Improvements to VBA's Heuristic

[DecimalTurn]

Description

3 small improvements to VBA's heuristic:

• Currently, there is a non-negligible share of VBA code that does not include the vb-module metadata. This PR relaxes the need for that metadata for a .bas code file to be considered VBA. (See sample named d_WordCmds.bas).

• This PR reverts commit 94e7b20. It wrongly presumed that Option Compare (Text|Binary) is exclusive to VBA, but only Option Compare Database is ¹. (See sample named modFrm.bas for a VB6 file using using Option Compare Binary).

• This PR adds Microsoft Word related objects to VBA's heuristic. (See sample named QuickCards.bas for a Word related sample).

Checklist:

• I am fixing a misclassified language
  • I have included a new sample for the misclassified language:
    • Sample source(s):
      • https://github.com/shapiromatron/funcVBA/blob/main/VBA/d_WordCmds.bas (License MIT)
      • https://github.com/ashtarcommunications/verbatim/blob/master/desktop/src/QuickCards.bas (License GPL)
      • https://github.com/haku/yuki.scr/blob/d77210935915ef5dc0ce887f0bb97d2ae485ba59/modFrm.bas (License Apache 2.0)
  • I have included a change to the heuristics to distinguish my language from others using the same extension.

Footnotes

1. Option Compare Database can only be used within Microsoft Access (source) ↩

[lildude]

Why are you removing the samples/VBA/AccUnitLoaderConfigProcedures.bas? We generally don't remove samples and when we do it's normally because they're complete rubbish or useless. This appears to be a legit sample.

[DecimalTurn]

This has to be done because I'm reverting the changes made to the heuristics in 94e7b20. If we leave AccUnitLoaderConfigProcedures.bas as a VBA sample, the test that checks if samples are identified correctly based on the heuristics will fail. That's because it doesn't contain anything specific to VBA, so it will be classified as Visual Basic 6.0. The only hint that it's part of a VBA project would be the use of the Access-related prefix Acc for certain variables, but that criteria is not good enough to be included in the heuristics, since this is pretty much unique to the repo where the sample is from.

For those reasons, I would say the sample is bad, because it fits the type of files that you'd expect to have to use an override for. It's simply VB6 code that is part of a VBA project.