fp(y)lll 5.x build fails on Cygwin

[embray]

Relevant portion of the build log:

[fplll-5.0.3rc0] depbase=`echo fplll.lo | sed 's|[^/]*$|.deps/&|;s|\.lo$||'`;\
[fplll-5.0.3rc0] /bin/sh ../libtool  --tag=CXX   --mode=compile g++ -std=c++11 -DHAVE_CONFIG_H -I. -I..  -DFPLLL_DEFAULT_STRATEGY_PATH=\"/home/embray/src/sagemath/sage/local/share/fplll/strategies\" -DFPLLL_DEFAULT_STRATEGY=\"/home/embray/src/sagemath/sage/local/share/fplll/strategies/default.json\" -I./.. -I/home/embray/src/sagemath/sage/local/include/  -I/home/embray/src/sagemath/sage/local/include/ -L/home/embray/src/sagemath/sage/local/lib -O0 -MT fplll.lo -MD -MP -MF $depbase.Tpo -c -o fplll.lo fplll.cpp &&\
[fplll-5.0.3rc0] mv -f $depbase.Tpo $depbase.Plo
[fplll-5.0.3rc0] libtool: compile:  g++ -std=c++11 -DHAVE_CONFIG_H -I. -I.. -DFPLLL_DEFAULT_STRATEGY_PATH=\"/home/embray/src/sagemath/sage/local/share/fplll/strategies\" -DFPLLL_DEFAULT_STRATEGY=\"/home/embray/src/sagemath/sage/local/share/fplll/strategies/default.json\" -I./.. -I/home/embray/src/sagemath/sage/local/include/ -I/home/embray/src/sagemath/sage/local/include/ -L/home/embray/src/sagemath/sage/local/lib -O0 -MT fplll.lo -MD -MP -MF .deps/fplll.Tpo -c fplll.cpp  -DDLL_EXPORT -DPIC -o .libs/fplll.o
[fplll-5.0.3rc0] In file included from ../fplll/nr/../defs.h:55:0,
[fplll-5.0.3rc0]                  from ../fplll/nr/nr.h:8,
[fplll-5.0.3rc0]                  from ../fplll/nr/numvect.h:19,
[fplll-5.0.3rc0]                  from nr/matrix.h:21,
[fplll-5.0.3rc0]                  from util.h:23,
[fplll-5.0.3rc0]                  from fplll.h:25,
[fplll-5.0.3rc0]                  from fplll.cpp:18:
[fplll-5.0.3rc0] ../fplll/nr/../nr/dpe.h: In function 'void dpe_ugly_log(dpe_struct*, const dpe_struct*)':
[fplll-5.0.3rc0] ../fplll/nr/../nr/dpe.h:775:42: error: 'M_LN2' was not declared in this scope
[fplll-5.0.3rc0]    dpe_set_d (x, ((double) DPE_EXP(y)) *  M_LN2 + log (DPE_MANT(y)));
[fplll-5.0.3rc0]                                           ^
[fplll-5.0.3rc0] In file included from ../fplll/nr/nr.h:10:0,
[fplll-5.0.3rc0]                  from ../fplll/nr/numvect.h:19,
[fplll-5.0.3rc0]                  from nr/matrix.h:21,
[fplll-5.0.3rc0]                  from util.h:23,
[fplll-5.0.3rc0]                  from fplll.h:25,
[fplll-5.0.3rc0]                  from fplll.cpp:18:
[fplll-5.0.3rc0] ../fplll/nr/nr_rand.inl: In static member function 'static void fplll::RandGen::init_with_time2()':
[fplll-5.0.3rc0] ../fplll/nr/nr_rand.inl:29:28: error: 'gettimeofday' was not declared in this scope
[fplll-5.0.3rc0]      gettimeofday(&time,NULL);
[fplll-5.0.3rc0]                             ^
[fplll-5.0.3rc0] In file included from bkz.h:24:0,
[fplll-5.0.3rc0]                  from fplll.h:28,
[fplll-5.0.3rc0]                  from fplll.cpp:18:
[fplll-5.0.3rc0] lll.h: In destructor 'fplll::LLLReduction<ZT, FT>::~LLLReduction()':
[fplll-5.0.3rc0] lll.h:39:21: error: 'LDConvHelper' has not been declared
[fplll-5.0.3rc0]    ~LLLReduction() { LDConvHelper::free(); }
[fplll-5.0.3rc0]                      ^
[fplll-5.0.3rc0] make[6]: *** [Makefile:851: fplll.lo] Error 1

This only started since #21221 was merged.

There are two issues here, one of which is definitely an upstream bug, and the other of which is more of an omission but could be addressed upstream as well.

