It's Buridan's ass situation with https://trac.sagemath.org/ticket/34152 - we have two competing solutions there, and cannot move with any of these. Other reviewers of this ticket somehow think it will solve itself.
The root cause of this is the idiotic gnulib state of affairs - gnulib does not provide an install target, it did not have any releases for years, although it is actively updated - it literally wants users to vendor huge chunks of itself, for reasons unclear to me. On Mon, 22 Aug 2022, 23:10 Nils Bruin, <nbr...@sfu.ca> wrote: > For other people who may run into this, it seems that `./bootstrap -D` > avoids this bug. I have no idea what it does > > why? it is documented in bootstrap... -- nor what the original problem is, but it was mentioned somewhere in > another thread where someone expressed their exasperation with another > break in the build-system. > > For Dima/Matthias and the other people responsible for breaking the build > process: > I understand that betas are not release quality, but they are what people > have to work with to base merge trees for tickets on (or, for that matter, > the develop tree in general!). So, if a change is found to break building > under normal circumstances, *back it out* and only re-merge it once it's > solved! It looks like the break was encountered in beta4, so it should have > been rolled back with beta5 until solved. We're on beta8 now. That's 4 beta > releases where people were locked out of being able to build the tree. > That's serious harm to normal development processes. > > I have a very standard build environment of Fedora 36 with all the system > packages installed that are recommended for sage, so if this is breaking > for me, it's breaking for a *lot* of people. > hmm, no, you mist probably don't have gettext installed. > On Monday, 22 August 2022 at 14:24:30 UTC-7 Nils Bruin wrote: > >> Thanks. So, what's the recommended course of action? Wait a month or so >> until this ticket has been merged? [reality is: the work will then probably >> be pushed back to January due to the term kicking in] >> >> On Monday, 22 August 2022 at 14:17:19 UTC-7 Matthias Koeppe wrote: >> >>> This is https://trac.sagemath.org/ticket/34152 >>> >>> On Monday, August 22, 2022 at 2:13:47 PM UTC-7 Nils Bruin wrote: >>> >>>> The error actually already occurs with a >>>> >>>> make distclean >>>> ./bootstrap >>>> >>>> I get: >>>> ---------------------------------------- >>>> ... >>>> m4/sage_spkg_configures.m4:608: the top level >>>> >>>> configure.ac:52: error: possibly undefined macro: AC_LIB_RPATH >>>> If this token and others are legitimate, please use >>>> m4_pattern_allow. >>>> See the Autoconf documentation. >>>> configure:26328: error: possibly undefined macro: AC_LIB_PREPARE_PREFIX >>>> configure:26334: error: possibly undefined macro: AC_LIB_LINKFLAGS_BODY >>>> configure:26342: error: possibly undefined macro: AC_LIB_APPENDTOVAR >>>> ---------------------------------------- >>>> >>>> On Monday, 22 August 2022 at 13:57:48 UTC-7 John H Palmieri wrote: >>>> >>>>> Maybe try some combination of `./bootstrap` and/or `make configure`, >>>>> followed by `./configure && make build`? >>>>> >>>>> -- >>>>> John >>>>> >>>>> >>>>> On Monday, August 22, 2022 at 1:05:07 PM UTC-7 Nils Bruin wrote: >>>>> >>>>>> I get a failure when I do >>>>>> git checkout tags/9.7.beta8 >>>>>> make build >>>>>> >>>>>> I get >>>>>> configure.ac:52: error: possibly undefined macro: AC_LIB_RPATH >>>>>> If this token and others are legitimate, please use >>>>>> m4_pattern_allow. >>>>>> See the Autoconf documentation. >>>>>> configure:26328: error: possibly undefined macro: >>>>>> AC_LIB_PREPARE_PREFIX >>>>>> configure:26334: error: possibly undefined macro: >>>>>> AC_LIB_LINKFLAGS_BODY >>>>>> configure:26342: error: possibly undefined macro: AC_LIB_APPENDTOVAR >>>>>> make[1]: *** [Makefile:365: configure] Error 1 >>>>>> make[1]: Leaving directory '/usr/local/sage/sage-git' >>>>>> make: *** [Makefile:39: base-toolchain] Error 2 >>>>>> >>>>>> Indeed the relevant lines in configure look like a complete mess: >>>>>> >>>>>> ---------------------------------------------- >>>>>> # END SAGE_SPKG_CONFIGURE_GSL >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> AC_LIB_PREPARE_PREFIX >>>>>> AC_LIB_RPATH >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> AC_LIB_LINKFLAGS_BODY(iconv) >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> am_save_CPPFLAGS="$CPPFLAGS" >>>>>> AC_LIB_APPENDTOVAR(CPPFLAGS, $INCICONV) >>>>>> >>>>>> { printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking for iconv" >>>>>> >&5 >>>>>> printf %s "checking for iconv... " >&6; } >>>>>> ---------------------------------------- >>>>>> when I do >>>>>> >>>>>> git checkout tags/9.6 >>>>>> make build >>>>>> >>>>>> things work nicely. >>>>>> >>>>>> This makes it a little hard to move along >>>>>> https://trac.sagemath.org/ticket/32297 (or any sage development, for >>>>>> that matter). >>>>>> >>>>>> Is there something wrong about my setup or is the tree messed up ATM? >>>>>> >>>>> -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/e215c8fe-04cd-4a48-8c8b-37845159c9d5n%40googlegroups.com > <https://groups.google.com/d/msgid/sage-devel/e215c8fe-04cd-4a48-8c8b-37845159c9d5n%40googlegroups.com?utm_medium=email&utm_source=footer> > . > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/CAAWYfq34n3qPiPcLmU_bkXKddLk_2O3%2BSfvitf2-wDtRSpEEng%40mail.gmail.com.
