It was said:
"embarassing bugs have this way of getting fixed faster..."
The last time I reported this bug, it was a bug in Maple 10, and I reported it in person to one of the GUI workers. If I wanted to report this bug privately again, I would use email@example.com; however, this forum seems just as effective a means of conveyance.
Anyway, there seems to be a bug, when using Maple's Classic interface, in the mechanism that redraws the screen. This is how the bug can be consistently replicated:
On a new worksheet, type in and execute a command which will at least fill the entire screen with output. For example: 10000!; fills my screen and does not trigger elision.
I've been laying low for a while now, finding new ways to break and repair Maple. With Maple11, I've found at least two disturbing bugs.
The first is in implicitplot(): Maple no longer allows for variable ranges that evaluate to numbers. In Maple10, the command
plots[implicitplot](x^3=y-x, x=-5..5, y=0..x);
plots correctly, whereas the same command in Maple11 results in the error message:
"Error, (in plots/implicitplot) invalid input: invalid range for second variable"
The second bug is yet another bug in Maple's eternally bug-ridden Worksheet package. I do not know if this bug exists in previous versions of Maple, but there is a problem reading in nested quotes. To see this bug, do the following:
We've encountered an interesting problem, outlined below:
f:= x->x; #anything, just so long as f is a function.
b:= leftsum( f, x=1..2, 6); # the range and number of rectangles aren't really important.
As you will see, b does not throw an error, and yet c does. The recursion error seems to be an untrappable error (i.e. it kicks you immediately out of all try-catch loops and back to the top level). It immediately breaks out of any function or procedure we throw at it. After analysis with printlevel := 100, it seems that the error arises in how evalf calculates the Sum in evalf/Sum1; evalf uses the variable r and adds each piece of the sum to r one by one.
According to the helpfile, the syntax for using the command ReplaceChild in the XMLTools package is
where spec is an equation which is acceptable to subsop(). However, when I try to do this:
XMLTools:-ReplaceChild( [1,1]="foo", doc);
where doc is an XMLtree I have previously defined, I receive this error message:
Error, (in ReplaceChild) child position must be specified as a positive integer, but received [1, 1].
Sounds like yet another oversight in XMLTools.
I'm in the process of manipulating Maple worksheets (both Classic .mws and Standard .mw formats) blindly from the command line. There are a surprising number of stumbling blocks in Maple that make this work difficult. While I've managed to worm my way around them one by one, I thought I'd take some time to just voice my frustration so that others are aware of these pitfalls.
I should note that I am using Maple 10.02 on a Gentoo Linux machine, feeding my commands into Maple through a standard pipe. I would update to Maple 10.04, but the updater program from the website continues to cause problems with Java. Another complaint? That might as well be...