Update PolyBoRi to upstream release 0.7-1

[AlexanderDreyer]

There was a small bugfix release of PolyBoRi.

Perhaps we can revert #11019.

Current spkg:

• ​http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.7.1.p4.spkg

[AlexanderDreyer]

The package can be reviewed (independently of reverting #11019).

[AlexanderDreyer]

I added a small patch, that makes PolyBoRi? compatible with upcoming M4RI release: ​http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.7.1.p1.spkg

It also fixed the issue that the while site-packages directory was recompiled. (Both backported from upcoming PolyBoRi? 0.8.)

[fbissey]

Probably not the proper place to report this but I have tried to compile polybori against m4ri-20100817 and it doesn't work:

gcc -o Cudd/util/pathsearch.os -c -O1 -march=native -pipe -ggdb -fPIC -DNDEBUG -DHAVE_GD -DHAVE_TR1_UNORDERED_MAP -DPACKED -DHAVE_M4RI -DHAVE_GD -DSIZEOF_VOID_P=8 -DSIZEOF_LONG=8 -DHAVE_IEEE_754 -DBSD -I/usr/include/python2.7 -Ipolybori/include -ICudd/obj -ICudd/util -ICudd/cudd -ICudd/mtr -ICudd/st -ICudd/epd Cudd/util/pathsearch.c
In file included from PyPolyBoRi/main_wrapper.cc:25:0:
/usr/include/m4ri/packedmatrix.h:75:3: error: 'mmb_t' does not name a type
/usr/include/m4ri/packedmatrix.h:105:3: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h: In function 'void mzd_row_swap(mzd_t*, size_t, size_t)':
/usr/include/m4ri/packedmatrix.h:174:3: error: 'word' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:174:9: error: 'a' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:174:16: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:175:9: error: 'b' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:175:16: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:176:8: error: expected ';' before 'tmp'
/usr/include/m4ri/packedmatrix.h:177:8: error: expected ';' before 'mask_begin'
/usr/include/m4ri/packedmatrix.h:178:8: error: expected ';' before 'mask_end'
/usr/include/m4ri/packedmatrix.h:181:5: error: 'tmp' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:182:21: error: 'mask_begin' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:191:29: error: 'mask_end' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:195:5: error: 'tmp' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:196:21: error: 'mask_begin' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:196:56: error: 'mask_end' was not declared in this scope
/usr/include/m4ri/packedmatrix.h: In function 'void mzd_col_swap_in_rows(mzd_t*, size_t, size_t, size_t, size_t)':
/usr/include/m4ri/packedmatrix.h:242:31: error: 'RADIX' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:247:3: error: 'word' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:247:8: error: expected ';' before 'a'
/usr/include/m4ri/packedmatrix.h:252:11: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h:253:11: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h:255:7: error: 'base' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:255:18: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:256:16: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h:257:16: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h:258:15: error: 'b' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:258:21: error: 'x' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:258:26: error: 'ai' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:258:38: error: 'bi' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:263:9: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h:264:9: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h:269:7: error: 'base' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:269:17: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:270:7: error: 'a' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:271:7: error: 'b' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:273:17: error: 'b_bm' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:274:17: error: 'a_bm' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:283:7: error: 'base' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:283:17: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:284:7: error: 'a' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:285:7: error: 'b' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:287:17: error: 'b_bm' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:288:17: error: 'a_bm' was not declared in this scope
/usr/include/m4ri/packedmatrix.h: At global scope:
/usr/include/m4ri/packedmatrix.h:308:15: error: 'BIT' does not name a type
/usr/include/m4ri/packedmatrix.h:324:86: error: 'BIT' does not name a type
/usr/include/m4ri/packedmatrix.h: In function 'void mzd_write_bit(mzd_t*, size_t, size_t, int)':
/usr/include/m4ri/packedmatrix.h:326:16: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:326:42: error: 'RADIX' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:326:73: error: 'SET_BIT' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:328:16: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:328:42: error: 'RADIX' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:328:73: error: 'CLR_BIT' was not declared in this scope
/usr/include/m4ri/packedmatrix.h: In function 'void mzd_row_add_offset(mzd_t*, size_t, size_t, size_t)':
/usr/include/m4ri/packedmatrix.h:362:38: error: 'RADIX' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:364:3: error: 'word' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:364:9: error: 'src' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:364:18: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:365:9: error: 'dst' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:365:18: error: 'struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:370:8: error: expected ';' before 'temp'
/usr/include/m4ri/packedmatrix.h:372:5: error: 'temp' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:372:60: error: 'RIGHTMOST_BITS' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:373:13: error: 'temp' was not declared in this scope
/usr/include/m4ri/packedmatrix.h: At global scope:
/usr/include/m4ri/packedmatrix.h:556:1: error: 'BIT' does not name a type
/usr/include/m4ri/packedmatrix.h:738:15: error: 'word' does not name a type
/usr/include/m4ri/packedmatrix.h:772:94: error: 'word' has not been declared
/usr/include/m4ri/packedmatrix.h: In function 'void mzd_xor_bits(const mzd_t*, size_t, size_t, int, int)':
/usr/include/m4ri/packedmatrix.h:773:3: error: 'word' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:773:9: error: 'temp' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:778:23: error: 'RADIX' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:780:16: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:787:8: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:788:8: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h: At global scope:
/usr/include/m4ri/packedmatrix.h:802:94: error: 'word' has not been declared
/usr/include/m4ri/packedmatrix.h: In function 'void mzd_and_bits(const mzd_t*, size_t, size_t, int, int)':
/usr/include/m4ri/packedmatrix.h:803:3: error: 'word' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:803:9: error: 'temp' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:808:23: error: 'RADIX' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:810:16: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:817:8: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:818:8: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h: In function 'void mzd_clear_bits(const mzd_t*, size_t, size_t, int)':
/usr/include/m4ri/packedmatrix.h:832:3: error: 'word' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:832:8: error: expected ';' before 'temp'
/usr/include/m4ri/packedmatrix.h:837:23: error: 'RADIX' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:839:5: error: 'temp' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:839:16: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:843:8: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:848:8: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:848:32: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:848:51: error: 'ONE' was not declared in this scope
/usr/include/m4ri/packedmatrix.h:849:8: error: 'const struct mzd_t' has no member named 'rows'
/usr/include/m4ri/packedmatrix.h:849:32: error: 'const struct mzd_t' has no member named 'rows'

