skinnable user-interface with Text Load and Parser

jorge · 10 Jul 2005

Great stuff rl, if you keep using it I'll implement new missing parsing rules if needed.

rl wrote: Questions (advice needed): - The parser in the example works only when the exact sequence of lines (key=value; settings) is written in the file. How could I change the parser, so that the lines could be in any sequence?

The following schematic would match either "aaa" or "bbb" but not "aaa bbb" and not "bbb aaa":

The following will match 0 or more matched of either "aaa" or "bbb", I think it's what you need in your schematic:

jorge · 10 Jul 2005

rl wrote: Can I setup the parser, so that comment lines could be written anywhere? (currently only the first line)?

I think if you use the Kleene start rule as I suggest in the last post it will sort that too. The comment will just become one of the optional lines that can appear anywhere zero or more times. I'm not sure but I think there is a small bug that might make your parsing not work after you delete a parsing component. If you save and load that bug doesnt show and it should always work.

jorge · 10 Jul 2005

rl wrote: How can I specify special (non Alnum) chars in "Rule Range" - I would like to specify something like a-zA-Z0-9,.:-_+()[]{} ?

There should be a new component to do that. For now you'll have to look at an ASCII table: http://www.lookuptables.com and do something like the following:

Shifrin · 10 Jul 2005

That's way cool rl 😀

Shif.

rl · 11 Jul 2005

jorge wrote: Great stuff rl, if you keep using it I'll implement new missing parsing rules if needed.

Oh yes, thank you for the quick reply. I will have a look on your examples when I'm at home....

lalo · 11 Jul 2005

nice stuff rl! yeah 🙂

rl · 11 Jul 2005

Thanks Jorge, the parser now works as expected, I've updated the schematic. Now it's your turn to fix the bugs 😉

Tom7777 · 11 Jul 2005

I was planning on making some similar things when i have time to do so, thanks for the example rl 🙂

infuzion · 13 Jul 2005

Has anyone noticed (w _e?) that the loaded config in the Data:Text is double-spaced?

rl · 13 Jul 2005

infuzion wrote: Has anyone noticed (w _e?) that the loaded config in the Data:Text is double-spaced?

it's this error (see above):

when loading file thru "After Load": error: synthmaker puts extra chars in string (after save: line ending is 0x0d 0x0d 0x0d 0x0a)

infuzion · 13 Jul 2005

jorge wrote: The following will match 0 or more matched of either "aaa" or "bbb", I think it's what you need in your schematic:

Perhaps I don't understand Rules, but it seems the 0 of "0 or more" is really "exactly matching":

or is it just me?

rl · 15 Jul 2005

AFAIK it makes sense. The rule means [0 ore more of (aa or bb)] but you have vv in your string. Should be True when you remove vv.

infuzion · 15 Jul 2005

rl wrote: AFAIK it makes sense. The rule means [0 ore more of (aa or bb)] but you have vv in your string. Should be True when you remove vv.

Perhaps I'm thinking back in my Math-Logic classes, but I would assume "0 or more" means the search condition (in this case "aa or bb") is found (eg subset) at all, or not at all (the "0") the the search is True. The example I posted makes me believe the Star-Rule is really an intersection, not an union (that the 0 indicates to me).

But I got a D trying to prove that 1+1=2, so what do I know? 🙂

rl · 15 Jul 2005

now I'm not so sure, Maybe you are right. I did a little regular expression test :

# succeds: lindner@lux1:~> echo aabb | grep "aa*\|bb*" aabb

# also succeds: lindner@lux1:~> echo aabb | grep "bb*\|aa*" aabb

# surprisingly also succeds ( but indeed: ZERO or more) lindner@lux1:~> echo aavv | grep "bb*\|aa*" aavv

# but did I write the right regular expression? Maybe it's: lindner@lux1:~> echo aavv | grep "\(bb\|aa\)*" aavv

now I'm totally confused: does the example implement

\(bb\|aa\)*

0 ore more of ( aa or bb ) or

bb*\|aa*

(0 or more of bb) or (0 or more of aa) ??? H e l p - Jorge, I need an explanation!!!

jorge · 15 Jul 2005

rl, you were right with:

rl wrote: The rule means [0 ore more of (aa or bb)] but you have vv in your string. Should be True when you remove vv.

rl · 15 Jul 2005

..then your schematic can be expressed by the regular expression:

\(bb\|aa\)*

and everything makes sense.