Well, I don't think it is an issue. Segments should be auto-propagated only if they are identical to make sure their translation is the same. Their status to auto-propagate based on regular expressions would make them similar but not identical. Thus auto-propagated translation from one segment might be incorrect for the other segments which are not the same.
There must be a big misunderstanding in our communication.
Let us conclude that I wasn't clear enough in describing the issue I'm experiencing.
Please allow me to give it another try, hopefully with a better result.
When I have ranges like this that only differ in numbers, I only have to translate the first one. CafeTran Espresso will adapt all numbers for my (thank you for that!):
However, when a glossary is loaded that contains a specific non-translatable that is a regular expression, the automatic adaption of numbers doesn't work anymore.
How come?
Here is a demo:
alwayslockyourbike
(This is a Trados project.)
If CafeTran Espresso can auto-propagate this, what are the settings that can prevent propagation?