bug(common): fix handling of CODE_CONTEXTEX and CODE_INDEX offsets in Developer, Core, and Web

[mcdurdin]

• Verify that keyboards referenced in bug(developer): context(n) counts the nul statement as a character in Web & android #13276 issue all function correctly with updated Keyman Core
• We also need to verify behaviour for context(n) in the context part of the rule as a follow-on for this.
• Review changes in fix(developer): produce correct index parameter value for context in kmw compiler #13003 against this.
  • My current guess: the unit tests added in fix(developer): produce correct index parameter value for context in kmw compiler #13003 are inadequate because all the any/index tests use the first character in the store (i.e. charindex=0) so would have accidentally passed.
  • Further analysis: First guess is wrong. I added notany() to the unit test very late. This bumped the minver of the keyboard to 14.0, which caused the compiler to use v10+ rules, so the rules with nul no longer worked (bug(developer): kmc-kmn kmw compiler generates an incorrect offset for nul ... > context(n) when in v10+ mode #13306). However I didn't manually retest the keyboard after adding notany() so I didn't pick this up.