker is correct in saying "please fix this up".

I also sent an email upstream along with the patch.

[tscrim]

comment:18

So do you think we should hold off on merging this until we hear back?

[vbraun]

Changed branch from public/packages/lrcalc_1_1_7-16560 to b124959

[kiwifb]

Changed commit from b124959 to none

[kiwifb]

comment:20

Replying to @tscrim:

So do you think we should hold off on merging this until we hear back?

Double dose of no. One for me and one for Volker :)