Bug Functions run times are inconsistent

[roovyshapiro]

Bug functions are inconsistent regarding how many times they are run. Some run very often and some run rarely.

Some functions exit without executing since they don't fulfill all conditionals
Ideally, if a function CAN apply to a file, it should execute once before exiting.

I think one way to solve this issue might be to generate some statistics of which bugs were run and how many times. It can then be appended to the answer key.

[sc1341]

I think this is a great idea... just thinking of some psudo code we could write to the end of the answer key by going down a few lines and then printing a "stats" variable that could be a dictionary. Each key could then be the name of the bug and the value could be the number of times it was run.

[roovyshapiro]

@sc1341
Besides being helpful for the end user, I think that in the process of creating these stats we can determine the root cause of why some functions are being applied more than others. If a bug function is called on, we should record the function and wether or not num_bugs was decremented. We should then be able to tell which functions are exiting without applying and can adjust their frequence accordingly. For example, missing_blanks() has a pretty loose conditional since it's not finding a specific character match, its just finding a line and the whole list is lines, so it is applied almost each time it is called. Contrast that with equal_swap() only applies if it finds '=' or '=='.

For a bit of background, in the original chronological order that the bugs were written, they got progressively "smarter" as new edge cases and other related information was discovered.

In an ideal scenario, a bug function tries up to 100 (or 50) times to apply a bug and if it can't, it removes itself from the function list. So a file without any "="'s shouldn't keep calling equal_swap().

In practice, not every bug function was created with that in mind. Some will try to apply and may exit without doing anything and without decrementing num_bugs. It's difficult to manually determine which ones were set up like this. But stats would help to even out the amount of bugs that are applied.