> On 23/05/2018, at 10:27, François Bissey <frp.bis...@gmail.com> wrote: > > > >> On 23/05/2018, at 10:02, Nils Bruin <nbr...@sfu.ca> wrote: >> >> On Saturday, May 19, 2018 at 12:08:22 AM UTC-7, Antonio Rojas wrote: >> >> >> OK, so how do we see whether the $SAGE_ROOT/sage script needs to be used? >> Can we just see if $SAGE_ROOT/sage is present and executable and if not, use >> $SAGE_LOCAL/bin/sage? The remark by François seems quite to the point: if >> sage is installed system-wide as a package of python2, then invoking python >> via $SAGE_LOCAL/bin/sage --python is unnecessary overkill. Then just calling >> python2 should be sufficient anyway. >> >> The thing is: the kernel.json does *not* do the right thing out of the box >> if you present it to a system jupyter and$SAGE_LOCAL/bin/sage can't run >> without help. It would be nice if we can get our jupyter kernel installation >> to do the right thing in both cases. >> >> >> IMO the right way to fix this would be to make $SAGE_LOCAL/bin/sage not >> complain if $SAGE_ROOT is not set (when used to simply launch sage, it would >> still be needed for build or install operations). But not sure if that would >> currently be possible without major changes. >> >> So what value does SAGE_ROOT have in the Arch distribution? And what value >> does it have in sage-on-gentoo? And where does it get its value? The vanilla >> $SAGE_LOCAL/bin/sage-env only figures out a value if "." or "../.." are good >> choices (and tests that by the presence of a "sage" file and a "build" >> directory, so it really ends up depending on its relative location to the >> "$SAGE_ROOT/sage" script again). >> > > In sage-on-gentoo I ship my own sage-env. I rejected patching or using the > vanilla one > from sage long ago. Too much useless stuff. Currently SAGE_ROOT points to > /usr/share/sage > where long ago I was keeping a copy of the “src” folder. I still keep > “COPYING.txt” there. > Which is probably the only real use of SAGE_ROOT in the whole of > sage-on-gentoo these days. > The other mentions of SAGE_ROOT are in the doc and conditional branches that > are never taken > in sage-on-gentoo. > > 99.999% of the stuff that use SAGE_ROOT is either patched to replace it with > something > “more adequate” or not shipped at all (packaging system). > > François
In fact this thread motivated me to look for the last few places where it looks like there may be a few more instances to eradicate. Either they crept back in since last time I sent patches to bring down the number or I dismissed them as harmless. Why create pickles in SAGE_ROOT/tmp/pickle_jar/ instead of DOT_SAGE/tmp/pickle_jar/ by default (sage_object.pyx)? François -- 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.