mattgodbolt / owlet-editor Goto Github PK

• type &1
• type A and note the suggestions...

Shouldn't be any at that point.

A significant feature the old editor had which owlet seems to currently lack is timing information. The bot waits 30 seconds then records for 3 seconds, and in many cases it's really useful to see what happens in those 3 seconds before sending the code to the bot.

Right now to test timing I use a local install of the bot code so I can run node client try and then copy and paste the code from owlet to a terminal, which is all a bit fiddly (especially because you seem to explicitly need to copy and I'm used to X-style select-to-copy). I doubt many users will bother with this.

The old editor's timing didn't entirely match what happened in the bot, but was good enough for many cases.

' in a PRINT emits a CR+LF, but the editor colours it red as if it's an error:

P.'"a"'@%''TIME'

Expanding out the whole thing from the (sometimes) compressed mass of P."moo";:?A=1:MO.2 into separate lines. Would need to handle line numbering cleverly.

It would be helpful to people less familiar with BBC BASIC if invalid operators were highlighted in red like many other invalid cases are - especially for cases which are valid in other BASIC dialects and/or other languages. E.g.:

A=1:B=1:C=1:ONERRORREPORT:P." at line ";ERL;:G.ERL+10
A+=B
A-=B
A/=B
A*=B
A^=B
A=B<<C
A=B>>C
IFA==B:P.
IFA!=B:P.
IFA=>B:P.
IFA=<B:P.
IFA><B:P.
END

Some of these even work in later BBC BASIC versions.

(The "backwards" =>, etc are supported by e.g. Apple II BASIC)

MODE2
FORI=0TO2

Having a new button (other than ctrl+a+backspace) would be good - on mobile deleting the whole program there is a pain

&7c00 = error, &7C00 = ok

Follow-on from #27:

The part after the E can just have a + or - and no digits:

A=9E+
B=9E-
C=.E-+.E+
P.A,B,C

Tokens with left and right versions (e.g. HIMEM) seem to all give duplicate entries in the completion list.

@8bitkick points out the editor is feeling a little sluggish. A super quick profile shows we're spending a bunch of time in the tokeniser, unsurprisingly probably due to the giant regex we use to tokenise. Most "sensible" languages use whitepsace to terminate tokens, and monaco is optimised for that. We have to disable it though as PRINTLEN is two tokens without whitespace :)

Adrian on slack points out that folks might just type print "hello" and wonder what's going on.

e.g. https://bbcmic.ro/#%7B%22v%22%3A1%2C%22program%22%3A%22MODE4%3AFORL%3D1TO8%3AREADS%24%3AMOVEFNG(1)%2CFNG(2)-100%3AFORI%3D1TOLEN(S%24)%2F2-1%3ADRAWFNG(I*2%2B1)%2CFNG(I*2%2B2)-100%3ANEXTI%3ANEXTL%5CnDEFFNG(A)%3A%3D14*ASC(MID%24(S%24%2CA))-400%5CnVDU31%2C12%2C4%3APRINT%5C%22----%20E%20L%20I%20T%20E%20----%5C%22%5CnDATA5QJKTG_%3FW%3FJD%3CK5Q%2C5Q8UIXRT%5DE_%3F%2CJKIX%2CTGRT%2CHJMIJFFHHJ%2COGSEQDMEOG%2CMVMX%2C(hqhq1(1(h%22%7D

Consider:

P.1.,.5,.

Here .5 is highlighted as a number (good), but in 1. the . is red, and . by itself is also red (but this is really the same as 0.0 as you can see if you run this program). I think 1. and . should both be highlighted as numbers.

TOP is a bit of an odd-ball keyword in that it doesn't actually have its own token but is actually tokenised as a TO token followed by a letter P.

The syntax highlighting doesn't seem to know about this quirk, e.g. this actually colours the P brown instead of blue:

PRINT TOP

You can't just colour any P after TO blue though, consider:

F.P=TOP TOP:P.P:N.

There the first TOP is TOP, but the second TOP is actually TO P.

