Last commit made on 2018-09-24
Get this branch:
git clone -b master

Branch merges

Branch information


Recent commits

32b78a4... by PeterPall 3 hours ago

The ChangeLog for the current maxima version ;-)

4f32e27... by PeterPall 4 hours ago Added documentation for (hopefully) all user-visible changes back to march 1, 2018.

edf055f... by Raymond Toy <email address hidden> 10 hours ago

Add some more entries from the commit logs

Look through commit logs and find interesting commits to add to the
ChangeLog. I ignored inernal changes.

5cbca8a... by Wolfgang Dautermann <email address hidden> on 2018-09-23

Add missing sheban lines to shell and perl scripts

bc6a78b... by Wolfgang Dautermann <email address hidden> on 2018-09-23

Update with the *user-relevant* changes in the Windows installer.

Many more changes were made, but spelling fixes, updates of bundeled Software
(etc.) are obvious and need not to be mentioned (in my opinion) - as well as
changes in the build system, which does not the affect normal users, which
just download an already compiled installer.

bae7dd2... by Wolfgang Dautermann <email address hidden> on 2018-09-23

Merge branch 'master' of ssh://

39e125c... by PeterPall on 2018-09-18

Make the testsuite tell which lisp the current test run is for

We have several build servers that generate Nightly Builds and run
the test bench, two of them being and

The first step for dealing with them was marking all tests that
failed there as "known bad" so we can see if a commit breaks
something for a lisp that is still used in the wild as the build
servers either use current lisp versions or ones that are used
in currently active versions of Ubuntu. As some tests occur only
sporadically and as testing is a side-effect of packaging and
packaging a non-c++ project needs non-standard measures making
that work required months.

The next (and potentially more long-running whilst being less
time-consuming) step will be finding out which tests occur only
in old lisp versions. With several caveats:

 - Our existing build servers already show that some test failures
   occur only on some computers (Intel/AMD? 32/64 Bit? Signal
   integrity problems? Weird virtual machines? Does all depend
   on the amount of free RAM, perhaps by triggering the gc
   in different places?)
 - Some lisp versions never change their version number even if
   they change the code.
 - And some Ubuntu versions use updated lisps without telling
   the lisp compiler to use a new version number
=> the 2nd step might not be the final one.

9dd0396... by Wolfgang Dautermann <email address hidden> on 2018-09-17

Nightly test: update CMUCL to the latest snapshot (requested by Raymond)

d6d7b1e... by Wolfgang Dautermann <email address hidden> on 2018-09-16

Nightly test: parallel test execution using GNU parallel

Another good speedup would be a parallel build using "make -j",
but then the Make-logfile would be too chaotic, so that is not

d6e4827... by PeterPall on 2018-09-12

Documented the variable MAXIMA_INITIAL_FOLDER.