On Friday, November 15, 2019 at 2:30:27 PM UTC-5, Markus Wageringel wrote: > > On CentOS 7.7, I see a large number of doctest failures due to the escape > sequence <CSI-?1034h> appearing in the output. For example: > > ./sage -t --long src/sage/combinat/tableau.py > ********************************************************************** > File "src/sage/combinat/tableau.py", line 2850, in sage.combinat.tableau. > Tableau.row_stabilizer > Failed example: > rs.order() == factorial(3)*factorial(2) > Expected: > True > Got: > <CSI-?1034h>True > ********************************************************************** > > This is a problem with readline in combination with xterm. A workaround is > to unset $TERM before running the tests. A similar problem was described in > #12263. > > This can also be avoided by running configure with > --with-system-readline=no, so that Sage builds readline 6.3. The system > version is 6.2 according to: > > sage: import readline > sage: readline._READLINE_LIBRARY_VERSION > 6.2 > > This problem has only started appearing recently (probably when the system > was upgraded from CentOS 7.6 to 7.7) as Sage always built its own readline > before. > > I am not sure whether this is something for Sage to solve, but maybe this > information is useful in case someone has the same problem. > > Markus >
On Friday, November 15, 2019 at 2:30:27 PM UTC-5, Markus Wageringel wrote: > > On CentOS 7.7, I see a large number of doctest failures due to the escape > sequence <CSI-?1034h> appearing in the output. For example: > > ./sage -t --long src/sage/combinat/tableau.py > ********************************************************************** > File "src/sage/combinat/tableau.py", line 2850, in sage.combinat.tableau. > Tableau.row_stabilizer > Failed example: > rs.order() == factorial(3)*factorial(2) > Expected: > True > Got: > <CSI-?1034h>True > ********************************************************************** > > This is a problem with readline in combination with xterm. A workaround is > to unset $TERM before running the tests. A similar problem was described in > #12263. > > This can also be avoided by running configure with > --with-system-readline=no, so that Sage builds readline 6.3. The system > version is 6.2 according to: > > sage: import readline > sage: readline._READLINE_LIBRARY_VERSION > 6.2 > > This problem has only started appearing recently (probably when the system > was upgraded from CentOS 7.6 to 7.7) as Sage always built its own readline > before. > > I am not sure whether this is something for Sage to solve, but maybe this > information is useful in case someone has the same problem. > > Markus > That 1034 stuff turns on *meta-mode*, which is not a new feature (about twelve years). The ncurses FAQ *Alt-keys do not work in bash* <https://invisible-island.net/ncurses/ncurses.faq.html#bash_meta_mode> gives some background. I'd expect unsetting TERM rather than setting it to some less featureful terminal description would be a step backward. -- 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 [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/5b39fcaa-f606-4e39-9fc9-b19b665e2b4b%40googlegroups.com.
