robert_dodier [Sun, 19 Nov 2006 16:48:01 +0000 (19 16:48 +0000)]
In PART+ and PART*, only copy the input expression instead of calling
$RATEXPAND or $FACTOR, respectively.
Rationale: (1) $RATEXPAND and $FACTOR change the structure of the expression under
consideration, which (1a) is inconsistent with the spirit of formal pattern matching,
and (1b) makes it even more difficult to predict what the result will be;
(2) $RATEXPAND and $FACTOR are potentially very expensive operations,
and it is hard to predict which ones will be expensive -- some simple
expressions may take a very long time.
COPY-TREE is called instead of those operations because PART+ and PART*
or their callees sometimes (not sure about exactly what circumstances)
modify the input expression destructively. (So we cannot simply omit the
$RATEXPAND and $FACTOR calls; they must replaced by something which
yields a fresh expression.) It is conceivable that COPY-TREE could
sometimes or always be avoided, but for now, just play it safe.
robert_dodier [Sun, 19 Nov 2006 15:34:15 +0000 (19 15:34 +0000)]
Make REMOVERULE respect *BUILTIN-$RULES*, which is a list of rules
which are to be preserved by kill. (*BUILTIN-$RULES* was introduced as part
of the revision of kill in r1.22 src/suprv1.lisp, April 2005.)
robert_dodier [Sun, 19 Nov 2006 08:15:35 +0000 (19 08:15 +0000)]
Make tests of rule functions yield expected results even in
the presence of other rules (do not assume all rules have been killed off).
robert_dodier [Sun, 19 Nov 2006 08:11:49 +0000 (19 08:11 +0000)]
Put kill(all) at beginning of file.
robert_dodier [Sun, 19 Nov 2006 08:00:40 +0000 (19 08:00 +0000)]
Test 44 in rtestsum.mac which failed previously now succeeds.
(Failure to match expected output was due to Maxima evaluation rules.
Test has been revised, using define to make a function instead of :=,
so now it yields the expected output.)
robert_dodier [Sun, 19 Nov 2006 07:55:25 +0000 (19 07:55 +0000)]
Put another example into comment header.
riotorto [Sun, 19 Nov 2006 07:52:31 +0000 (19 07:52 +0000)]
Adding new file lbfgs.texi in Spanish
riotorto [Sat, 18 Nov 2006 23:23:53 +0000 (18 23:23 +0000)]
Adding statistical inference package: stats
riotorto [Sat, 18 Nov 2006 19:48:50 +0000 (18 19:48 +0000)]
Correcting some errors
belanger [Thu, 16 Nov 2006 19:39:20 +0000 (16 19:39 +0000)]
Added an "inline" option.
robert_dodier [Thu, 16 Nov 2006 05:30:55 +0000 (16 05:30 +0000)]
New .texi file for lbfgs package.
belanger [Wed, 15 Nov 2006 18:02:55 +0000 (15 18:02 +0000)]
Make sure right font is used in session outputs.
belanger [Wed, 15 Nov 2006 17:54:51 +0000 (15 17:54 +0000)]
(tex-stripdollar): Update function using Andrey Grozin's texmacs fix.
riotorto [Tue, 14 Nov 2006 19:09:00 +0000 (14 19:09 +0000)]
Updating translations and some improvements
rtoy [Tue, 14 Nov 2006 17:10:31 +0000 (14 17:10 +0000)]
src/limit.lisp:
o The change in rev 1.24 breaks limit(exp(-1/x)/x^4,x,0,plus). Remove
that change in lhop-numden and do something different in
%e-right-placep.
tests/rtest16.mac:
o Add a test for limit(exp(-1/x)/x^4,x,0,plus).
robert_dodier [Tue, 14 Nov 2006 05:50:51 +0000 (14 05:50 +0000)]
Committing quaternion.mac verbatim as received from Gosei Furuya 2006/11/13.
rtoy [Mon, 13 Nov 2006 22:36:53 +0000 (13 22:36 +0000)]
Add a few comments, fix a typo.
rtoy [Mon, 13 Nov 2006 19:38:21 +0000 (13 19:38 +0000)]
Rename stringdisp to $stringdisp. Update files accordingly, and
update documentation to match.
