OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

cgmo-webcgm message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re[3]: [cgmo-webcgm] ISSUE: delimited string inconsistency?


At 10:57 AM 5/6/2005 -0400, Benoit Bezaire wrote:
[...]
Now do the same exercise with type 3 and/or 4, and you'll notice it
doesn't work. Type does _not_ tell you how many points there are, not
for polygon and polybezier.

I don't feel like writing code to figure out what this is suppose to
be :)
"3 0 0 100 0 100 200 0 200 4 0 0 0 0 100 0 100 0 100 0 100 200 100
200 100 200 200 0 200 0"

LH> I guess I could live with it either way, but don't see any reason to apply
LH> more powerful syntax to a simple problem like region.
I'm thinking we don't have much of a choice, it has to be a delimited
string (for multiple sub-regions).

Zounds!  You're right of course.  I was only thinking of the fixed-data rectangle and ellipse cases.  (And not thinking further, 'cause it's all taken care of by the SDR coding for WebCGM instances.)

[...]
... actually we agree in Munich that delimited string
would only be used if multiple strings are return, else it's a single
string.

I didn't remember that.  So you're saying that things like 'name' will be simple string if 1 name, delimited string if >1 names?  I haven't thought it all through completely, but ... are we sure that is well-defined for "get/setApp..." and parsing purposes?  I guess 'name' is the only potential issue.  Looking at 3.1.1.3, I think we're safe, and we could detect the single versus multiple cases as follows...

If the first non-wsp character of the content is QUOTE or APOSTROPHE, then it's a sub-string delimiter and we're dealing with delimited string and there must be more than one sub-string (else it breaks the rule to use simple string).

Btw, as long as we're on this topic, a related but slightly different issue...  In 5.5, there is a note, "need a link for ValidChar".  That character set would be the union of:

1.) valid characters of 'name' ApsAttr
2.) valid characters occurring in linkuri ApsAttr, of 3 subparts:
a. URI (or IRI, if we should decide to go that way)
b. 'linkTitle'
c. behavior
3.) valid characters for "number" (in 'region').

Does #2b present us with any problems?  E.g., what would be the JS call for setAppStructureAttr for the 'linkuri':
-- uri:  http://www.itedo.com/tc.cgm
-- title:  Benoit's test case
-- behavior:  _replace

Whose rule do we use to handle the APOSTROPHE in "Benoit's"?  JS escaping rules?  Or do we need an escaping rule in our definition of 5.5 delimited string?  (My instinct is the latter.)

-Lofton.


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]