1. Compiling with -std=c++11 fails on Cygwin. I haven't pinpointed the exact cause, but this is not the first time (e.g. #21185) this has been an issue on Cygwin. Really this is a bug with newlib, which has bugs in its headers w.r.t. C standard versions. Compiling with -std=gnu++11 fixes the first two issues above. This could be addressed either in Sage's spkg-install or in a patch to fplll's configure.ac.

2. The second issue is simply a bug. fplll does not define the macro FPLLL_WITH_LONG_DOUBLE on Cygwin, so the definition of the destructor ~LLLReduction() should be be bracketed with #ifdef FPLLL_WITH_LONG_DOUBLE

Upstream bug report for fplll: ​https://github.com/fplll/fplll/issues/212 Upstream bug report for fpylll: ​https://github.com/fplll/fpylll/issues/40

(P.S. I can't wait to get that cygwin build bot up and running :D)

[embray]

Here's an initial fix (against #21675) which includes the patch to lll.h and a patch to spkg-install.

I'm also about to try a patch to the configure.ac for supporting Cygwin.

---

New commits:

​88a7a70	Make sure dowgrading from fplll to libfplll works.
​3951f91	Rename fplll -> libfplll
​618a581	sagelib no longer depends on fplll
​91814c2	Fix issues with building fplll 5.x on Cygwin as reported in

[embray]

Looks like fpylll also has a problem on Cygwin. I'm going to address that as part of this ticket while I'm at it. It appears to be another issue with long double.

[embray]

My patch for the issue in fpylll is here: ​https://github.com/fplll/fpylll/pull/41 I've confirmed that it solves the problem for me and passes the existing tests. I could update this ticket to include the patch, but there's no longer any urgency on this for me (as I can proceed with the build) so I'll wait and see.

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. Last 10 new commits:

​27cc968	Add license notice.
​6c9d6d1	Warn but don't fail if there are no patches to apply
​229199f	Apply sage-apply-patches standard on all packages during spkg-install
​5b00b25	Fixup
​5343f3b	Use sage-apply-patches in more packages.
​2f728e2	Make sure dowgrading from fplll to libfplll works.
​1a372f4	Rename fplll -> libfplll
​32aad0a	sagelib no longer depends on fplll
​719a5ac	Fix issues with building fplll 5.x on Cygwin as reported in
​855752d	Add patch for fpylll that addresses the build failure on Cygwin

[embray]

Added additional patch needed for fpylll to build/work on Cygwin. Rebased on #20692 which I'm making a dependency of this ticket, just so I don't have to rebase that branch again. This ticket can wait on #20692.

[cheuberg]

does not merge with 7.5.rc2

[embray]

Hm, I wouldn't complain but this one never even got a review.

[jpflori]

Please update the branch and I'll review it.

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

​1517ea7	Fix issues with building fplll 5.x on Cygwin as reported in
​24abf0d	Add patch for fpylll that addresses the build failure on Cygwin

[embray]

Rebased.

[jpflori]

Looks ok to me. I trust Erik tested it properly.

[vbraun]

Applying ../patches/ticket-21690.patch
patching file setup.py
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file setup.py.rej
patching file src/fpylll/config.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file src/fpylll/config.pyx.rej
patching file src/fpylll/fplll/bkz.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
16 out of 16 hunks ignored -- saving rejects to file src/fpylll/fplll/bkz.pyx.rej
patching file src/fpylll/fplll/decl.pxd
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file src/fpylll/fplll/decl.pxd.rej
patching file src/fpylll/fplll/enumeration.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
8 out of 8 hunks ignored -- saving rejects to file src/fpylll/fplll/enumeration.pyx.rej
patching file src/fpylll/fplll/gso.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
30 out of 30 hunks ignored -- saving rejects to file src/fpylll/fplll/gso.pyx.rej
patching file src/fpylll/fplll/lll.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
9 out of 9 hunks ignored -- saving rejects to file src/fpylll/fplll/lll.pyx.rej
patching file src/fpylll/fplll/pruner.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
3 out of 3 hunks ignored -- saving rejects to file src/fpylll/fplll/pruner.pyx.rej
patching file src/fpylll/numpy.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
2 out of 2 hunks ignored -- saving rejects to file src/fpylll/numpy.pyx.rej
patching file src/fpylll/util.pyx
Reversed (or previously applied) patch detected!  Assume -R? [n] 
Apply anyway? [n] 
Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file src/fpylll/util.pyx.rej
Error applying '../patches/ticket-21690.patch'

[embray]

Um, this obviously used to work. Was the fpylll package updated?

[git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

[embray]

Updated the branch to drop to patch for fpylll since it is no longer needed since the fpylll version was bumped in #21728.