rtoy [Mon, 13 Nov 2006 18:29:57 +0000 (13 18:29 +0000)]
Fix typo (wont).
vvzhy [Sun, 12 Nov 2006 18:24:09 +0000 (12 18:24 +0000)]
Package CHM (Compressed HTML) instead of plain HTML
vvzhy [Sun, 12 Nov 2006 18:22:43 +0000 (12 18:22 +0000)]
On Windows use maxima.chm (Compressed HTML) as Maxima Help instead of plain HTML
vvzhy [Sun, 12 Nov 2006 14:59:47 +0000 (12 14:59 +0000)]
CHM files are created for all languages instead
vvzhy [Sun, 12 Nov 2006 14:56:58 +0000 (12 14:56 +0000)]
CHM file renamed header.chm -> maxima.chm
vvzhy [Sun, 12 Nov 2006 10:02:40 +0000 (12 10:02 +0000)]
HTML Help Workshop project files for CHM creation
rtoy [Fri, 10 Nov 2006 21:50:48 +0000 (10 21:50 +0000)]
Bug [ 965926 ] trigsimp exponentially slow on lists
Modify trigsimp3 to handle each element of the list one at a time
instead of trying to do the whole list all at once. This provide
linear growth now.
rtoy [Fri, 10 Nov 2006 19:45:32 +0000 (10 19:45 +0000)]
Bug [
1469411 ] limit(t^2*exp(-4*t/3-8*exp(-t)),t,inf) is wrong
Bug [
1594330 ] limit(x*(atan(x)-%pi/2),x,inf)
These are bugs in the L'Hospital heuristic. Maxima was sometimes
converting the form n/d to (1/d)/(1/n) which actually makes the
problem harder.
src/limit.lisp (lhop-denom):
o If the numerator has an exponential and the denominator is a
polynomial, don't invert them.
o If the denominator has an exponential and the numerator is a
polynomial, don't invert them.
o If the numerator involves asin, acos, or atan, and the denominator
or the reciprocal of the denominator is a polynomial, don't invert
them.
tests/rtest16.mac:
o Add tests for
1469411 and
1594330.
rtoy [Fri, 10 Nov 2006 16:28:53 +0000 (10 16:28 +0000)]
Bug [
1590528 ] Does debugmode(true) actually do anything?
src/merror.lisp:
o Remove "Quitting"
o When entering the debugger, tell user to enter :h for help
src/mdebug.lisp:
o Clean up help output so things line up nicely
o Clean up help strings so line up nicely. (Is there some other way
to do this than inserting spaces into the help strings?)
riotorto [Thu, 9 Nov 2006 23:10:27 +0000 (9 23:10 +0000)]
Updating translations
riotorto [Thu, 9 Nov 2006 22:09:08 +0000 (9 22:09 +0000)]
Defined 'status' function as suggested by Barton in the
mailing list (sept/25/2005) to avoid error message when
loading this package
rtoy [Thu, 9 Nov 2006 14:00:24 +0000 (9 14:00 +0000)]
Bug [ 626721 ] logarc of atan2 wrong
Fixed as suggested
src/logarc.lisp:
o Use suggested formula (with slight correction).
tests/rtest16.mac:
o Add test for this.
rtoy [Thu, 9 Nov 2006 03:26:46 +0000 (9 03:26 +0000)]
Add tests from bug
1548643, 671574,
1152668, 593344.
rtoy [Thu, 9 Nov 2006 02:58:48 +0000 (9 02:58 +0000)]
Bug [ 626697 ] limit(atan2(y,x),y,minf) => -FALSE
src/limit.lisp:
o Fixed as suggested
tests/rtest16.mac:
o Add test.
rtoy [Wed, 8 Nov 2006 15:41:35 +0000 (8 15:41 +0000)]
o Add one more integral from bug
1073338
o Load sqdnst so we can use sqrtdenest to simplify the result of
integrate(1/((x-3)^4+1/2),x,0,1).
robert_dodier [Wed, 8 Nov 2006 05:04:35 +0000 (8 05:04 +0000)]
New item %i (imaginary unit). Apparently undocumented before.
dtc [Wed, 8 Nov 2006 03:59:36 +0000 (8 03:59 +0000)]
* tex-bigfloat: be consistent with the output TeX case, irrespective of
the CL case mode - for the support of the Scieneer CL lower case mode and
the ACL modern mode.
dtc [Wed, 8 Nov 2006 03:40:02 +0000 (8 03:40 +0000)]
* Fix a bunch of type errors that show up when compiled with type checking
enabled on the Scieneer CL and CMU CL.
