On Thu, Sep 21 2023, Ihor Radchenko wrote: > Leo Butler writes: > >>>> +(defconst org-babel-header-args:maxima >>>> + '((batch . :any) >>> >>> Why :any? Only two values are allowed here. >>> >>>> + (graphics-pkg . :any)) >>> >>> Same. The value is a closed list. >> >> I have made the change requested. >> >> However, beyond the documentation value, I don't see this variable's >> values being used. > > In completion. See `pcomplete/org-mode/block-option/src'. > Also, in org-lint. See `org-lint-wrong-header-argument'. Thanks. > >> Fixed. I don't use `apropos', so I didn't see the problem (and I had >> failed to absorb everything in §D.6). > > M-x checkdoc checks for many of the §D.6 conventions. Yes, I discovered that from the documentation. I must have done something wrong, because it didn't work for me when I tried it, but after re-starting emacs it is working. > >> Please see the attached, revised patch. It fixes the problems that you >> have identified–except for the way that the gnuplot terminal is >> determined. I would like to leave that for a subsequent patch, mainly >> because this one has become quite large. > > Ok. Thanks! > > I have one more comment on the tests: > >> +(ert-deftest ob-maxima/batch+verbatim+eof-error () >> + "Exercise the `:batch' header argument. Send invalid input to >> +Maxima." >> + (unwind-protect > > What is the purpose of this `unwind-protect'? > In its current form, it does nothing because no UNWINDFORMS are > provided. I copied the forms from tests I had written for ob-octave, without thinking about why those tests were wrapped in an `unwind-protect' form. The attached patch removes them, otherwise it is identical to the previous patch. Leo