Posts: 96
Threads: 20
Joined: Sep 2011
Hi Marcus
I have found an error (finger trouble probably) in the matrix library - namely in wp34s-1.dat.
The first 'M-' label should read 'M-0'. The real 'M-' appears further down.
For Pauli - in the 'M-y' routine you execute a [cmplx]DROP, which does not leave
a descriptor on the stack. Should this not be just a DROP?
Regards John
BTW- Thanks for these routines. The LU routine (if included) would be great :)
Edited: 19 Oct 2011, 4:25 a.m.
Posts: 3,229
Threads: 42
Joined: Jul 2006
Quote:
I have found an error (finger trouble probably) in the matrix library - namely in wp34s-1.dat.
Won't be finger trouble -- these are generated by the assembler from the sources in the library. Still, the source only has one 'M-' label so something has gone awry.
Quote:
For Pauli - in the 'M-y' routine you execute a [cmplx]DROP, which does not leave a descriptor on the stack. Should this not be just a DROP?
The matrix descriptor is saved in L. The value on the stack isn't the descriptor anymore -- it is a loop index for ISG that has finished. Thus, complex DROP is correct here.
Quote:
BTW- Thanks for these routines. The LU routine (if included) would be great :)
The LU routine will be back if we can find enough space for the wrapper. We're working on it.
- Pauli
Posts: 96
Threads: 20
Joined: Sep 2011
Thanks Pauli, should have checked L - I just looked at the comments which said the descriptor was returned in X.
Thanks for the prompt response. I am really enjoying this calculator :)
Regards John
Posts: 3,229
Threads: 42
Joined: Jul 2006
There wasn't a comment on this routine about stack and register use????
I've added some clarification nonetheless. The documentation for these is in the doc directory.
- Pauli
Posts: 96
Threads: 20
Joined: Sep 2011
My apologies then, I mistook the comments above, applying to the functions below. Still love this wp34s - great effort and thanks!
Regards John
Posts: 3,283
Threads: 104
Joined: Jul 2005
Quote:
Still, the source only has one 'M-' label so something has gone awry.
This must have been my fault when changing names from M. to M- in the (wrong) hope that would ease the navigation of the built-in matrix commands in the catalogs. I assume it has been fixed by Pauli now.
Posts: 4,587
Threads: 105
Joined: Jul 2005
... and at least some of these routines/commands will be renamed soon.