langford@reed.UUCP (Chris Langford) (10/07/88)
Hi,
I have been writing in hyperTalk for a while, and there is a problem I
have run into several times. When I try to run scripts that contain
if statements with this format:
on mouseUp
repeat with i = 1 to the number of cards
if field 1 of cd i contains "RTFM" then
foobar
end if
end repeat
end mouseUp
-- where foobar is any command or set of commands
I get an error message that says "No such card". The problem is fixed if
I write the script:
on mouseUp
repeat with i = 1 to the number of cards
get field 1 of cd i
if it contains "RTFM" then
foobar
end if
end repeat
end mouseUp
or like this:
on mouseUp
repeat with i = 1 to the number of cards
go cd i
if field 1 contains "RTFM" then
foobar
end if
end repeat
end mouseUp
The second correction takes longer to run than the first, but I think
the original script would run even faster, if it worked (one less line).
Am I missing something completely, or is this a problem with HC?
Has anyone else run into this problem?
Thanks for any help.
--
Chris Langford |
tektronix!reed!langford | "And to everyone else out there,
langford@reed.bitnet | the secret is to bang the rocks
| together, guys." Scot.Kamins@f555.n125.z1.FIDONET.ORG (Scot Kamins) (10/09/88)
The solution to your problem is to put the field reference within parentheses:
if (field 1 of cd i) contains "RTFM" then
Hypertalk gets confused when you give it an expression to evaluate as a field address if that expression is anyplace but at the end of a
line. So if the line were something like
else put foo into field 1 of cd 1
it would have no problem with it.
The solution I suggest will work every time.
--
Scot Kamins - via FidoNet node 1:125/406
UUCP: ...!sun!hoptoad!fidogate!555!Scot.Kamins
ARPA: Scot.Kamins@f555.n125.z1.FIDONET.ORGtgl@zog.cs.cmu.edu (Tom Lane) (10/10/88)
In article <10537@reed.UUCP>, langford@reed.UUCP (Chris Langford) writes: > [...] When I try to run scripts that contain if statements with this format: > > if field 1 of cd i contains "RTFM" then > > I get an error message that says "No such card". The problem is fixed if > I write the script: > > get field 1 of cd i > if it contains "RTFM" then > > or like this: > > go cd i > if field 1 contains "RTFM" then > The problem goes away if you add some parentheses: if (field 1 of cd i) contains "RTFM" then Evidently, when you write the command without parentheses, HC parses it as if field 1 of cd (i contains "RTFM") then hence the error "no such card" (unless you have a card named "false"). It would seem that after "card" HC looks for a general expression, whereas after "field" it looks for a more restricted expression---since Chris's last example is not parsed as if field (1 contains "RTFM") then <flame on> This sort of undocumented inconsistency is the reason why people have been complaining since day one about HyperCard's weird syntax. Has the HC team ever heard of formal syntax specifications (e.g., BNF notation)? If so, why do they still refuse to publish one? There is nothing about the relative precedence of arithmetic operators and chunk-expression operators in the HyperCard Script Language Guide, and certainly nothing to suggest that arithmetic expressions in different parts of a chunk expression will be parsed differently. I get the impression that Apple believes that formal specifications would scare away potential scripters who don't understand formal methods. I beg to differ. Failing to publish a formal specification for HyperTalk syntax is not going to make scripting any easier; it just creates more room for problems of the kind shown above. (A truly scary thought is that Apple doesn't *have* a formal specification, and will be as surprised as the rest of us by problems of this ilk.) <flame off> I don't want to offend anybody at Apple, but I do think it's high time to publish a *complete* specification of the language. Failure to do so suggests that they do not understand their product any better than the rest of us. -- tom lane Internet: tgl@zog.cs.cmu.edu UUCP: <your favorite internet/arpanet gateway>!zog.cs.cmu.edu!tgl BITNET: tgl%zog.cs.cmu.edu@cmuccvma
tgl@zog.cs.cmu.edu (Tom Lane) (10/14/88)
Here's some more info about the problem Chris Langford posed, namely that
if field 1 of cd i contains "RTFM" then
fails while
if field 1 contains "RTFM" then
does not.
It seems that following "card", "card id", "background", or "background id"
(or abbreviations of same), HyperCard expects to find an arbitrary expression.
This expression is everything up to the next "of", "in", "then", or similar
keyword, end-of-line, or right parenthesis that matches a prior left
parenthesis. Hence
if (field 1 of cd i) contains "RTFM" then
works.
The same rule applies for expressions in a chunk-selector (i.e., following
"char", "word", "item", or "line").
BUT: following "button", "button id", "field", or "field id", HyperCard
expects to find a "factor" as defined on p. 49 of the Script Language Guide.
This is:
* a simple source of value (constant or variable name);
* a parenthesized expression;
* a factor preceded by "-" or "not".
This explains why Chris's second example works: the "contains ..." is not
part of the factor which is the field number. If you try something like
put the name of field 1+1
you get a complaint, not the name of field 2.
A factor is what is taken to be the argument of a parenthesis-less function;
for instance, "the sqrt of 4+12" means "sqrt(4)+12", not "sqrt(4+12)".
While I can think of arguments in favor of using either general expressions
or factors as the value selecting an object, I find it hard to see why
fields and buttons should be treated differently from cards and backgrounds
in this regard. I also don't find anything about this topic in the Script
Language Guide. (I derived the above rules by experiment.) Anybody from
Apple care to comment?
--
tom lane
Internet: tgl@zog.cs.cmu.edu
UUCP: <your favorite internet/arpanet gateway>!zog.cs.cmu.edu!tgl
BITNET: tgl%zog.cs.cmu.edu@cmuccvma