Henry-- The transcripted excerpts below show the latest instance of my problems with lost values.
Happy hunting (not haunting). --Art > On 1Nov 2020, at 4:46 PM, Henry Rich <[email protected]> wrote: > If you give me a script that reproduces the problem, I will fix it. The less > you give, the less my chances of fixing it. With a report as sketchy as the > one you provided I can do nothing. ------------------- Debugging J902beta-j Program steps: fb2=. (<"0 fw) {."0 1 each"1 fb1 NB. Setting common internal widths in each column fb3=. < CTicl & CTTrncAry eachfb2 NB. Trimming items if needed y2=. ,"2 each/"1 ,"1 each/"1 fb2 NB. Merging to level-1 rank-m character table(s) After stop: $each fb2 ┌────┬────┬────┬────┐ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │5 17│5 22│5 22│5 22│ ├────┼────┼────┼────┤ │6 17│6 22│6 22│6 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ └────┴────┴────┴────┘ ┌────┬────┬────┬────┐ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ └────┴────┴────┴────┘ ┌────┬────┬────┬────┐ │6 17│6 22│6 22│6 22│ ├────┼────┼────┼────┤ │6 17│6 22│6 22│6 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ └────┴────┴────┴────┘ ┌────┬────┬────┬────┐ │6 17│6 22│6 22│6 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │6 17│6 22│6 22│6 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ └────┴────┴────┴────┘ ┌────┬────┬────┬────┐ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │6 17│6 22│6 22│6 22│ ├────┼────┼────┼────┤ │6 17│6 22│6 22│6 22│ ├────┼────┼────┼────┤ │6 17│6 22│6 22│6 22│ └────┴────┴────┴────┘ ┌────┬────┬────┬────┐ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ ├────┼────┼────┼────┤ │7 17│7 22│7 22│7 22│ └────┴────┴────┴────┘ $each fb3 $&.> fb3 $each y2 NB. A problem value, from earlier runs, but should be local ┌───┬───┬───┐ │3 4│3 4│3 4│ ├───┼───┼───┤ │3 4│3 4│3 4│ └───┴───┴───┘ $each fb3 $&.> fb3 fb3 < 80 40&CTTrncAry eachfb2 ] fb3=. < CTicl & CTTrncAry eachfb2 ] (< 80 40&CTTrncAry eachfb2) NB. Note error report malformation $ fb2 NB. fb2 has shape, but apparently no good value 2 3 4 4 ------ Routine called dozens of times before this point (probably not the culprit): CTTrncAry=: 4 : 0 NB. Truncates shape of one rank-n level-m array--contained in y--to fit NB. within shape x, marking any loss of data NB. if. CTStops do. 'Stop' 13!:8 [ 2 end. yval=. > y NB. Beware mixed types! xrev=. (|. $yval) <. ($$yval) {.!._ |. x NB. Adjusts (reversed) shape not to over-rank or over-fill NB. xrev=. xrev <. (# xrev) {.!.1 (CTRankLim # _) NB. Flattens excess axes (without notice) if. +./ cuts=. xrev < |. $ yval do. NB. Axes exceeding bounds marks=. > ((((<0),(<' '),(<a:)) i. ([: <@{. [: >@(0&{.@,) ])) { CTTrncMks&[)yval for_axis. 1 + i. $$yval do. if. (axis-1){ cuts do. yval=. (((axis-1) { xrev) {."axis yval) ,"axis (axis > 1) { marks end. end. else. < yval NB. Boxed return. end. ) JVERSION Engine: j902/j64avx2/darwin Beta-j: commercial/2020-10-29T16:26:56 Library: 9.02.06 Platform: Darwin 64 Installer: J902 install InstallPath: /applications/j902 Contact: www.jsoftware.com ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
