Ladislav,
Thanks for the comments and catching the 'yield reference I missed in
the doc string.
collect: func [
[throw]
{Collects values, returning them as a block.}
'emit "Word used to collect values"
block [any-block!] "Block to evaluate"
/into dest [series!] "Where to append results"
/only "Insert series values as series"
] [
; Create a new context containing just one word (the EMIT
; argument) and set the 'emit word to refer to the new context
; word. Note the care taken to suppress possible conflicts and
; undesired evaluations.
emit: reduce [emit]
emit: first use emit reduce [emit]
use [dst] copy/deep [
; copy/deep lets us use just as the fallback value [] here.
dst: any [:dest []]
; The only difference in the function bodies is insert[/only].
set emit func [value [any-type!]] either only [
[insert/only tail :dst get/any 'value get/any 'value]
][
[insert tail :dst get/any 'value get/any 'value]
]
do bind/copy block emit
head :dst
]
]
GS> Hmm, but this is a different matter, being able to address
GS> different output blocks at the same time. So we need to figure out
GS> if it's more useful to be able to do the above, or to be able to
GS> do:
GS> do-something: does [
GS> repeat i 10 [emit i]
GS> ]
GS> collect [do-something]
I see the value in both. Some things, like CGI scripts, that build up
a single output buffer benefit from the global approach, but it's not
nearly as convenient for use in other functions that need to build up
a local result and return it, which was my main target.
I also like the context you get by specifying the "word" you want to
collect, rather than saying 'collect one place, but knowing that 'emit,
used somewhere else, applies to it. Localized thinking I guess.
-- Gregg
--
To unsubscribe from the list, just send an email to
lists at rebol.com with unsubscribe as the subject.
