On 12/29/08, Alexey Beshenov <al at beshenov.ru> wrote:
> Was:
>
> (%i1) declare(j, integer, e, even, o, odd)$
> (%i2) sin(x + (e + 1/2)*%pi);
> (%o2) cos(x)
> (%i3) sin(x + (o + 1/2)*%pi);
> (%o3) - cos(x)
>
> (Is it a regression?)
Looks to me like those simplifications (which existed since time
immemorial) were accidentally lost; I see no reason to disable them.
Can you please make a bug report about that.
> 2. The input lines in Differential.texi are highlighted like
>
> (%i1) @b{@t{x^2*'diff(y,x) + 3*y*x = sin(x)/x;}}
>
> But it (1) conflicts with typesetting of the other sections and
> (2) will be lost after update_examples.
Hmm, it makes sense to display inputs differently than outputs.
But any such markup for input text shouldn't be done by hand.
At this point, I would say just clobber the existing markup.
Maybe at some point someone can equip update_examples with
suitable output (be it for Texinfo markup or a style sheet or whatever).
> (%i5) conjugate (xx + yy);
> (%o5) yy + xx
> (I guess the new result is correct since xx and yy are assumed real.)
Yeah --- I guess the output is OK.
> 4. For simplifications.texi, ./update_examples skips some input lines:
>
> < (%o2) a >= 4
> < (%i3) (b>c)+%; /* add a second, strict inequality */
> I'm not sure what happens here. I thought it's a bug in
> update_examples, possibly caused by my modifications, but the old
> version of update_examples gives the same result. Maybe it is because
> of some problem with my Maxima installation.
My guess is that the trailing comments are causing trouble.
Commentary should be in the body of the text anyway.
Does the problem go away if you move the commentary from
the comments to the body of the text? (i.e. outside of @example)
> 5. The examples in simplex.texi are missing load(simplex) which is bad
> for automatically generated examples and possibly for a reader.
Yes, examples which need it should have the appropriate load(foo) incantation.
> 6. orthopoly.texi splits Maxima sessions into different @example
> blocks. I think @c ===beg=== / @c ===end=== is not correct here.
I guess those examples which extend over two or more @example
blocks should not have beg/end. Maybe there is a way to extend
update_examples to process such examples.
> 7. In stringproc.texi "openw("E:/file.txt");" on input gives "The path
> maxima/doc/info/E:/file.txt does not exist." Maybe it is because of a
> strange path with "E:".
Maybe the file name could be changed to "tmp.txt" (i.e. without
any leading stuff) --- I hope that could work on a greater variety of systems.
> 8. The linear algebra stuff contains a lot of messages from "rat" in
> the output. Should we set ratprint : fase for update_examples?
No --- I would put ratprint:false into those particular examples which need it.
Thanks a lot for working on the documentation.
Robert