Quote:
Originally Posted by Sabardeyn
Silentguy,
Are you talking about this: Calibre Manual: Creating AZW3, EPUB, MOBI Catalogs>Prefixes? (I know that seems like I'm asking you to repeat yourself, again. But I'm ignorant here so I'm trying to make sure we're talking about the exact same thing.)
So what I think you're saying is, when it comes to Prefix tests, and values returned by them, the results should be: - Missing test: #read:yes (which would find all yes/true values)
- Missing test: #read:no (which would find all no/false values; but not null/unspecified)
- Incorrect logic: #read:true (which should return yes/true if a Y or N value is present; false if null/unspecified)
- Incorrect logic: #read:false (which should return yes/true if there is a null or unspecified value only; false if there is any other value present)
- Incorrect logic: #read:"" or Undefined (which should return the same results as #read:false)
Did I get any part of that correct? If I did not, could you just state which test # is wrong?
I still think this has (peripherally) to do with the decision to go with a tri-state Y/N field. I vaguelly recall that there was some question about logical comparison's at that time, but I cannot recall any other details.
Edit: AHA! My Search-Fu is strong today! Possibly relevant topics:
Opinions wanted on details of new GUI features
Default value for read field (Yes/No-field)
default value for defined y/n column
|
Your list is pretty much dead on! Well, for 5. I would think that the empty string is some weird default and not returning anything is acceptable. And 1. is only missing in name, cause 3. currently returns what I would expect for 1.
And you guys are correct, used on an two state box it would make sense the way it is, cause the value that is dropped is undefined, which means true=yes false=no is a sensible comparison. But it does not match up with the normal search.