Are you meaning a later release of m4ri that will fix this?

[AlexanderDreyer]

It seems that the header m4ri/packedmatrix.h is not self-contained in that version. I'll add a patch for PolyBoRi which includes m4ri/m4ri.h instead. I'll also re-bundle the spkg with the Sage release. (Maybe #11331 could be release first.)

[AlexanderDreyer]

@fbissey Switching to the development version at bibucket should also fix that issue.

[fbissey]

I didn't know about the polybori page on bitbucket, thanks for pointing it out. Actually if you make the release available on bitbucket it would be easier to automatically fetch than sourceforge.

Thanks for the patch.

[AlexanderDreyer]

I meant the bitbucket repo of m4ri, but of course you may also have a look at PolyBoRi? there. The "Downloads" section of bitbucket is automatically generated from mercurial tags. I just realized that the releases 0.7.0 and 0.7.1 are missing albeit the tags do exist. Maybe I'll have to contact the support there.

[fbissey]

Just to be picky the patch hasn't been made with the git compatibility mode which means that "plain patch" doesn't like it. It shouldn't get in the way of merging this ticket I am just mentioning.

[AlexanderDreyer]

Backporting m4ri header issue patch from upcoming PolyBoRi 0.8

[AlexanderDreyer]

I hope this patch applies better.

[fbissey]

Thx.

[AlexanderDreyer]

This should be rebased on the spkg from #11331 when the latter is accepted.

[AlexanderDreyer]

After Sage 4.7 being out (a while, I must admit), I rebased it the spkg to the most recent one from Sage. Please have a look at it: ​http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.7.1.p2.spkg

[AlexanderDreyer]

BTW all tests pass on a SLES 11 Linux (64 bits).

[fbissey]

I should do some test on vanilla sage but I shamelessly decided to skip 0.7 altogether in favor of that one for sage-on-gentoo for sage-4.7 and over. Works a treat on x86 and amd64. Still an annoying problem on macos that is not mirrored in the vanilla version for some reason.

[AlexanderDreyer]

I have often several times that installing one package breaks some completely unrelated tests, while replacing the spkg and building all Sage from scratch works fine. Does sage -ba always suffice?

[fbissey]

Replying to AlexanderDreyer:

I have often several times that installing one package breaks some completely unrelated tests, while replacing the spkg and building all Sage from scratch works fine. Does sage -ba always suffice?

You are not entirely coherent there. Anyway in most case sage -b will suffice, sage -ba if you want to be sure for the one with real widespread use (pari for example). One thing that sage -b(a) doesn't do is rebuild dependent packages I think. So if update numpy I should also rebuild scipy. If I am not mistaken sage -b(a) won't do that for you.

If your comment was aimed at the problems I mentioned I have in sage-on-gentoo on macos, don't worry that's irrelevant to the present issue. Just expect me to mention it on bitbucket.

[AlexanderDreyer]

Backport from 0.8: polybori.plot uses jinja2, where available

[AlexanderDreyer]

I backported another thing from 0.8: polybori.plot may use jinja2, where available ​http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.7.1.p3.spkg

See the attached patch for details.

[kcrisman]

Passes relevant tests on PPC OS X 10.4.

[AlexanderDreyer]

To push this forward: it builds/installs on Solaris (t2) and passes all tests in sage/rings/polynomial/. (I tested it with the official Sage 4.7 build for Solaris.)

[burcin]

I've also used this version for a while and it seems to work fine. Given the reports of extensive tests above, I am switching this to positive review.

[jdemeyer]

The spkg contains files which are not world-readable. In order to avoid another #11660, this should be fixed by a simple

chmod -R "og+rX" .

[jdemeyer]

See also #11664 to do the same in sage-4.7.1 with polybori-0.7.

[AlexanderDreyer]

I rebased the package on #11664 and did the chmod here: ​http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.7.1.p4.spkg

[malb]

Files are world-readable now. returning to positive review.