empty table cell in PGML

[Alex-Jordan]

Consider this MWE:

DOCUMENT();
loadMacros(qw(PGstandard.pl PGML.pl));
BEGIN_PGML
[#
    [..] [.A.]*
    [.B.] [.C.]
#]
END_PGML
ENDDOCUMENT();

This renders as if the [..] cell were never there in the first place, putting the "A" in the top left cell.

The top left cell is empty, and there is a legitimate reason to support that when you consider tables with column headers and row headers and it is natural for that top left cell to be empty. You can hack it by replacing [..] with [. .], but I think we should get to work even with [..].

PGML is effectively eradicating that [..] cell at some stage and I can't pinpoint where I could adjust that behavior. @dpvc , are you able to point me in the right direction?

[drgrice1]

One way to make this work would be to change line 273 of PGML.pl from
if (scalar(@{ $block->{stack} }) == 0) { $prev->popItem }
to
if (!$block->{isContainer} && scalar(@{ $block->{stack} }) == 0) { $prev->popItem }.

Note that would also allow empty "tags" (i.e., [<>]) as well. I think that could be useful also.

Another way to do this would be to add an allowEmpty => 1 flag to the table-cell block definition, and then make that line
if (!$block->{allowEmpty} && scalar(@{ $block->{stack} }) == 0) { $prev->popItem }.

[dpvc]

I don't see any downside to your first proposal, so I'd probably go with that rather than add a new flag.

I don't recall seeing the new [< ... >] tag before, so had to look it up. Looks good. I'm glad you have been able to add new functionality to PGML. Sometimes when I look at the code (nearly 20 years old now, I think), I am embarrassed by what a mess it is, but other times it seems that it does the job pretty well. Originally, it was only meant as an experiment to see what a Markdown-inspired syntax might look like for PG, and was never really intended for production use. It is gratifying that it has taken off as much as it has.