BASIC compiling issues about applecommander HOT 5 OPEN

Actually I should have just read the code :) It scans forward from each character position, not backwards. http://www.txbobsc.com/scsc/scdocumentor/D52C.html is a well-commented disassembly and the parsing code is right at the top. On Sat, Sep 26, 2020 at 11:16 AM Kris Kennaway ***@***.***> wrote: > Cool project! > > bt is a command-line wrapper that uses the same parsing code underneath > (FWIW it just needs a JRE and you can download the compiled .jar file which > is self-contained, so there shouldn't be an additional installation > requirement). > As you found, this is mis-parsed as follows: > > % java -jar bastools-tools-bt-0.3.1.jar --tokens foo.bas > NUMBER(0.000000), IDENT(FORX), = (d0), NUMBER(1.000000), IDENT(TO100), > SYNTAX(:), IDENT(?X), SYNTAX(:), IDENT(NEXTX), EOL > EOL > > The issue is that > https://github.com/AppleCommander/bastools/blob/master/api/src/main/java/io/github/applecommander/bastools/api/TokenReader.java#L85 > does not tokenize character-at-a-time as the real applesoft interpreter > does. Instead it consumes entire words at a time. This should probably be > changed to iterate character-wise to build up a string and emit tokens as > they match. But this doesn't use a prefix match, e.g. here's another case: > > "10 ATO=1" --> "10 A TO = 1" > > I suspect the way this is implemented is by searching *backwards* from > the current position in the input buffer towards the end of the previous > token. It might be worth cross-referencing the actual Applesoft parsing > code to confirm, and to make sure there are no other edge cases. e.g. ISTR > there is a special case for a token that is a substring of another token > and would otherwise prevent from matching it. > > > On Sat, Sep 26, 2020 at 4:02 AM Kay Savetz ***@***.***> > wrote: > >> I am creating a Twitter bot where people can tweet Applesoft BASIC code, >> then get a response tweet with a video of the code running. I'm using >> AppleCommander -bas to get the tweet-submitted code into a DSK file. >> >> My issue is: AppleCommander seems to be sensitive to the spacing in the >> input file. A line like: >> 0FORX=1TO100:?X:NEXTX >> will create a compiled BASIC program that results in ?Syntax Error. For >> a couple of different reasons. e.g it turns to: >> FORX = 1TO100. >> >> Is it possible to make AppleComputer create syntactically valid code, >> when it comes it spaces around keywords? >> >> Does the bt tool handle this sort of thing better? (Because man, that >> looks hard to install.) >> >> Thanks >> Kay >> >> — >> You are receiving this because you are subscribed to this thread. >> Reply to this email directly, view it on GitHub >> <#43>, or >> unsubscribe >> <https://github.com/notifications/unsubscribe-auth/AGSNUV3L3YINKOU46LH6SRDSHVKV5ANCNFSM4R2TYKEA> >> . >> >

On Sat, Sep 26, 2020 at 4:02 AM Kay Savetz ***@***.***> wrote: I am creating a Twitter bot where people can tweet Applesoft BASIC code, then get a response tweet with a video of the code running. I'm using AppleCommander -bas to get the tweet-submitted code into a DSK file. My issue is: AppleCommander seems to be sensitive to the spacing in the input file. A line like: 0FORX=1TO100:?X:NEXTX will create a compiled BASIC program that results in ?Syntax Error. For a couple of different reasons. e.g it turns to: FORX = 1TO100. Is it possible to make AppleComputer create syntactically valid code, when it comes it spaces around keywords? Does the bt tool handle this sort of thing better? (Because man, that looks hard to install.) Thanks Kay — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#43>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGSNUV3L3YINKOU46LH6SRDSHVKV5ANCNFSM4R2TYKEA> .

Cool project! bt is a command-line wrapper that uses the same parsing code underneath (FWIW it just needs a JRE and you can download the compiled .jar file which is self-contained, so there shouldn't be an additional installation requirement). As you found, this is mis-parsed as follows: % java -jar bastools-tools-bt-0.3.1.jar --tokens foo.bas NUMBER(0.000000), IDENT(FORX), = (d0), NUMBER(1.000000), IDENT(TO100), SYNTAX(:), IDENT(?X), SYNTAX(:), IDENT(NEXTX), EOL EOL The issue is that https://github.com/AppleCommander/bastools/blob/master/api/src/main/java/io/github/applecommander/bastools/api/TokenReader.java#L85 does not tokenize character-at-a-time as the real applesoft interpreter does. Instead it consumes entire words at a time. This should probably be changed to iterate character-wise to build up a string and emit tokens as they match. But this doesn't use a prefix match, e.g. here's another case: "10 ATO=1" --> "10 A TO = 1" I suspect the way this is implemented is by searching *backwards* from the current position in the input buffer towards the end of the previous token. It might be worth cross-referencing the actual Applesoft parsing code to confirm, and to make sure there are no other edge cases. e.g. ISTR there is a special case for a token that is a substring of another token and would otherwise prevent from matching it. On Sat, Sep 26, 2020 at 4:02 AM Kay Savetz ***@***.***> wrote: > I am creating a Twitter bot where people can tweet Applesoft BASIC code, > then get a response tweet with a video of the code running. I'm using > AppleCommander -bas to get the tweet-submitted code into a DSK file. > > My issue is: AppleCommander seems to be sensitive to the spacing in the > input file. A line like: > 0FORX=1TO100:?X:NEXTX > will create a compiled BASIC program that results in ?Syntax Error. For a > couple of different reasons. e.g it turns to: > FORX = 1TO100. > > Is it possible to make AppleComputer create syntactically valid code, > when it comes it spaces around keywords? > > Does the bt tool handle this sort of thing better? (Because man, that > looks hard to install.) > > Thanks > Kay > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#43>, or > unsubscribe > <https://github.com/notifications/unsubscribe-auth/AGSNUV3L3YINKOU46LH6SRDSHVKV5ANCNFSM4R2TYKEA> > . >