Last commit made on 2018-04-25
Get this branch:
git clone -b master

Branch merges

Branch information


Recent commits

27510aa... by PeterPall 15 hours ago

Check for installed maxima flavour before falling back to MAXIMA_DEFAULT_LISP

The nightly linux build is split into maxima-gcl, maxima-sbcl, maxima-ecl,
maxima-clisp and in the future maybe others. If maxima-sbcl isn't
installed currently but was the default at compile time we don't want this
to produce an error but want this to cause maxima to use one of the
installed flavours instead.

a308d09... by PeterPall on 2018-04-25

Another test that might fail when using ECL in i386, but apparently not on x64.

72a64a2... by PeterPall on 2018-04-25

The known bad tests order seems to be important.

8b3809f... by PeterPall on 2018-04-24

More errors we trigger in the ECL nightly build.

6096c45... by PeterPall on 2018-04-24

ECL: Another test failure we get with the nightly build.

8280c9b... by PeterPall on 2018-04-24

Marked more tests that fail in ECL depending on the version as "known bad".

485e08d... by PeterPall on 2018-04-23

Marked a few tests as "known bad" for ecl.

Seems like the last few times always the same tests resulted in errors
on ecl.

ee8addc... by PeterPall on 2018-04-23

Draw: Gnuplot supports replotonresize since version 5.0.

Thanks to Robert Dodier for finding this out.

f20ea19... by PeterPall on 2018-04-23

GCL can allow concurrent processes to negotiate how much memory each one uses.

This commit makes this new behavior the default.

Thanks to Camm Maguire for reminding me that this option is needed.

effcd02... by PeterPall on 2018-04-23

An experimental go at the random ECL build failures.

What if we don't use uninitialized memory but exceed the stack limits
instead? I guess there are canarys in this case. But it depends on the
type of canaries and the data we overwrite them with if they ever detect
us overwriting memory.