Bue Vester-Andersen

Ok, the "endcapsphrase, " opcode apparently has no effect at the moment. The endcaps indicator is apparently always used, and apparently always after the last word. If endcaps has not...

Further to capsphrase: It looks like begcapsphrase simply replaces begcaps, if it has been defined, even if the requirements for a phrase are not met. In the below example, we...

You wrote: > The idea is to not mix begcaps/endcaps with begcapsword/endcapsword and begcapsphrase/endcapsphrases. Yes, I think I understand, but AFAICS that is not what is happening. Begcapsphrase seems to...

Short test to demonstrate the state of things. [computer_braille.yaml.txt](https://github.com/liblouis/liblouis/files/2974561/computer_braille.yaml.txt)

Least i can do when I actually complained about it. (smile)

Hmm, what kind of test would you like to have? Perhaps "context" could be used, but I thought "alwaysletsign" would fall nicely into the family of "noletsign", "noletsignbefore" and "noletsignafter".

Here is a little test to illustrate the use for alwaysletsign. It is inspired by the Danish tables, but I hope it makes sense. [alwaysLetsign.yaml.txt](https://github.com/liblouis/liblouis/files/4302531/alwaysLetsign.yaml.txt)

I just meant that if we decided to implement the various noletsign opcode using attributes, it would be a good occasion to also implement the alwaysletsign opcode this way

I don’t know about the German requirements, but I guess I can live with the current situation. However, having a alwaysletsign opcode to supplement the other letsign related opcodes would...

Both in Windows, Linux and macOS, environment variables can be set with the scope of the current process (and subprocesses), the current user or system wide. However, it is not...