Documentation: Tried to make the entry for "for" more novice-friendly.
Novices normally need loads of time before they determine that if the
result of a loop should be a list "makelist" is the right command.
Most other manual entries have the form
Head-Description-See also-Examples.
Here the description is accompanied by examples. Also this manual entry
is long enough that not everybody will reach the traditional location
=> Did add the cross-ref at the head of the entry.
Other changes:
- Omitted the line that tells that the elements of the for loop can
be shuffled around a bit as I agreed with the commentary that this
line isn't necessary
- Moved the general form of loops to a block at the beginning of the
entry (which should make things earlier for the hasty that feel to
just need one example)
- Added a link to the chapter "performance considerations for lists"
(it is easy to create O(n^2) complete constructs using for, after
all)
- Reformulated the first paragraphs a bit.
Hope it was OK not to split this into 8 separate commits (one for each
change) as I originally planned: This approach felt like flooding the
repo with commits this time and - after all only one manual entry was
changed.
f364084...
by
Robert Dodier <email address hidden>
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code
6b0f2a3...
by
Robert Dodier <email address hidden>
In GET-DIRS, omit trailing "/" from path for GCL.
Follow-on work for previous commit 271a4b9;
these commits are intended to make the --list-avail option work for GCL.
This should stop the testbench from failing in ECL.
For some reason ECL wants maxima-objdir to exist or it fails at
more or less random places. I assume this to be a bug in ECL. But
creating the directory at startup seems to resolve the problem so
we probably should do this for now.
271a4b9...
by
Robert Dodier <email address hidden>
Revise GET-DIRS as suggested by Camm Maguire 2017-09-18:
"Suggested patch to get-dirs in init-cl.lisp"