LiveCode Forums

Whatever else you changed in the javascript, what I can see here is correct and it doesn't use special new features of javascript. You could check again: 1. At what point do you set the javascriptHandlers and the htmltext of the widget? This should be done on openstack. Mobile is slow and needs some...

Your code do "var g=" & sudoku4JS & ";var r=solve_sudoku(g);liveCode.JSresult(g," & T0 & ");" in widget "Browser" fails, if T0 is undefined. Just write liveCode.JSresult(g," & 0 & "); instead or enclose T0 with single quotes liveCode.JSresult(g,'" & T0 & "'); A possible problem with mobile is genera...

This has nothing to do with javascript: Alerts on mobile are different from alerts on desktop. To have a unique code for the platforms you could use (instead of "answer ..."): put tLevel & "-Puzzle solved." into tA do "alert('" &tA& "');" in widget "browser" This javascript dialog works also with a ...

Here are some more sudokus for testing, see https://attractivechaos.github.io/plb/kudoku.html . "Hardest 20" 000000000000003085001020000000507000004000100090000000500000073002010000000040009 000000012000000003002300400001800005060070800000009000008500000900040500470006000 020050700400100006800003000...

Thierry, this is a really fine optimization you have worked out for that. On the other hand, the dancing links algorithm X (DLX) by D. Knuth is known to be very fast for solving even difficult puzzles. Find enclosed a stack that does a JS implementation of that algorithm in JavaScript (via a browser...

Related to images LCB is at least as fast as LCS when drawing rectangular blocks of images and is faster with deliberate affine transforms of images. Moreover LCB has features that you don't have in LCS, e.g. (sub-pixel) drawing of affine transformed pathes, especially SVG Pathes. LCB is still far a...

With accessing single bytes of imageData it is even slower than LC Script...

LiveCode's byte-version is still 3-5 times slower than the (for that purpose non-optimized) JS version what is nevertheless a "bump" compared to the char-version. Here is the script of a button (pure LCScript) you can use in my imageStatistics-stack for comparison (yields the same output using 256 b...

The stacks were (nearly) all there, some since years. The problems started after the 'server change' last year. The first limit was for up-/download stacks of size > 2-3 MByte. I reported to support, Robin removed that limit. The limit changed again to (currently) stacks of size > 5-6 MByte. Reporte...

"Sample Stacks"/livecodeshare/revOnline doesn't work for me any more (impossible to upload/download large stacks) since several months. Thus I removed all my 64 stacks from there.
So: Sorry, but given time, in a few weeks, I'll make an own download page with some of these.

These are the results of tests with severals of my use cases: If you use textstyle only then certainly the array method is faster than using the htmltext (or the ordinary styledText). But if you do also other text styling that textstyle can't do (and have already the htmltext/styledText in a variabl...

Moreover you could try to use "call" instead of "send" (what changes the context).

Now it is clear (for me) what you really want.
The following could be done also with other tags as <b> or <i> instead of <box>.

-- ON set the textstyle of char x to y of fld "name" to "box" -- OFF set the textstyle of char x to y of fld "name" to empty To preserve other textstyles when removing, you could also put the effective textstyle of <chunk> into <variable>. Then filter items of <variable> without "box" and then set ...

Here is for your comparisons a complete "brute-force-solver" for 9x9 sudokus (=the method you tried to code from python) in 65 lines, the essential solver in 25 lines. It is implemented into LC based on https://samirhodzic.github.io/sudoku-solver-js/ (which is more than 100 times faster than Livecod...