Re: Please vote for Syntax check in /FREE! -- RPG400-L

Anton Gombkötö wrote:

Hi Hans!

The efforts to "stuff the ballot box" have been noted, and hopefully
that won't negatively influence the prioritization of future work.

Somehow this is funny. There is a feature request database, people are
allowed to vote, and when one advertises this (cool!) feature and
people do actually vote (and i expect the members of the mailing lists
to be smart enough to know whether they want to vote for this or for
other features) and then there's this response...

Actually, i didn't feel to do a bad thing in advertising FRED and my
feature request. Have i been a bad boy? Maybe this sounds silly, but
that's the way i feel, thanks for your patience, please make me a
little bit smarter! Maybe it has something to do with local customs
and conventions, we have some weird things here in austria, too! :-)

Anton: I've been to Austria a couple of times, and I must say that there are more weird things in North American society than European society. ;-)

No, asking other people to support your position is not a bad thing. But I think it only makes sense to take that into account when counting the votes. After all, other worthy proposed enhancements do not get the benefit of a campaign. Some weighting has to be done to make any comparison between the alternatives fair.

...

In the meantime, remember that the RPG program verifier already does
a (more) complete check of the RPG source code.

Yep, but not @home, where i imported an exported iSeries project from
the workstation at work.

And i don't like pressing "verify" on every line. What gets me most is
matching the ' and the (). When i get one one the fingers when i want
to leave the line, it's OK, i still know what i initially wanted to
express. Several lines later this knowledge more or less slowly
vanishes.

No argument there. But consider that once you have difficulty matching the apostrophes and parentheses, your expression is probably getting too complex. At that point, you should probably consider splitting the expression into multiple statements anyways.

Also note that the syntax checker can't do much more of a check than
the error messages that get listed embedded within the source in the
compile listing.

I expect the free form syntax checker to be able to do the same as the
fixed form syntax checker. There hasn't been so many changes, it's
still the same language.

IMHO, the fixed form syntax checker has a harder task than the free
form syntax checker, because the fixed one hasn't to be told when a
statement ends via the semicolon.

No, the rules for fixed calcs are easier for the syntax checker to handle. When scanning backwards through the specs, you can easily find the beginning of the statement by finding the C-Spec with a non-blank opcode entry. (If it's got AN or OR in positions 7-8, just back up a little bit more.)

As I mentioned elsewhere, if you back up and find a semi-colon in /FREE calcs, the meaning of it is different depending on context. It's not a statement terminator if it's coded in a comment or a literal, and deciding that is not an easy thing.

Keeping track of the statement terminator semi-colons can be done. It's just that the editor has to keep track of them. And the resources need to be allocated to implement the functionality. Within the current syntax checker architecture, this kind of thing can't currently be done easily.

Cheers! Hans

Cheers! is said when one lifts a glass with a liquid in it, or? Then a
(virtual) Cheers! to you, too, Hans!

I'm happy that you are around, although your mails aren't making me
happy all the time, but whose do? :-)

"Cheers!" is said when lifting a glass with a certain specific liquid in it, eh? One particular brown beverage made in a town just north of Regensburg comes to mind. But several other potables found in that part of the world would do just as well!

Prost! Hans