Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
Hi, Sorry for not getting pack to this thread before. I'm still having this issue that newly compiled Sage is missing many packages. Overnight I did `make distclean && make --trace --debug=bvjm` on Sage 7.3.beta7 with #20926 applied (gcc6.1 C++ workarounds). I had downgraded my Make to version 4.1. The result is the same: Building the docs crashed due to uninstalled jmol. Other packages are also missing: palp, sympy, database_cremona_ellcurve, etc. The missing packages are listed in the beginning of the build: Makefile:24: update target 'build/make/Makefile' due to: They are also mentioned a few pages below: ... checking for sqrt in -lm... yes checking multiprecision library... MPIR checking BLAS library... atlas checking package versions... 4ti2-1.6.7 alabaster-0.7.8 ... However, none of the not-installed packages are mentioned again in the logs! Those that are installed are of course in the log, a few examples: ... [arb-2.8.1.p0] Finished installing arb-2.8.1.p0.spkg Successfully remade target file '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/arb-2.8.1.p0'. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/atlas-3.10.2.p2' does not exist. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/python2-2.7.10.p2' does not exist. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/readline-6.3.008' does not exist. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/ncurses-5.9.20131221' does not exist. Must remake target '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/ncurses-5.9.20131221'. Makefile:2142: target '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/ncurses-5.9.20131221' does not exist sage-logger -p 'sage-spkg ncurses-5.9.20131221' '/home/jsrn/mat/sage_coding_project/logs/pkgs/ncurses-5.9.20131221.log' ... [ipython_genutils-0.1.0] Finished installing ipython_genutils-0.1.0.spkg Successfully remade target file '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/ipython_genutils-0.1.0'. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/decorator-4.0.6' does not exist. Must remake target '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/decorator-4.0.6'. Makefile:1436: update target '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/decorator-4.0.6' due to: /home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/setuptools-21.0.0 sage-logger -p 'sage-spkg decorator-4.0.6' '/home/jsrn/mat/sage_coding_project/logs/pkgs/decorator-4.0.6.log' [decorator-4.0.6] Found local metadata for decorator-4.0.6 ... [sphinx-1.4.1.p1] Finished installing sphinx-1.4.1.p1.spkg Successfully remade target file '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/sphinx-1.4.1.p1'. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/sagenb-0.13' does not exist. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/flask-0.10.1' does not exist. File '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/werkzeug-0.11.9' does not exist. Must remake target '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/werkzeug-0.11.9'. Makefile:2804: update target '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/werkzeug-0.11.9' due to: /home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/python2-2.7.10.p2 sage-logger -p 'sage-spkg werkzeug-0.11.9' '/home/jsrn/mat/sage_coding_project/logs/pkgs/werkzeug-0.11.9.log' ... Makefile:2828: target '/home/jsrn/mat/sage_coding_project/local/var/lib/sage/installed/zlib-1.2.8.p0' does not exist sage-logger -p 'sage-spkg zlib-1.2.8.p0' '/home/jsrn/mat/sage_coding_project/logs/pkgs/zlib-1.2.8.p0.log' It looks to me like some explicit dependencies to the not-installed packages, as has been discussed with Jmol recently, must have been removed or altered, and that's why Make is not picking up those targets. Any thoughts? Best, Johan leif writes: > Johan S. H. Rosenkilde wrote: >>> Except for the Sage library (and probably docbuilding, not sure), yes. >>> If you explicitly use 'make -j1', then also the Sage library will be >>> built sequentially, still using Python's multi-processing though, as >>> does docbuilding. >> >> I've done `make distclean && make -j1`. Let's see how that one goes... > > If that still omitted standard packages, I'd suggest to experiment with > GNU make's debug options (you can also add '-n' aka '--dry-run'). > > '-d' (='--debug=a') is far too verbose, but e.g. '--debug=bvjm', > probably also with '--trace' may give you a clue what's happening (or > why 'make' /doesn't/ build some targets). > > >