* Symbol name downcasing - for support of the Scieneer lower case mode and
the ACL modern mode.
dtc [Wed, 8 Nov 2006 03:37:15 +0000 (8 03:37 +0000)]
* Symbol name downcasing - for support of the Scieneer lower case mode and
the ACL modern mode.
dtc [Wed, 8 Nov 2006 03:34:48 +0000 (8 03:34 +0000)]
* File new_cartan_test4.bat has been renamed to new_cartan_test4.mac.
dtc [Wed, 8 Nov 2006 03:33:57 +0000 (8 03:33 +0000)]
* Fix some spelling typos. File cartan_init.bat has been renamed to
diff_form.mac, and new_cartan_test4.bat to new_cartan_test4.mac.
dtc [Wed, 8 Nov 2006 03:31:29 +0000 (8 03:31 +0000)]
* Symbol case downcasing - for support of the Scieneer lower case mode and
the ACL modern mode.
rtoy [Tue, 7 Nov 2006 15:54:11 +0000 (7 15:54 +0000)]
Bug [
1073338 ] integrate yields incorrect result on rational function
This is a workaround for this bug. We basically simplify some of the
intermediate results so that maxima doesn't get confused. I think
this is a gcd problem.
src/residu.lisp:
o Fix a few comments and add a few more.
o For the case of simple poles, use RES1 instead of the more general
$RESIDUE. (Workaround 1)
o In RES1, call $RECTFORM on each pole to simplify somewhat so that
maxima doesn't get confused. (Workaround 2)
tests/rtest15.mac:
o Add a test for
1073338, but disable it for now. The integral takes
a bit of time to compute, and the result is quite messy. If
sqrtdenest were available, the result can be simplified quite a bit.
belanger [Mon, 6 Nov 2006 21:11:58 +0000 (6 21:11 +0000)]
(inferior-maxima-wait-for-output): Avoid infinite loop by replacing
"while" by "when".
(maxima-mode): Comment out old lines.
rtoy [Mon, 6 Nov 2006 15:28:47 +0000 (6 15:28 +0000)]
src/defint.lisp
src/residu.lisp
o Update and correct some of the comments.
rtoy [Mon, 6 Nov 2006 02:51:38 +0000 (6 02:51 +0000)]
src/defint.lisp:
src/residu.lisp:
o Rename semirat* to *semirat*. Add defvar for *semirat*.
rtoy [Mon, 6 Nov 2006 02:28:02 +0000 (6 02:28 +0000)]
src/defint.lisp:
o Add a few minor comments
src/residu.lisp:
o Document better what POLELIST and RES do.
riotorto [Sun, 5 Nov 2006 21:00:48 +0000 (5 21:00 +0000)]
Updating translations
robert_dodier [Sun, 5 Nov 2006 18:27:48 +0000 (5 18:27 +0000)]
Create stubs for Italian translation. These .texi files have the @menu,
@node, @section, @deffn, and @defvr tags but all of the text is cut out
by processing each English .texi file with the following awk script:
/^@menu/ {in_menu=1;}
/^@end menu/ {in_menu=0; print; print "";}
/^@node/ || /^@deffn/ || /^@defvr/ {print;}
/^@section/{print; print "";}
/^@end deffn/ || /^@end defvr/ {print "THIS ITEM NEEDS TO BE TRANSLATED."; print; print "";}
!/^@node/ && !/^@section/ && !/^@deffn/ && !/^@defvr/ && !/^@end deffn/ && !/^@end defvr/ {if (in_menu) print;}
The idea here is that all of the structure and none of the text is
preserved, so that will hopefully make it easier to create the translation.
``makeinfo maxima'' successfully creates maxima.info with these files.
robert_dodier [Sun, 5 Nov 2006 17:45:53 +0000 (5 17:45 +0000)]
Cut out uninformative section "Introduction to Bug Detection and Reporting".
ficmatinfmag [Sun, 5 Nov 2006 07:55:15 +0000 (5 07:55 +0000)]
augmented_lagragian.texi -> correcting *constraints* translation
robert_dodier [Sun, 5 Nov 2006 03:07:27 +0000 (5 03:07 +0000)]
Enclose @fnindex and @vrindex in @ifinfo / @end ifinfo,
because texi2html complains about @fnindex and @vrindex.
