|
There was unfortunately a bug in the <TAG> tag on the development
system when the U.G. was processed. I spotted and fixed the bug,
but wasn't aware until just recently that that was the version
we distributed for BL8.
All tag arguments must follow the tag, on the same line and with
no intervening spaces.
Sorry for any trouble this caused.
patti
|
| I also encountered this problem by following the format of the
examples in the Document User's Guide. I think that allowing an
end of line delimiter between a tag and an argument list would
improve Document because the practice makes .SDML files more
readable. In addition, DSR developers used to recommend this
practice and customers converting from DSR to Document are therefore
likely to code files that way.
In my case, putting the end of line between the tag and the argument
generated an internal error as well as the undefined tag error.
Any time software generates a message with the words "internal error,"
you are guaranteed SPRs, regardless of what the documentation says
in the way of rules. (Page 2-2 of User Manual Vol. 1)
I'm curious. What is the technical reason for requiring the
( to be the next character after the >?
|
| There are some tags that have alternate forms:
form 1: <tagname>(probably a short argument)
form 2: <tagname>
probably a long argument
<endtagname>
When the tag translator reads "<tagname>" it does not want
to look more than one character ahead to decide which form
it is dealing with. (I'm not saying that it CAN'T, I'm saying
that it does not WANT to -- primarily for efficiency's sake.
Furthermore, if you look at form 2 in the above example, suppose
the long argument is supposed to be surrounded by parentheses?
How would the tag translator know whether the parentheses were
PART OF the argument as opposed to be DELIMITERS of the argument?
Example:
<literal>
(Here is the text that I want to be taken literally.)
And I would like this to be taken literally, too.
<endliteral>
The tag translator won't know what to do with those parentheses
until it encounters the <endliteral> tag.
SO the user would be forced to enter it as:
<literal>
((Here is the text that I want to be taken literally.)
And I would like this to be taken literally, too.)
bill
|