[sage-devel] Re: spkgs assumed for doctesting but not default installed
Johan S. H. Rosenkilde wrote: >> Except for the Sage library (and probably docbuilding, not sure), yes. >> If you explicitly use 'make -j1', then also the Sage library will be >> built sequentially, still using Python's multi-processing though, as >> does docbuilding. > > I've done `make distclean && make -j1`. Let's see how that one goes... If that still omitted standard packages, I'd suggest to experiment with GNU make's debug options (you can also add '-n' aka '--dry-run'). '-d' (='--debug=a') is far too verbose, but e.g. '--debug=bvjm', probably also with '--trace' may give you a clue what's happening (or why 'make' /doesn't/ build some targets). -leif >> Forgot to ask: >> >> How does at all local/var/lib/sage/installed/ look like? > > See below. It doesn't contain the packages I know are missing. I don't > know if other packages are also missing. > > alabaster-0.7.8 > arb-2.8.1.p0 > atlas-3.10.2.p2 > babel-2.3.4 > backports_abc-0.4 > backports_shutil_get_terminal_size-1.0.0.p0 > backports_ssl_match_hostname-3.5.0.1 > boost_cropped-1.58.0 > brial-0.8.4.3 > bzip2-1.0.6-20150304 > cephes-2.8.p1 > certifi-2016.2.28 > cliquer-1.21.p3 > conway_polynomials-0.4.p0 > cycler-0.10.0 > cysignals-1.1.1 > cython-0.24.p0 > dateutil-2.2 > decorator-4.0.6 > docutils-0.12 > ecl-16.1.2.p0 > eclib-20150827 > ecm-6.4.4 > fflas_ffpack-1.6.0.p2 > flask-0.10.1 > flask_autoindex-0.6 > flask_babel-0.9 > flask_oldsessions-0.10 > flask_openid-1.2.5 > flask_silk-0.2 > flint-2.5.2.p0 > freetype-2.6.3 > future-0.15.2 > gap-4.8.3 > gc-7.2f.p0 > gf2x-1.1.p1 > givaro-3.7.1 > glpk-4.55.p1 > gsl-2.1 > iconv-1.14 > imagesize-0.7.1 > iml-1.0.4p1.p1 > ipykernel-4.3.1 > ipython-4.2.0 > ipython_genutils-0.1.0 > itsdangerous-0.24 > jinja2-2.8 > jupyter_client-4.2.2 > jupyter_core-4.1.0 > lcalc-1.23.p14 > libfplll-20160331 > libgap-4.8.3 > libgd-2.1.1.1.p0 > libpng-1.2.51.p0 > linbox-1.3.2.p1 > lrcalc-1.2.p0 > m4ri-20140914 > m4rie-20150908 > markupsafe-0.23 > mathjax-2.5 > matplotlib-1.5.1 > maxima-5.35.1.p2 > mpc-1.0.3.p0 > mpfi-1.5.1 > mpfr-3.1.4 > mpir-2.7.2 > mpmath-0.19 > ncurses-5.9.20131221 > networkx-1.11 > ntl-9.8.1.p0 > numpy-1.11.0.p0 > pari-2.8-2771-gb70b447.p0 > pari_galdata-20080411 > pari_seadata_small-20090618 > patch-2.7.5 > pathlib2-2.1.0 > pathpy-7.1 > pexpect-4.0.1.p1 > pickleshare-0.7.2 > pillow-3.2.0 > pip-8.1.2 > pkgconf-0.9.7 > pkgconfig-1.1.0 > planarity-2.2.0 > ppl-1.1 > ptyprocess-0.5.1 > pygments-2.1.3 > pynac-0.6.7 > pyparsing-2.1.4 > python2-2.7.10.p2 > python_openid-2.2.5 > pytz-2016.4 > pyzmq-15.2.0 > ratpoints-2.1.3.p3 > readline-6.3.008 > rw-0.7 > sagenb-0.13 > scipy-0.17.1 > setuptools-21.0.0 > setuptools_scm-1.11.0 > simplegeneric-0.8.1 > singledispatch-3.4.0.3 > singular-3.1.7p1.p1 > six-1.10.0 > snowballstemmer-1.2.1 > speaklater-1.3 > sphinx-1.4.1.p1 > sqlite-autoconf-313 > symmetrica-2.0.p9 > tachyon-0.98.9.p5 > tornado-4.3 > traitlets-4.2.1 > twisted-16.2.0 > werkzeug-0.11.9 > zeromq-4.0.5 > zlib-1.2.8.p0 > zn_poly-0.9.p11 > zope_interface-4.1.3 > > > Best, > Johan > > > leif writes: > >> Johan S. H. Rosenkilde wrote: $ make --version >>> GNU Make 4.2.1 >>> Built for x86_64-pc-linux-gnu >>> Did you build in parallel, and if so, does probably building sequentially fix the problem (such that the missing packages then get built)? >>> >>> I just did "make"; the default is still sequential build, right? I have >>> no suspect environment variables. >> >> Except for the Sage library (and probably docbuilding, not sure), yes. >> If you explicitly use 'make -j1', then also the Sage library will be >> built sequentially, still using Python's multi-processing though, as >> does docbuilding. >> >> >> Forgot to ask: >> >> How does at all local/var/lib/sage/installed/ look like? >> >> >> -leif > > -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