Too bad about that, really; the extra index items would be very helpful.
rtoy [Sun, 5 Nov 2006 02:52:42 +0000 (5 02:52 +0000)]
Fix for bug [
1582625 ] integrate(t^2*log(t)/((t^2-1)*(t^4+1)), t, 0, 1) wrong?
src/defint.lisp:
o Call $RECTFORM before simplifying plog(x) in LOG-IMAG-0-2%PI. This
is what fixes the bug. The issue is probably really in the
simplifier for plog, but that's much harder to fix, so workaround
the issue by calling $RECTFORM to make it easier for
LOG-IMAG-0-2%PI.
o Minor modification of KEYHOLE not to modify its argument N, and some
reindentation.
o Add some comments to RECTZTO%PI2.
tests/rtest15.mac:
o Add test for bug
1582625.
rtoy [Sun, 5 Nov 2006 02:27:01 +0000 (5 02:27 +0000)]
Fix for bug [
1504505] integrate( 1/(x^8-1),x,0,1/2) => internal error.
src/residu.lisp:
o Reindent $residue
o In RESM1 call $RATSIMP to simplify the pole. The underlying error
is in $TAYLOR, but calling $RATSIMP works around the issue.
tests/rtest15.mac:
o Add test for bug
1504505.
ficmatinfmag [Sat, 4 Nov 2006 21:07:26 +0000 (4 21:07 +0000)]
Enclosing vrindex and fnindex with ifinfo-endifinfo to keep alert messages during doc compilation clean. orthoply.texi -> minor bugs
ficmatinfmag [Sat, 4 Nov 2006 21:05:30 +0000 (4 21:05 +0000)]
orthopoly.texi minor bugs
ficmatinfmag [Sat, 4 Nov 2006 19:31:21 +0000 (4 19:31 +0000)]
Manual Portuguese Translation. Enclosing vrindex and fnindex with ifinfo-endifinfo to keep alert messages during doc compilation clean
ficmatinfmag [Sat, 4 Nov 2006 10:41:44 +0000 (4 10:41 +0000)]
Manual Portuguese Translation. -r 1.1 -r 1.2 augmented_lagrangian.texi,-r 1.41 -r 1.42 Command.texi,-r 1.11 -r 1.12 Constants.texi,-r 1.23 -r 1.24 Equations.texi,-r 1.18 -r 1.19 Floating.texi,-r 1.46 -r 1.47 Input.texi,-r 1.38 -r 1.39 Itensor.texi,-r 1.24 -r 1.25 Matrices.texi,-r 1.18 -r 1.19 Number.texi,-r 1.21 -r 1.22 Numerical.texi,-r 1.32 -r 1.33 Operators.texi,-r 1.28 -r 1.29 Plotting.texi
willisbl [Wed, 1 Nov 2006 23:22:27 +0000 (1 23:22 +0000)]
appending regression tests for meqp.
willisbl [Wed, 1 Nov 2006 23:20:02 +0000 (1 23:20 +0000)]
new version of meqp
willisbl [Wed, 1 Nov 2006 21:17:37 +0000 (1 21:17 +0000)]
Either we need to quote the correct output of assume, or not check it. I decided to not check it.
willisbl [Wed, 1 Nov 2006 21:03:31 +0000 (1 21:03 +0000)]
Fix for unreported bug featurep(x[1](2),integer) --> Lisp error. New code from Stavros
robert_dodier [Wed, 1 Nov 2006 14:54:52 +0000 (1 14:54 +0000)]
Removing coeflist.lisp and format.lisp from share/diff_form
because these two files are functionally identical to files of the same names
in share/contrib/format (only differences are in comments).
