[v1.7.0.10] "Wait For Spoken Response" can CTD VoiceAttack on profile load

[Pfeil]

If the "Wait For Spoken Response" action is given a large numeric range in the "Responses" field(E.G. "[1..1000000000]"), VoiceAttack will crash to desktop when the change is applied to the profile.

Subsequent profile loads will cause VoiceAttack to crash, making the profile "poisoned", and effectively useless.
When this occurs, VoiceAttack can only be started by using the "-profile" command line parameter with a working profile, or by removing the database file/replacing it with a working backup.

Unlike normal commands, there is no warning given about the "Large number of dynamic command created".


Does this action always generate command phrases, with VoiceAttack ignoring them if they're recognized when the action is not running?
If something like this is the case, should the action register on the "derived commands" counter?


As an aside, this action cannot be tested before the change is saved to the profile(thus my assumption about the inner workings above).


EDIT: Fixed in v1.7.0.11

[Gary]

I see what you are saying.  No, there are no commands being added, rather, entries into the speech dictionary.  Back to the drawing board o_O

[Pfeil]

there are no commands being added, rather, entries into the speech dictionary.

I used the term

command phrases

To refer to that, though thinking about it VoiceAttack probably keeps an identical list itself so it can match what the speech recognition engine recognizes with the command(or in this case, action) it belongs to.


The reason I brought it up is because it will still affect profile loads times, as the phrases have to be generated regardless.

[Gary]

I may have to put a hard cap on the number of allowable responses per action.

[Gary]

Hard-capped at 250.