Hi, I just noticed, in the very beginning when doing "make" after distclean, the build process prints a list of packages: checking package versions... 4ti2-1.6.7 alabaster-0.7.8 arb-2.8.1.p0 ... This list *does* contain the uninstalled packages: database_cremona_ellcurve-20160211 jmol-14.2.11_2015.01.20 palp-2.1.p1 sympy-1.0.p0 ... This list is not piped out to install.log. Best, Johan -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
> Except for the Sage library (and probably docbuilding, not sure), yes. > If you explicitly use 'make -j1', then also the Sage library will be > built sequentially, still using Python's multi-processing though, as > does docbuilding. I've done `make distclean && make -j1`. Let's see how that one goes... > Forgot to ask: > > How does at all local/var/lib/sage/installed/ look like? See below. It doesn't contain the packages I know are missing. I don't know if other packages are also missing. alabaster-0.7.8 arb-2.8.1.p0 atlas-3.10.2.p2 babel-2.3.4 backports_abc-0.4 backports_shutil_get_terminal_size-1.0.0.p0 backports_ssl_match_hostname-3.5.0.1 boost_cropped-1.58.0 brial-0.8.4.3 bzip2-1.0.6-20150304 cephes-2.8.p1 certifi-2016.2.28 cliquer-1.21.p3 conway_polynomials-0.4.p0 cycler-0.10.0 cysignals-1.1.1 cython-0.24.p0 dateutil-2.2 decorator-4.0.6 docutils-0.12 ecl-16.1.2.p0 eclib-20150827 ecm-6.4.4 fflas_ffpack-1.6.0.p2 flask-0.10.1 flask_autoindex-0.6 flask_babel-0.9 flask_oldsessions-0.10 flask_openid-1.2.5 flask_silk-0.2 flint-2.5.2.p0 freetype-2.6.3 future-0.15.2 gap-4.8.3 gc-7.2f.p0 gf2x-1.1.p1 givaro-3.7.1 glpk-4.55.p1 gsl-2.1 iconv-1.14 imagesize-0.7.1 iml-1.0.4p1.p1 ipykernel-4.3.1 ipython-4.2.0 ipython_genutils-0.1.0 itsdangerous-0.24 jinja2-2.8 jupyter_client-4.2.2 jupyter_core-4.1.0 lcalc-1.23.p14 libfplll-20160331 libgap-4.8.3 libgd-2.1.1.1.p0 libpng-1.2.51.p0 linbox-1.3.2.p1 lrcalc-1.2.p0 m4ri-20140914 m4rie-20150908 markupsafe-0.23 mathjax-2.5 matplotlib-1.5.1 maxima-5.35.1.p2 mpc-1.0.3.p0 mpfi-1.5.1 mpfr-3.1.4 mpir-2.7.2 mpmath-0.19 ncurses-5.9.20131221 networkx-1.11 ntl-9.8.1.p0 numpy-1.11.0.p0 pari-2.8-2771-gb70b447.p0 pari_galdata-20080411 pari_seadata_small-20090618 patch-2.7.5 pathlib2-2.1.0 pathpy-7.1 pexpect-4.0.1.p1 pickleshare-0.7.2 pillow-3.2.0 pip-8.1.2 pkgconf-0.9.7 pkgconfig-1.1.0 planarity-2.2.0 ppl-1.1 ptyprocess-0.5.1 pygments-2.1.3 pynac-0.6.7 pyparsing-2.1.4 python2-2.7.10.p2 python_openid-2.2.5 pytz-2016.4 pyzmq-15.2.0 ratpoints-2.1.3.p3 readline-6.3.008 rw-0.7 sagenb-0.13 scipy-0.17.1 setuptools-21.0.0 setuptools_scm-1.11.0 simplegeneric-0.8.1 singledispatch-3.4.0.3 singular-3.1.7p1.p1 six-1.10.0 snowballstemmer-1.2.1 speaklater-1.3 sphinx-1.4.1.p1 sqlite-autoconf-313 symmetrica-2.0.p9 tachyon-0.98.9.p5 tornado-4.3 traitlets-4.2.1 twisted-16.2.0 werkzeug-0.11.9 zeromq-4.0.5 zlib-1.2.8.p0 zn_poly-0.9.p11 zope_interface-4.1.3 Best, Johan leif writes: > Johan S. H. Rosenkilde wrote: >>> $ make --version >> GNU Make 4.2.1 >> Built for x86_64-pc-linux-gnu >> >>> Did you build in parallel, and if so, does probably building >>> sequentially fix the problem (such that the missing packages then get >>> built)? >> >> I just did "make"; the default is still sequential build, right? I have >> no suspect environment variables. > > Except for the Sage library (and probably docbuilding, not sure), yes. > If you explicitly use 'make -j1', then also the Sage library will be > built sequentially, still using Python's multi-processing though, as > does docbuilding. > > > Forgot to ask: > > How does at all local/var/lib/sage/installed/ look like? > > > -leif -- -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: spkgs assumed for doctesting but not default installed