No other changes in share/diff_form are needed since load("coeflist.lisp") and
load("format.lisp") find the files successfully in share/contrib/format .
robert_dodier [Tue, 31 Oct 2006 05:37:21 +0000 (31 05:37 +0000)]
Append end-of-line character at end of file. (Lack of end-of-line character
might cause a problem for some version or versions of some Lisp implementations.)
robert_dodier [Mon, 30 Oct 2006 04:11:37 +0000 (30 04:11 +0000)]
Correct minor typo (b22 --> b2) in final input expression.
robert_dodier [Mon, 30 Oct 2006 00:00:25 +0000 (30 00:00 +0000)]
Change "for i from dim step -1 thru 0 do ..."
to "for i from dim step -1 thru 1 do ...".
With this change, d(translist) succeeds and the rest of the example
works just as shown in surface_example.txt.
robert_dodier [Sun, 29 Oct 2006 18:00:47 +0000 (29 18:00 +0000)]
Remove load(..) and infix(...) from this example script.
robert_dodier [Sun, 29 Oct 2006 17:09:38 +0000 (29 17:09 +0000)]
Commit files submitted by Gosei Furuya in message dated Oct 29, 2006 3:16 AM
with subject line "Re: [Maxima] differencial form package".
All files are the same as submitted, except that two have been renamed:
rename cartan_init.bat --> diff_form.mac (make initialization file name same
as the name of the package, and make it a .mac file), and
rename new_cartan_test4.bat --> new_cartan_test4.mac (make it a .mac file).
robert_dodier [Sat, 28 Oct 2006 20:15:51 +0000 (28 20:15 +0000)]
Added example @deffn and @defvr definitions.
robert_dodier [Thu, 26 Oct 2006 14:44:26 +0000 (26 14:44 +0000)]
In texput, handle user-defined nary and nofix operators.
Minor adjustments elsewhere accordingly.
-- tests/rtest.mac: More examples of user-defined TeX properties
-- src/mactex.lisp:
(1) TEX-NARY: treat TeX symbol as a list (for consistency with other
TeX output functions)
(2) TEX-NARY: call APPEND instead of NCONC (avoid in-place modification)
(3) TEX-NOFIX: grab TeX symbol (instead of printing symbol verbatim)
(4) MNCTIMES, MTIMES, MAND, MOR: TeX symbol is a list (see (1))
(5) MOR: Cut redundant (DEFPROP MOR TEX-NARY TEX)
(6) $TEXPUT: Convert input TeX symbols to Lisp strings
(7) $TEXPUT: Added cases for $NOFIX and $NARY
(8) $TEXPUT: Return TeX symbols
robert_dodier [Mon, 23 Oct 2006 01:32:42 +0000 (23 01:32 +0000)]
Fixed minor typo in example (x range specified twice instead of x range and y range).
robert_dodier [Sun, 22 Oct 2006 16:09:36 +0000 (22 16:09 +0000)]
Put in @fnindex for operators and @vrindex for some symbols,
to make it possible to find the operator or symbol by entering its
name for describe or ?.
robert_dodier [Fri, 20 Oct 2006 05:41:25 +0000 (20 05:41 +0000)]
In MDEFARRAY, ensure that loop variable U is an atom before
attempting to test its first character.
Without this change, expressions like f[x + y](z) := foo cause a Lisp
error "STRING: argument ((MPLUS) $X $Y) should be a string ...".
Now f[x + y](z) := foo does not cause an error (x + y is treated
as a hash key, which I believe is just as it should be).
robert_dodier [Thu, 19 Oct 2006 03:51:03 +0000 (19 03:51 +0000)]
Revise and expand description of realroots.
billingd [Tue, 17 Oct 2006 14:05:10 +0000 (17 14:05 +0000)]
Update solution for Kamke equation 1.558. Order of terms has changed.
billingd [Tue, 17 Oct 2006 14:02:03 +0000 (17 14:02 +0000)]
Update result for Kamke equation 1.400
van_nek [Tue, 17 Oct 2006 13:07:30 +0000 (17 13:07 +0000)]
update and restriction to already translated chapters
van_nek [Tue, 17 Oct 2006 13:06:36 +0000 (17 13:06 +0000)]
changing the appearance of umlauts and es-zets
van_nek [Tue, 17 Oct 2006 13:04:16 +0000 (17 13:04 +0000)]
just for completition
van_nek [Tue, 17 Oct 2006 13:03:33 +0000 (17 13:03 +0000)]
German translation of Help.texi CVS 1.19
robert_dodier [Sat, 14 Oct 2006 16:25:15 +0000 (14 16:25 +0000)]
Documentation for the newton function (in share/numerical/newton1.mac).
