This is fixed for the next beta.
Henry Rich
On 3/28/2021 2:22 PM, 'Michael Day' via Beta wrote:
Thanks, though I don't recall "PPPP;" Wiki reveals some unlikely
matches, including
Pigeons Playing Ping Pong.
I agree that the timing numbers are "correct." I don't know the same
about the numbers
replacing parts of the code. The display is still quite useful, but
it is of course preferable
to see what one has typed.
I've just checked back in J902 - the display is ok for F[14] and I
didn't manage to get jpm
to misbehave otherwise as it had done for me recently! I'll try to
remember to report when
and if I encounter the problem I have in mind.
Happy hunting!
Mike
On 28/03/2021 17:44, Henry Rich wrote:
Thanks for taking up beta-g so promptly!
I believe you have hit a problem related to PPPP. Starting with
beta-g the sequence (2*+/@:e.~) is detected at function-definition
time and parsed then, saving time during execution of the verb. I
have no reason to suspect an error in that, and I think the numbers
in the result are correct.
The error is in the display generated by showdetail. Before PPPP,
the only things that could possibly be in an explicit definition were
primitives and names. Now, they include anonymous compounds, and
something doesn't know how to format them. I thought I had found
where the explicit definition was displayed but there must be another
place.
I'll have to look into showdetail and see where it picks up its view
of the executing definition. If anybody knows that & can tell me,
I'd be obliged.
Henry Rich
On 3/28/2021 12:25 PM, 'Michael Day' via Beta wrote:
I haven't noticed this before using jpm, using beta g here.
The showdetail_jpm_ display is now sometimes showing apparent
constants replacing parts of an expression. I doubt the constants
are constant, but haven't checked. I also doubt whether it matters
much what the function is trying to do.
It is however satisfying that jpm works at all; I've had trouble
avoiding
errors preventing showdetail recently, not reported because
difficult to
pin down, explain what I'm doing.
JVERSION
Engine: j903/j64avx2/windows
Beta-g: commercial/2021-03-27T13:01:30
Library: 9.03.01
Qt IDE: 1.9.2/5.12.10(5.12.10)
Platform: Win 64
Installer: J903 install
InstallPath: c:/d/j903-beta
Contact: www.jsoftware.com
start_jpm_ 100000000
1785713
timer'F p:300'
+---------+-------+
|0.0170059|1326672|
+---------+-------+
NB. look at line [14] in this display cf the Function listing
further below.
NB. I've tried setting fixed width below, without apparent success!
showdetail_jpm_ 'F'
Time (seconds)
+--------+--------+----+--------------------------------------+
|all |here |rep |F |
+--------+--------+----+--------------------------------------+
|0.000001|0.000001| 1|monad |
|0.000001|0.000001| 1|[0] p=.y |
|0.000006|0.000001| 1|[1] by=.p mtimes |
|0.000004|0.000001| 1|[2] ad=.p madd |
|0.000001|0.000001| 1|[3] echo |
|0.000024|0.000017| 1|[4] all3=.}.a3=.a by*:a=.i.p |
|0.000010|0.000010| 1|[5] if. p=#~.a3 do. (-~2*:)<:p return.|
|0.000000|0.000000| 0|[6] else. |
|0.000000|0.000000| 1|[8] end. |
|0.000000|0.000000| 1|[9] echo |
|0.000015|0.000009| 1|[10] t=.+/}.a3 e.~ad~a3 |
|0.000104|0.000104| 1|[11] for_i. i.<:p do. |
|0.001148|0.001148|1992|[12] a3=.}.a3 |
|0.005897|0.001223|1992|[13] try=.({.ad}.)a3 |
|0.005955|0.005955|1992|[14] t=.t+ti=.all3(23*)try |
|0.000090|0.000090|1992|[16] end. |
|0.000003|0.000003| 1|[17] t |
|0.013259|0.008565| 1|total monad |
+--------+--------+----+--------------------------------------+
F
3 : 0
p =. y
by=. p mtimes
ad=. p madd
echo
all3 =. }. a3=. a by *: a =. i.p
if. p = #~.a3 do. (-~*:) <: p return.
else.
end.
echo
t =. +/ }. a3 e.~ ad~ a3
for_i. i.<:p do.
a3 =. }. a3
try =. ({.ad}.)a3
t =. t + ti =. all3 (2*+/@:e.~) try NB. where does
"23" come from
end.
t
)
Cheers,
Mike
--
This email has been checked for viruses by AVG.
https://www.avg.com
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
