On 2014-07-15 20:09:40, Juergen Sauermann wrote:
> Except that (not so obviously) :
>
>
> 1E¯15 = 1E¯14
> 0
Sorry, my mistake. You're, of course, right.
Current (r376) behaviour seems to be correct with one exception:
1E¯15 = 1E¯14
0
1E¯15J0 = 1E¯14J0
1
Both should be zero,
Hi Kacper.
thanks, fixed in SVN 376.
/// Jürgen
On 07/15/2014 12:18 PM, Kacper Gutowski wrote:
2⎕TF re-uses real part for imaginary part:
X←5J0
10⎕CR'X' ◊ 2⎕TF'X'
X←5J0
X←5J5
-k
Hi Kacper,
thanks , fixed in SVN 376.
Except that (not so obviously) :
* 1E¯15 = 1E¯14
0
*
**/// Jürgen
On 07/15/2014 11:55 AM, Kacper Gutowski wrote:
When either of arguments of = has imaginary part of magnitude greater
than comparison tolerance, a domain error is thrown.
⎕CT
2⎕TF re-uses real part for imaginary part:
X←5J0
10⎕CR'X' ◊ 2⎕TF'X'
X←5J0
X←5J5
-k
When either of arguments of = has imaginary part of magnitude greater
than comparison tolerance, a domain error is thrown.
⎕CT
1E¯13
0J1 = 0J1
DOMAIN ERROR
0J1=0J1
^ ^
There are no conditions under which = is allowed to throw a domain
error at all so this is serious.
W
The latest version of the SQL library has renamed the ResultValue files:
https://github.com/lokedhs/apl-sqlite
Regards,
Elias
On 15 July 2014 16:30, Jay Foad wrote:
> The attached patch lets me build without libsqlite3-dev again, by
> moving both ResultValue.* and apl-sqlite.* into the "if SQ
The attached patch lets me build without libsqlite3-dev again, by
moving both ResultValue.* and apl-sqlite.* into the "if SQLITE3" part
of the makefile.
Jay.
On 15 July 2014 03:12, Elias Mårtenson wrote:
> It's not supposed to. I think the error is that it still tries to compile
> ResultValue.cc