We should consider the C bit of the token table for e.g. FALSE, EXT... http://www.benryves.com/bin/bbcbasic/manual/Appendix_Tokeniser.htm

Concretly, TIMER=1 is parsed as a variable, not <TIME>R=1

See #34

Currently Z%=2E4 highlights 2 as a number but E4 as a variable. I think 2E4 should be highlighted as a number.

cosmetic enhancement: when you shrink code, you see a mess of accented characters and blanks in the editor. A bunch of the byte tokens for keywords fall into the C1 control character block, so just show up as spaces. It makes it hard to tell what's going on (and I know, I'm reading the compressed code - I shouldn't really care)

Making the spans that contain the tokens have title attributes with the expanded keyword would be ideal, since then mouseover would reveal what is going on. However, this doesn't seem possible with monaco. Another solution does work tho:

@font
-face {
 font-family: 'Unprintable';
 src: local('Unicode BMP Fallback SIL');
 unicode-range: U+80-9F,U+A1-FF;
} 

declare a fallback font, 'Unprintable', for the high-byte characters. SIL BMP displays hexadecimal instead of the normal glyphs https://scripts.sil.org/cms/scripts/page.php?site_id=nrsi&id=unicodebmpfallbackfont
... the full font covers all of the basic multilingual plane, subsetting it to just the ones used would be ideal. Here I'm using a local font but the font would have to be imported for most users. A0 is excluded, since I see the editor uses non-breaking spaces to display normal spaces in the code.
Then just use Unprintable as the first font, so it replaces that character range:

.monaco-editor {
    font-family: Unprintable,-apple-system,BlinkMacSystemFont,Segoe WPC,Segoe UI,HelveticaNeue-Light,system-ui,Ubuntu,Droid Sans,sans-serif;
    --monaco-monospace-font: Unprintable,"SF Mono",Monaco,Menlo,Consolas,"Ubuntu Mono","Liberation Mono","DejaVu Sans Mono","Courier New",monospace;
}

While playing around and running the code it seems like the "undo" doesn't go back as far as I like. Sometimes only one step. Not sure what's causing this, maybe something is setValue()-ing the code unnecessarily?

See #36 - abbreviated tokens are "MISTAKE"s, but putting a space fixes it. Put wigglies under it instead of just making it red, and then have an action to fix it.

Following on from #20...

The current state is better than before the fix for #20, but ~ now seems to be treated as an operator like +, etc but really it's more like ' as it can't be used in most expressions. I think it only works in PRINT and as the next token after STR$ (so A$=STR$~(12) and B$=STR$~12 work, but C$=STR$(~12) doesn't.)

Admittedly a minor issue, but ideally A=~B should show ~ in red.

e.g. in the line:

220 IF INKEY(-83) colour%=colour%+1:IF colour%>7 colour%=1

Currently the highlighter seems to switch back to BASIC if it sees a : in a * command:

*FX19:PRINT

 MODE 2:!-512=720907:$@%="1022":!80=&2030605: DIM S(18),C(18): FOR L=0 TO 18:A= RAD(21*L-9):S(L)= SINA*23:C(L)= COSA*28: NEXT : PRINT "���PRESS CONFERENCE";
 FOR C= FNb(640,-240,9,9) TO 4:C=3 AND C: VDU 31,6- INKEY75;: PRINT  MID$("AUTUMNWINTERSPRINGSUMMER",C*6+1,6): VDU 19,1,@%?C;0;19,2,C?80;0;: NEXT 
 DEF  FNb(X,Y,A,D) IF D MOVE X,Y:U=S(A)*D:V=C(A)*D: VDU 18;1+D/2,281;-U;-V;18;D/2+.5: PLOT 49,U,V:X=X-U:Y=Y-V:= FNb(X,Y,A+ FNb(X,Y,A-1,D-1),D-1)
=1

Lots of issues show up here (wordwrapping; for which "bounded" doesn't work properly even), comments.

In the ‘examples’ tab on the right, the text doesn’t wrap. Presumably because it’s better to maintain a separate ‘this is the first line, this is the second line’ concept… however it did make it a tad painful to select the text when I wanted to copy-paste it into the left-hand editor pane.

e.g. "unable to tokenise" if you paste a long line in.

Add support to run in overclocked mode. In jsbeeb as-is it's limited to probably 2.5-3x real speed.

It appears as a big empty box instead of a nice list of commands and editor options

^ should get highlighted as an operator, but it's currently red:

F.A=H. TO8^5:?A=42:N.

Some keywords don't work in a program so it's not helpful to suggest them as completions.