At one time there existed documentation for newton, but it was inaccurate and confusing
and so it was cut out. The new documentation attempts to be clearer.
robert_dodier [Sat, 14 Oct 2006 16:18:37 +0000 (14 16:18 +0000)]
Added notes about how to dump Maxima image for SBCL and Allegro.
robert_dodier [Fri, 13 Oct 2006 05:47:05 +0000 (13 05:47 +0000)]
Added some text to describe recently invented plot option plot_realpart.
robert_dodier [Thu, 12 Oct 2006 15:18:45 +0000 (12 15:18 +0000)]
(1) Work around GCL number of arguments limitation in $matrix.
src/mlisp.lisp: change $matrix from defun to defmspec, and evaluate arguments within $matrix.
From point of view of caller, $matrix should work same as before (but without bug).
tests/rtest2.mac: new test cases for matrix function.
(2) Avoid unintended upper/lower case inversion in declare(foo, alphabetic) (in src/mlisp.lisp).
Also process all characters in a string instead of just one at a time.
robert_dodier [Thu, 12 Oct 2006 14:46:33 +0000 (12 14:46 +0000)]
Added specfun.lisp (place-holder which prints a message about how specfun went away).
specfun.lisp should have been added in the 5.10.0 release; an oversight.
riotorto [Thu, 12 Oct 2006 11:20:18 +0000 (12 11:20 +0000)]
Removing tab characters
willisbl [Wed, 11 Oct 2006 23:53:13 +0000 (11 23:53 +0000)]
fixing xreduce(f,[a,b,c]) bug and xreduce(union, set(set(a),set(b))) bug.
(Bug report and fixes due to Stavros).
rtoy [Tue, 10 Oct 2006 21:13:17 +0000 (10 21:13 +0000)]
Bug
1571454: plot2d(0.99,[x,0,5]) is very slow.
Modify slow-enough-p as mentioned in the bug report. This may need
some tweaking.
billingd [Tue, 10 Oct 2006 13:29:54 +0000 (10 13:29 +0000)]
Comment out three tests. These now prompt for input and hang the testsuite.
Tests are clearly marked with FIXME and should be reviewed at some stage.
willisbl [Tue, 10 Oct 2006 11:09:23 +0000 (10 11:09 +0000)]
Fixing bug in noncommutingring
willisbl [Mon, 9 Oct 2006 21:09:25 +0000 (9 21:09 +0000)]
Changing div back to mncexpt, but keeping the conditional check for
matrixp. The div is wrong for symbolic lu facotorization; and mncexpt
is wrong for double-blocked matrices.
andrejv [Mon, 9 Oct 2006 08:55:37 +0000 (9 08:55 +0000)]
Fixed a problem with subscripted values in change_to_shift (reported on
the list by Laurent Couraud).
andrejv [Mon, 9 Oct 2006 08:41:56 +0000 (9 08:41 +0000)]
Don't print warnings about the rank of A.
willisbl [Mon, 9 Oct 2006 01:03:22 +0000 (9 01:03 +0000)]
Fix for bug
1573247; also append tests for double-blocked matrices in
test-lu.mac and test-cholesky.mac
riotorto [Sat, 7 Oct 2006 09:05:13 +0000 (7 09:05 +0000)]
Rewriting translation of function 'express'
riotorto [Sat, 7 Oct 2006 07:47:10 +0000 (7 07:47 +0000)]
Removed paragraph in English
van_nek [Wed, 4 Oct 2006 11:06:33 +0000 (4 11:06 +0000)]
German translation of Command.texi CVS 1.41
van_nek [Wed, 4 Oct 2006 11:05:22 +0000 (4 11:05 +0000)]
German translation of Floating.texi CVS 1.19
van_nek [Wed, 4 Oct 2006 11:04:16 +0000 (4 11:04 +0000)]
German translation of Bugs.texi CVS 1.9
van_nek [Wed, 4 Oct 2006 11:01:44 +0000 (4 11:01 +0000)]
updating
van_nek [Wed, 4 Oct 2006 11:00:41 +0000 (4 11:00 +0000)]
minor changes
van_nek [Wed, 4 Oct 2006 10:51:36 +0000 (4 10:51 +0000)]
minor changes