Moreover, integral values in range 1+2⋆31 to (2⋆32)-1
are incorrectly interpreted as their two's complement
negative equivalents:
0⎕SVO'X'
1
X←1+2⋆31
X
¯2147483647
-k
Hi again,
There are still (r148) some problems similar to previously reported.
When passing a big array as shared variable, instead of being
truncated, now its tail is corrupted:
0⎕SVO'X'
1
X←489⍴2
¯2↑X
2 ¯1254096894
This suggests some nasty out-of-bounds read.
And, if array is
Hi,
providing descriptions of functions is a lot of work. Most people know
APL after
a short while, and those who need it are probably better off with a good
APL book.
I wanted to wait with the next release until the rate of error reports
goes down a little.
Right now it seems to get a littl
Thanks. Looks easy enough.
What do you think of the idea of having descriptions of the functions
embedded in the APL interpreter instead of in the Emacs mode?
Secondly, when do you plan to release 1.3?
Regards,
Elias
On 1 March 2014 23:50, Juergen Sauermann wrote:
> Hi Elias,
>
> for commands
Hi Elias,
for commands, you can #define macro cmd_def() and then #include Command.def.
For Quad-vars you can #define macros ro_sv_def() and rw_sv_def and then
#include SystemVariable.def.
For example, to get the strings for all commands:
#define cmd_def(cmd_str, code, _arg) #cmd_str
#include
I've recently added smart expansion of symbols in the Emacs mode (press TAB
and get a list of suggested variable and function names). I'd like it to be
able to expand system commands and quad commands as well. Of course, I
could simply hard-code the list of available commands, but it would be
neate