Johan S. H. Rosenkilde wrote: >> $ make --version > GNU Make 4.2.1 > Built for x86_64-pc-linux-gnu > >> Did you build in parallel, and if so, does probably building >> sequentially fix the problem (such that the missing packages then get >> built)? > > I just did "make"; the default is still sequential build, right? I have > no suspect environment variables. Except for the Sage library (and probably docbuilding, not sure), yes. If you explicitly use 'make -j1', then also the Sage library will be built sequentially, still using Python's multi-processing though, as does docbuilding. Forgot to ask: How does at all local/var/lib/sage/installed/ look like? -leif -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
> $ make --version GNU Make 4.2.1 Built for x86_64-pc-linux-gnu > Did you build in parallel, and if so, does probably building > sequentially fix the problem (such that the missing packages then get > built)? I just did "make"; the default is still sequential build, right? I have no suspect environment variables. Best, Johan -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: spkgs assumed for doctesting but not default installed
Johan S. R. Nielsen wrote: > Hi, > > I've now recompiled Sage after rolling back llvm-libs to 3.7.1-1 (was > 3.8.0.1). That fixed the GLIBCXX errors spewing out of ar and ranlib, > but it didn't fix the problem: my newly compiled version of Sage is > still born without most of the default spkgs, including sympy, jmol, > palp, cremona-minidatabase, R, and probably many other. > > I don't see any problems in install.log. llvm is not mentioned anywhere, > and neither is any of the above packages. In logs/pkgs I don't have any > log files for the above packages (I do have for many other packages > though, all of which I guess are installed.) > > I'm at a total loss... $ make --version ? Did you build in parallel, and if so, does probably building sequentially fix the problem (such that the missing packages then get built)? -leif -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
Hi, I've now recompiled Sage after rolling back llvm-libs to 3.7.1-1 (was 3.8.0.1). That fixed the GLIBCXX errors spewing out of ar and ranlib, but it didn't fix the problem: my newly compiled version of Sage is still born without most of the default spkgs, including sympy, jmol, palp, cremona-minidatabase, R, and probably many other. I don't see any problems in install.log. llvm is not mentioned anywhere, and neither is any of the above packages. In logs/pkgs I don't have any log files for the above packages (I do have for many other packages though, all of which I guess are installed.) I'm at a total loss... Best, Johan -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
Hi leif, > Well, it's apparently just that your toolchain isn't using binutils > consistently; usually LLVM's tools get installed as 'llvm-ar' and > 'llvm-ranlib', or they don't get installed into the default PATH. I don't have llvm-ar or llvm-ranlib installed AFAIK (they're in the package llvm which I don't have. I have llvm-libs, lib32-llvm-libs and clang, though. Also locate shows up nothing.). > You may try to reinstall binutils or create appropriate symlinks along > PATH (such that 'ar' and 'ranlib' originate from them), but I have no > idea whether that's at all related to your main problem. "pacman -Qo `which ar`" tells me binutils already owns the main ar (and ranlib). > P.S.: Downgrading GCC of course (in this case at least) also downgrades > libstdc++ (to 6.0.21), so programs and libraries built against the newer > one are borked. If the newer libstdc++ didn't get removed, it's > probably just a matter of symlinking libstdc++.so.6 back to the more > recent one, presumably libstdc++.so.6.0.22 (from GCC 6.1). Oh yeah, that probably explains why llvm expects a GLIBCXX which is not in my current /usr/lib/libstdc++.so.6. I've downgraded llvm and trying a "make distclean && make" again. Best, Johan -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: spkgs assumed for doctesting but not default installed
