Subject: | |
From: | |
Reply To: | VANCE,JEFF (HP-Cupertino,ex1) |
Date: | Thu, 15 Aug 2002 18:55:35 -0400 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Wirt wrote:
> 1. If line begins with comment symbol, skip to next line
> immediately.
So the CI grabs the first token, upshifts it (not inline) and if it
is '#' or 'COMMENT' does the appropriate echoing to $STDLIST and reads
the next line.
> 2. Line is scanned for "%" symbols. If found, and the
> macrovariable is
> determined to be legitimate, the macrovariable is excised and
> its value substituted into the line.
For the CI it does the variable substitution step here, which is
recursive.
> 3. Line is scanned for "*" symbols. If found, and the
> relative cell
> address is determined to be legitimate, the row, page, column
> address is made real and substituted into the line.
Not applicable to the CI.
> 4. The line is then handed off to QueryCalc's command
> interpreter, just
> as if the user had typed the substituted (if any) line in by hand.
This will work and should be fully backwards compatible -- well, noting
the fact that $STDLIST output can be different for comment commands than
it is today. Today, comment lines in $STDLIST have all variables
resolved. The approach above would cause those lines to appear without
any variables resolved. I don't know how important this difference is,
but it *may* impact spoolfile scanning tools -- don't know.
Step 1 is what is not done by the CI today -- we don't parse the line
until after substituting all variables. Since IO redirection is not
acted upon in comment commands this scheme should work fine, given my
$STDLIST caveat above.
Jeff Vance, "CSY"
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
|
|