On Fri, Jan 31, 2020 at 8:47 PM Josh Bevan <jbe...@bu.edu> wrote: > > Thanks for the help so far; a couple comments/questions: > The "spkg-configure.m4" for sqlite worked and was able to use an installed > version, but I then got the same error with brial-1.2.5 which also does not > recognize the --disable-maintainer-mode option. I suspect continuing down > this route would just be playing whack-a-mole with packages that do not > recognize that option. > > The installation is indeed happening over a networked filesystem, so I get > how timestamps may not match. It *is* possible to build from a local disk > from a particular node (/scratch as noted). After reading more carefully > through the install instructions I noted: > "Note that once you have built Sage (by running make, as described below), > you will not be able to move or rename its directory without likely breaking > Sage." > > Ok, so I can't actually have it put the built install on /scratch and move it > where I want after. Then I noted: > "SAGE_BUILD_DIR - the default behavior is to build each spkg in a > subdirectory of $SAGE_ROOT/... If this variable is set, then build in > $SAGE_BUILD_DIR/... instead" > "By default, `make install' will install all the files in ... You can > specify an installation prefix other than ... using `--prefix', for instance > `--prefix=$HOME'." > > So I set the build dir on /scratch too and set "prefix" to be the correct > final install location.
Good! > The install failed part-way through again with: > Error building Sage. > > The following package(s) may have failed to build (not necessarily > during this run of 'make all-start'): > > * package: pari_seadata_small-20090618.p0 > log file: > /scratch/sage/sage-9.0/logs/pkgs/pari_seadata_small-20090618.p0.log > build directory: /scratch/sageb/pari_seadata_small-20090618.p0 > > The log file was empty, hmm. So I just simply ran make again on the hopes > that some transient issue popped up. This time it successfully installed. > > I'm able to load up sage and do the simple suggested "2+2" and "factor(2005)". > However when I do > [jbevan@scc-he1 sage-9.0]$ sage --testall > > I get: > no stored timings available > Running doctests with ID 2020-01-31-13-28-13-f917e3d0. > Git branch: develop > Using --optional=build,dochtml,memlimit,mpir,sage > Doctesting entire Sage library. > Doctesting 3796 files. > sage -t src/sage/__init__.py > [11 tests, 0.05 s] > sage -t src/sage/env.py > [44 tests, 0.36 s] > sage -t src/sage/all.py > [16 tests, 1.34 s] > sage -t src/sage/version.py > [0 tests, 0.00 s] > sage -t src/sage/all_cmdline.py > [0 tests, 0.00 s] > sage -t src/sage/all_notebook.py > [0 tests, 0.00 s] > sage -t src/sage/monoids/automatic_semigroup.py > ********************************************************************** > File "src/sage/monoids/automatic_semigroup.py", line 156, in > sage.monoids.automatic_semigroup.AutomaticSemigroup > Failed example: > N.cardinality() == G5.cardinality() > Expected: > True > Got: > <CSI-?1034h>True This must be something funny with locale, or terminal, or shell... How exactly are you accessing the node? > ********************************************************************** > 1 item had failures: > 1 of 80 in sage.monoids.automatic_semigroup.AutomaticSemigroup > [268 tests, 1 failure, 1.28 s] > > It continues on with more tests, most of which pass and maybe 1 in 10-20 have > a failure. They all look like: > Expected: > n^2*log(x) > Got: > <CSI-?1034h>n^2*log(x) > and > Expected: > pi*x > e > Got: > <CSI-?1034h> pi*x > e > > It seems that these test actually are passing, but some garbage is > pre-pended. What's causing this, and should I care (i.e. will it affect > normal Sage usage)? > > Thanks! > Josh > > On Friday, January 31, 2020 at 6:13:05 AM UTC-5, Dima Pasechnik wrote: >> >> On Fri, Jan 31, 2020 at 3:20 AM Isuru Fernando <isu...@gmail.com> wrote: >> > >> > > " WARNING: 'aclocal-1.15' is missing on your system." Which seems >> > > unimportant since it's followed with the text "You only need it if..." >> > > with reasons that don't pertain to me. >> > >> > One possible reason is that somehow the timestamps for the files got >> > messed up. Easiest solution would be to install automake 1.15. You can >> > also do a make clean and try again. >> > >> if you have or can install sqlite development package on your system, then >> the current Sage beta (9.1.beta2) already has mechanisms to use it, instead >> of building its own. You can actually don't need to switch from your 9.0, >> just save >> https://git.sagemath.org/sage.git/plain/build/pkgs/sqlite/spkg-configure.m4?id=3f9ef468a2a708f7c3607ed4200c912bba50f93d >> to build/pkgs/sqlite/spkg-configure.m4 >> run >> ./boostrap # needs autotools installed >> ./configure >> >> and in the output of the latter you'd see >> >> sqlite-3290000 will not be installed (configure check) >> >> > Isuru >> > >> > On Thu, Jan 30, 2020 at 9:00 PM Josh Bevan <jbe...@bu.edu> wrote: >> >> I'm trying to build Sage 9.0 from source on CentOS 7.5.1804 and it fails >> >> with: >> >> "Error installing package sqlite-3290000" >> >> Here is the ".../sage-9.0/logs/pkgs/sqlite-3290000.log" log file that the >> >> error message says to include: >> >> https://pastebin.com/SCy98UXK >> >> >> >> The two main messages that jump out to me are: >> >> "configure: WARNING: unrecognized options: --disable-maintainer-mode" I >> >> had seen that option listed in a post from a decade ago as a solution to >> >> a Sage Sqlite build error, although I doubt that is still relevant >> >> and >> >> " WARNING: 'aclocal-1.15' is missing on your system." Which seems >> >> unimportant since it's followed with the text "You only need it if..." >> >> with reasons that don't pertain to me. >> >> >> >> For some additional context: >> >> This is being built on a cluster with SGE (Sun Grid Engine) and uses Lmod >> >> for a module system. I have gcc 8.3.0, python 3.6.9, and blis 0.6.0 >> >> (OpenBLAS alternative with better performance) modules loaded, any other >> >> dependencies (e.g. Perl) I've left as the system defaults and they have >> >> not been overridden by the module system. I unpacked the zip and simply >> >> ran "make" with no additional messing with config or modifications. >> >> >> >> If there is any additional info I should supply that would prove helpful >> >> feel free to let me know. > > -- > 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/f4472011-c13f-4580-8b77-08af6acf75fb%40googlegroups.com. -- 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/CAAWYfq02%2Bpsv8qOd%2BdbsL4ep1LcQK8NCOAD71sj9UBos4N1ZiQ%40mail.gmail.com.