Johan S. H. Rosenkilde wrote: >>> ar: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required >>> by /usr/lib/libLLVM-3.8.so) >> >> >> Are you using llvm? As far as I know that doesnt' work, though I haven't >> tried it myself. > > I was baffled about that too. I haven't consciously done anything that > would use llvm. I'm using a fully updated Arch linux, but with gcc > rolled back to 5.3. Then I just cloned the git repo and did 'make'. I > don't seem have any environment variables pointing to llvm either. Well, it's apparently just that your toolchain isn't using binutils consistently; usually LLVM's tools get installed as 'llvm-ar' and 'llvm-ranlib', or they don't get installed into the default PATH. You may try to reinstall binutils or create appropriate symlinks along PATH (such that 'ar' and 'ranlib' originate from them), but I have no idea whether that's at all related to your main problem. If 'ar' etc. fail but that error isn't catched that's of course a bug. -leif P.S.: Downgrading GCC of course (in this case at least) also downgrades libstdc++ (to 6.0.21), so programs and libraries built against the newer one are borked. If the newer libstdc++ didn't get removed, it's probably just a matter of symlinking libstdc++.so.6 back to the more recent one, presumably libstdc++.so.6.0.22 (from GCC 6.1). -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
>> ar: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required >> by /usr/lib/libLLVM-3.8.so) > > > Are you using llvm? As far as I know that doesnt' work, though I haven't > tried it myself. I was baffled about that too. I haven't consciously done anything that would use llvm. I'm using a fully updated Arch linux, but with gcc rolled back to 5.3. Then I just cloned the git repo and did 'make'. I don't seem have any environment variables pointing to llvm either. Best, Johan -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
On Thursday, June 30, 2016 at 2:26:47 PM UTC+2, Johan S. R. Nielsen wrote: > > ar: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required > by /usr/lib/libLLVM-3.8.so) Are you using llvm? As far as I know that doesnt' work, though I haven't tried it myself. -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: spkgs assumed for doctesting but not default installed
> sounds like the first "make" failed... That's not clear to me. The compile of Sage itself seemed to work but then make failed during doc-building. This has happened so often during the years I've worked with Sage, however, that I've learned to ignore it. Inspecting the error now, I see that this time it has to do with the missing jmol-package. Inspecting install.log, I it seems that everything before seems to have compiled. jmol, sympy, etc. are never mentioned once in the install.log. A disturbing thing though is that the following line appears a lot: ar: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required by /usr/lib/libLLVM-3.8.so) as well as: ranlib: /usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required by /usr/lib/libLLVM-3.8.so) However, the packages being compiled report that everything went alright afterwards. I don't know what the errors mean. I have the Arch linux package gcc-libs 5.3.0-5 installed. The install.log is huge and I don't know what might be relevant. I've put the whole thing here: http://atuin.dk/install.log.zip Best, Johan -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: spkgs assumed for doctesting but not default installed
sounds like the first "make" failed... On Wednesday, June 29, 2016 at 5:38:24 PM UTC+2, Johan S. R. Nielsen wrote: > > Hi sage-devel > > Today I made a fresh clone of the develop branch, ran "make", and then > tried to run "./sage -t". To my surprise, this didn't work at all! > > First off, the doctest framework wouldn't even run, since sympy wasn't > installed. After doing "./sage -pip install sympy", I could now actually > perform doctests. > > But I then get thousands of failures all around Sage's library, due to > other uninstalled spkgs: jmol, palp, Cremona's mini database, and > others. > > Am I doing something wrong? Or do we currently have a problem with > default spkgs not being installed as they should? > > Best, > Johan > -- 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 post to this group, send email to sage-devel@googlegroups.com. Visit this group at https://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
