I was wondering whether non-translatables glossaries should default to whole words?
In the current situation, when I want to add 'db' to my non-translatables glossary, I have to surround it with |\b and \b to ensure that words like 'tidbit', 'wordbook' aren't flagged.
Especially when you have many of these non-translatables, it can be quite some work to mark them up.
Wouldn't it be more logic to let CafeTran Espresso 10 Croissant interpret all expressions in non-translatables glossaries as whole words, except when the line starts with a pipe (making it a regular expression)?
Another possible improvement of non-translatable glossaries would be that 'everything that's present in the glossary' and stands between brackets (of any type) will automatically be interpreted as a new 'spanning' non-translatable:
Same for variations with *, + etc.
So, if COMFORT is in your non-translatables glossary and your source text contains COMFORT+, COMFORT*, COMFORT® etc, the whole fragment should be detected as non-translatable.