r3wp [groups: 83 posts: 189283]
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

World: r3wp

[Parse] Discussion of PARSE dialect

BrianH
1-Jul-2006
[1240x5]
Graham, yes I did. I know parse well enough to run it in my head.
expand-macros: func [data [string!] macros [block!]
    /local whitespace macro-rule macro expansion here there
] compose [
    whitespace: (charset " ^/")
    macro-rule: make block! 2.5 * length? macros
    foreach [macro expansion] macros [
        macro-rule: insert macro-rule
            compose [here: (macro) there: [whitespace | end] '|]
    ]
    macro-rule: head remove back macro-rule
    parse/all data [some [
        macro-rule (
            macro: copy/part here there

            there: change/part here select/skip macros macro 2 there
        ) :there |
        skip
    ]]
    macro-rule: none
    data
]
Whoops, small fix...
expand-macros: func [data [string!] macros [block!]
    /local whitespace macro-rule macro expansion here there
] compose [
    whitespace: (charset " ^/")
    macro-rule: make block! 2.5 * length? macros
    foreach [macro expansion] macros [
        macro-rule: insert macro-rule
            compose [here: (macro) there: [whitespace | end] |]
    ]
    macro-rule: head remove back macro-rule
    parse/all data [some [
        macro-rule (
            macro: copy/part here there

            there: change/part here select/skip macros macro 2 there
        ) :there |
        skip
    ]]
    macro-rule: none
    data
]
Wait, there is another error.
Tomc
1-Jul-2006
[1245]
I am glad to see someone else using here and there ;)
BrianH
1-Jul-2006
[1246x2]
expand-macros: func [data [string!] macros [block!]
    /local whitespace macro-rule macro expansion here there
] compose [
    whitespace: (charset " ^/")
    macro-rule: make block! 2.5 * length? macros
    foreach [macro expansion] macros [
        macro-rule: insert macro-rule
            compose [here: (macro) there: [whitespace | end] |]
    ]
    macro-rule: head remove back macro-rule
    parse/all data [some [
        macro-rule (
            macro: copy/part here there

            there: change/part here select/skip macros macro 2 there
        ) :there |
        thru whitespace
    ] to end]
    macro-rule: none
    data
]
That should do it.
Graham
1-Jul-2006
[1248]
I don't feel so bad that my rule wasn't working !
BrianH
1-Jul-2006
[1249x2]
; Now whitespace is dealt with
expand-macros: func [data [string!] macros [block!]
    /local whitespace macro-rule macro expansion here there
] compose [
    whitespace: (charset " ^/")
    macro-rule: make block! 2.5 * length? macros
    foreach [macro expansion] macros [
        macro-rule: insert macro-rule
            compose [here: (macro) there: [whitespace | end] |]
    ]
    macro-rule: head remove back macro-rule
    parse/all data [some [any whitespace [
        macro-rule (
            macro: copy/part here there

            there: change/part here select/skip macros macro 2 there
        ) :there |
        to whitespace
    ]] to end]
    macro-rule: none
    data
]
You might want to add tabs to the whitespace charset.
Graham
1-Jul-2006
[1251x2]
why do you need to reset macro-rule to none?
it's a local so memory will be released anyway ..
BrianH
1-Jul-2006
[1253]
You don't need to, but if you don't it will sit in memory until the 
next time you run the function.
Graham
1-Jul-2006
[1254]
Oh ... ok.
BrianH
1-Jul-2006
[1255]
Memory for locals is not released.
Graham
1-Jul-2006
[1256]
when the function is exited ?
BrianH
1-Jul-2006
[1257]
Nope.
Graham
1-Jul-2006
[1258]
Hmm.
BrianH
1-Jul-2006
[1259]
Only on recursed functions.
Graham
1-Jul-2006
[1260]
I guess it can't be otherwise we wouldn't need to use 'copy on series
BrianH
1-Jul-2006
[1261]
It's a speed optimization. This might change with REBOL 3.
Graham
1-Jul-2006
[1262]
memory use is a large with Rebol.
BrianH
1-Jul-2006
[1263]
The series copy thing is something different.
Graham
1-Jul-2006
[1264]
speed or memory ..
BrianH
1-Jul-2006
[1265x2]
Most of the excessive memory overhead of REBOL is just sloppy (no 
offense Carl). It's not much of a problem for most, but I have run 
into memory limits when running on embedded or handheld platforms, 
or running hundreds of instances on servers.
You have to design your platform for memory efficiency or it won't 
be.
Graham
1-Jul-2006
[1267]
Little problem :(

** Script Error: Invalid argument: make bitset! #{
0004000001000000000000000000000000000000000000000000000000000000
}
** Where: expand-macros
** Near: parse/all data [some [any whitespace [
            macro-rule (
                macro: copy/part here there
           ...
Tomc
1-Jul-2006
[1268x3]
whi the parens on   whitespace: rule?
nevermind it is in compose
but that can just be a static rule outside of compose
BrianH
1-Jul-2006
[1271]
More efficiency. I build my charsets once, rather than on every call.
Graham
1-Jul-2006
[1272]
should probably build the rule outside the function as well since 
it only should be rebuilt when the macro list changes
Tomc
1-Jul-2006
[1273x3]
cant say to whitespace
wish you could
tipicaly get around with
Graham
1-Jul-2006
[1276]
that's where my rule got stuck .. on to whitespace
Tomc
1-Jul-2006
[1277x2]
bs: complement ws
any bs
BrianH
1-Jul-2006
[1279x3]
Yeah, I'm changing my code now.
expand-macros: func [data [string!] macros [block!]
    /local ws non-ws macro-rule macro expansion here there
] compose [
    ws: (charset " ^/") non-ws: (complement charset " ^/")
    macro-rule: make block! 2.5 * length? macros
    foreach [macro expansion] macros [
        macro-rule: insert macro-rule
            compose [here: (macro) there: [ws | end] |]
    ]
    macro-rule: head remove back macro-rule
    parse/all data [some [any ws [
        macro-rule (
            macro: copy/part here there

            there: change/part here select/skip macros macro 2 there
        ) :there |
        some non-ws
    ]] to end]
    macro-rule: none
    data
]
See, I don't have to test. Graham has the data, he can test :)
Tomc
1-Jul-2006
[1282]
that is the part I can't allways be bothered with either
Graham
1-Jul-2006
[1283]
LOL ...
BrianH
1-Jul-2006
[1284x2]
When it's my data, I test it myself. After the major logc bugs are 
worked out in my head of course...
REBOL catches spelling errors (logc -> logic).
Graham
1-Jul-2006
[1286x2]
Have you guys made any submissions to RT for parse improvements for 
R3?
Any further submissions ie.
Tomc
1-Jul-2006
[1288]
RT can just use the pending suggeations from before R3
BrianH
1-Jul-2006
[1289]
Gabriele and I have worked extensively on such submissions.