I have to say that, as often as not I find the tool tip with the syntax of the current construct more distracting than helpful, especially since it often covers the few lines prior, which is what one is most likely to want to look at, but it is particularly annoying when the tip is completely irrelevant to the situation. E.g., type a define variable for a class and it comes up with the tip for class, i.e., the word which starts a class definition, not something appropriate to a define variable context.
Two more examples, tips for MESSAGE and FOR *in a quoted string*!
Not to mention the prompts for completion every time one types a colon ... not only a PITA for blocks, but when writing OO code it happens all the time!
tamhas schrieb:
Not to mention the prompts for completion every time one types a colon ... not only a PITA for blocks, but when writing OO code it happens all the time!
You are aware, that you can turn that off in the preferences?
Selectively? No, I am not actually aware of where to turn it off ... doesn't seem to be under POE/Editor, but the point is not that I object to the feature as that I object to inappropriate tips. If I am in the middle of a quoted string ... editiing it, no less, so the closing quote is there too ... there is no purpose in prompting me about keywords. Likewise, if I type a colon in a place where accept-changes is not a valid construct, then there is no positive function in prompting me for it and other illegal things to insert.
Whereas, meaningful tips and meaningful completion prompts are just fine.
You complaint about "Not to mention the prompts for completion every time one types a colon ... not only a PITA for blocks, but when writing OO code it happens all the time!" That's a feature you can selectivly turn off in the preferences. Progress OpenEdge -> Editor -> Assistance -> Automatically propose completion on . and :
Well, yes, but I can't turn it off selectively. It is fine to prompt me with possible method or property names, especially after I have typed a few letters so the list is short. It is prompting accept-changes for the : in a Do or Repeat or a For that I object to because it is not a legal completion there.
that would fit the bill... hmmm..
this works nicely for me as a template. no more DO:accept-changes..!
DO:
$
END.
thanks for mentioning it as a problem - I too often live with these annoyances rather than fixing them.
Templates are fine for some things, but if one types reasonably well, such a trivial template as this is not worth the effort to remember to use it, really. The big thing it "fixes" is the loss of fluidity one gets when the d*mn accept-changes gets added and one has to go back and undo it.
tamhas schrieb:
Well, yes, but I can't turn it off selectively. It is fine to prompt me with possible method or property names, especially after I have typed a few letters so the list is short. It is prompting accept-changes for the : in a Do or Repeat or a For that I object to because it is not a legal completion there.
Turn it off completely - if you call it that way.
You can still trigger the autocompletion with CTRL-SPACE like in most modern IDE's
Just because there is some kind of workaround doesn't make it not an error to offer tips or completion which is inappropriate to the context.
tamhas schrieb:
Just because there is some kind of workaround doesn't make it not an error to offer tips or completion which is inappropriate to the context.
Sorry for giving you advices then to make it "workable". Will not happen again.
The topic of DO:ACCEPT-CHANGES has been discussed at every Info-Exchange session in the past 5 years.
You may log a bug if you care enough.
I have been trying the workaround and I think it is better than the alternative.
I am raising the point exactly because Sunil has recently been talking about some syntax aware improvements which might be in the works, like syntax aware search and replace. This would seem to fall in the same category. I'd like to create awareness so that it doesn't get missed.