I thought master looks for upper case characters that match the regex and ALSO forms a special regex based on the supplied acronyms that are not just made of upper case characters.

https://github.com/lsst/lsst-texmf/blob/master/bin/generateAcronyms.py#L42

CAP_ACRONYM = re.compile(r"\b[A-Z][A-Z]+\b")

Just 2 or more upper case characters.

Yes, but I'm talking about https://github.com/lsst/lsst-texmf/blob/master/bin/generateAcronyms.py#L344

That's the list of defined entries in glossarydefs.csv+myacronyms-skipacronyms.It is used to find all the matches between the glossary definitions and what is used.

The regular expression I am referring to, is used on line 352 to find all potential acronyms-like strings (as per line 349) in the text. And it is looking on only full uppercase strings.

From the difference with the matches, the missing ones are found.

I have checked this for both glossary and acronyms and it looks good. One small comment on github was fixed. There is still one comment from Tim on the regexp.. it could perhaps be simplified but I would not delay much on merging this. It is an improvement and it is good to flag acronyms separately to glossary entries.