It seems the list is AUTO, LIST, RENUMBER, NEW, OLD, LOAD, SAVE but I may have missed some (or contexts where some of those work). (Note CHAIN and RUN are valid in a program.)

https://bbcmic.ro/#%7B%22v%22%3A1%2C%22program%22%3A%22DIM%20MC%25%20400%3AFOR%20O%25%3D0%20TO%203%20STEP%203%3AP%25%3DMC%25%3A%5B%3AOPT%20O%25%3A.H%25%20EQUW%20%2655AA%3A.S%25%20EQUW%20%263000%3A.Y%25%20EQUD%20%26FCF0C000%3AEQUD%20%26030F3FFF%3A.g%20LDA%20S%25%3ASTA%20%2670%3ALDA%20S%25%2B1%3ASTA%20%2671%3ALDY%20%230%3A.t%20LDA%20H%25%3ASTA%20(%2670)%2CY%3AINY%3ALDA%20H%25%2B1%3ASTA%20(%2670)%2CY%3AINY%3ABNE%20t%3ALDY%20%230%3AINC%20%2671%3ABPL%20t%3A.c%20LDA%20S%25%3ACLC%5CnADC%20%2332%3ASTA%20%2676%3ALDA%20S%25%2B1%3ASTA%20%2677%3ALDA%20%230%3ASTA%20%2672%3A.a%20ASL%20A%3AAND%20%237%3ASTA%20%2673%3ALDA%20%2676%3ASTA%20%2678%3ALDA%20%2677%3ASTA%20%2679%3ALDA%20%230%3ASTA%20%2674%3A.l%20LDA%20%2678%3ASTA%20%267A%3ALDA%20%2679%3ASTA%20%267B%3ALDA%20%2673%3ACLC%3AADC%20%267C%3AAND%20%23%267%3ATAY%3ALDA%20Y%25%2CY%3ALDY%20%230%3ASTA%20(%267A)%2CY%3AINY%3ASTA%20(%267A)%2CY%3AINY%3ASTA%20(%267A)%2CY%3AINY%3ASTA%20(%267A)%2CY%5CnINY%3ASTA%20(%267A)%2CY%3AINY%3ASTA%20(%267A)%2CY%3AINY%3ASTA%20(%267A)%2CY%3AINY%3ASTA%20(%267A)%2CY%3ALDA%20%23%26F0%3ABIT%20%2674%3ABNE%20e%3ALDA%20%23%262%3ABIT%20%2672%3ABNE%20k%3AJMP%20f%3A.e%20LDA%20%23%262%3ABIT%20%2672%3ABNE%20f%3AJMP%20k%3A.f%20LDA%20%2673%3ACLC%3AADC%20%231%3AAND%20%237%3ASTA%20%2673%3AJMP%20i%3A.k%20LDA%20%2673%3ACLC%3AADC%20%237%3AAND%20%237%3ASTA%20%2673%3A.i%20CLC%3ALDA%20%2678%3AADC%20%23%2680%3ASTA%20%2678%5CnLDA%20%2679%3AADC%20%232%3ASTA%20%2679%3ALDA%20%2674%3ACLC%3AADC%20%231%3ASTA%20%2674%3ACMP%20%2332%3ABNE%20l%3ALDA%20%2676%3ACLC%3AADC%20%2364%3ASTA%20%2676%3ALDA%20%2677%3AADC%20%230%3ASTA%20%2677%3ALDA%20%2672%3ACLC%3AADC%20%231%3ASTA%20%2672%3ACMP%20%2310%3ABEQ%20r%3AJMP%20a%3A.r%20LDA%20%267C%3ACLC%3AADC%20%23%261%3AAND%20%23%267%3ASTA%20%267C%3AJMP%20c%3ARTS%3A%5D%3ANEXT%3AMODE%200%3AVDU23%2C1%2C0%3B0%3B0%3B0%3B%3A*FX19%5CnCALL%20g%22%7D

Worth a look to get inline asm syntax looking nicely.

DIM MC% 400:FOR O%=0 TO 3 STEP 3:P%=MC%:[:OPT O%:.H% EQUW &55AA:.S% EQUW &3000:.Y% EQUD &FCF0C000:EQUD &030F3FFF:.g LDA S%:STA &70:LDA S%+1:STA &71:LDY #0:.t LDA H%:STA (&70),Y:INY:LDA H%+1:STA (&70),Y:INY:BNE t:LDY #0:INC &71:BPL t:.c LDA S%:CLC
ADC #32:STA &76:LDA S%+1:STA &77:LDA #0:STA &72:.a ASL A:AND #7:STA &73:LDA &76:STA &78:LDA &77:STA &79:LDA #0:STA &74:.l LDA &78:STA &7A:LDA &79:STA &7B:LDA &73:CLC:ADC &7C:AND #&7:TAY:LDA Y%,Y:LDY #0:STA (&7A),Y:INY:STA (&7A),Y:INY:STA (&7A),Y:INY:STA (&7A),Y
INY:STA (&7A),Y:INY:STA (&7A),Y:INY:STA (&7A),Y:INY:STA (&7A),Y:LDA #&F0:BIT &74:BNE e:LDA #&2:BIT &72:BNE k:JMP f:.e LDA #&2:BIT &72:BNE f:JMP k:.f LDA &73:CLC:ADC #1:AND #7:STA &73:JMP i:.k LDA &73:CLC:ADC #7:AND #7:STA &73:.i CLC:LDA &78:ADC #&80:STA &78
LDA &79:ADC #2:STA &79:LDA &74:CLC:ADC #1:STA &74:CMP #32:BNE l:LDA &76:CLC:ADC #64:STA &76:LDA &77:ADC #0:STA &77:LDA &72:CLC:ADC #1:STA &72:CMP #10:BEQ r:JMP a:.r LDA &7C:CLC:ADC #&1:AND #&7:STA &7C:JMP c:RTS:]:NEXT:MODE 0:VDU23,1,0;0;0;0;:*FX19
CALL g

(raw output)

The ASCII table (also for VDU codes) and the PLOT codes.

Either float text in editor or table on the right

See #47 - we can probably compromise on cycle accuracy in an "overclock" mode and see if that buys us more performance.

Specifically:

• Running more video clocks back to back instead of per instruction.
• Disabling the "cycle accurate" ticking inside instructions.
• Mayyyyybe some magical webasm trickery - see mattgodbolt/jsbeeb#150

Cut and paste this into owlet:

F.F=F TOF S.F:READF$:F%=F*LENF$:P.F$,F%:N.
D.F

Then right click on one of the F variables in the code and select "Change All Occurrences", then type X. The program now looks like this:

X.X=X TOF S.X:READF$:X%=X*LENF$:P.X$,X%:N.
D.X

I'd expect it only to change uses of the variable F, but it's also changed:

• F. is FOR abbreviated
• F$ and F% are also matched when not preceded by a keyword
• F in D.F is really a string constant not the variable F in this program (though if it were read into a numeric variable it would be a the variable F so this one more forgivable)

It's also failed to match variable F when preceded by a keyword (TOF). I guess it assumes C-like tokenisation rules.

I can see this feature would be handy for golfing code if it actually worked, but if it can't be persuaded to understand BBC BASIC tokenisation rules then perhaps it would be better turned off in owlet. When I first tried it on a real example it looked like it had worked, but actually it had missed a use of the variable, and so broke the code.

I figured we should hide the 10 / 20 / 30 etc if the user doesn't type them. Also if the user loads an example with 1/2 etc

Some lines are too long to expand. We should avoid doing so.

I haven't checked if the 'official' BBC BASIC byte tokens we generate fail when parsed by Twitter, but Rheolism did a fair amount of testing and discovered certain byte tokens need ANDing (doh) ORing with a high byte to make in through unscathed.

As shown in https://github.com/8bitkick/BBCMicroBot/blob/7bf7435d70cde7892d69c12bf5b934b0ee068ee2/tools/bbcbasictokenise#L17 by @ojwb

This may also fix issues with some bytes not being rendered as characters in the editor as he also reported

TIMER=1
P.TIMER

is tokenised as TIMER and not TIME R.

http://www.benryves.com/bin/bbcbasic/manual/Appendix_Tokeniser.htm

Though...I just tried this and filtering on the "continuation" bit doesn't account for the fact that TIME appears twice in the tokenisation table. Once as TIME (PMC bits set) and TIME (no special bits set). I'm not sure how BASIC handles this case. I am confuse.

    it("should not tokenise continuation tokens followed by alphanum", () => {
        checkTokens(["TIMER"], [{offset: 0, type: "variable"}]);
    });

should pass

E.g.

PRINT"I have a new line"'"here"

I'd expect @% to be highlighted just like A% (since they're both resident integer variables) but in A%=@%:@%=A% @% is green (looks like a slightly paler green than numbers).

It's true that the value of @% is used for formatting numbers, but other resident integer variables have special properties too (A%, C%, X%, Y% in CALL and USR; O% and P% in assembler).

If the code is

GOTO 30
PRINT "Doeesn't get here"
GOTO 30

or something, hitting enter at the top won't renumber the GOTOs which basically breaks the program

e.g. ? % . and ,` all appear as red in the editor which I think is Monaco's "this isn't handled" character?

From the HN: https://news.ycombinator.com/item?id=25256853

I got an old game from a magazine and pasted into the editor, didn't like for some reason.
edit:it was leading spaces before the line numbers - got rid of them and it works. Well, sortof, you have to press the JSBeeb button on the right.
Its not a very good game though
https://pastebin.com/dntnb8es

More likely a fault in jsbeeb's tokeniser, but one we can fix.