[Compose key in 1.27 to follow X11, not Hacker's]

[Spike-from-NH]

[Compose key in 1.27 to follow X11, not Hacker's]

[Julow]

Thanks for looking at this!

The compose key feature request is tracked by #204
I planned on using xkb's sequences but adding Hacker's Keyboard's ones is a good idea.

The unicode entry feature request is tracked by #191

I'll close this issue as duplicate but I'll make sure to keep your nice ideas.

[Julow]

I've found out that Hacker's Keyboard's compose sequences are very close to X11's ones that I included so I did not include both. Feedback appreciated.

[Spike-from-NH]

My feedback is as above: The characters I use often, I add to my custom layouts. The ones I use rarely, I don't mind if they take several keystrokes, provided the keystrokes are defined somewhere. In fact, if a rare character of interest is defined in a file, I'd just as soon copypaste it from the file into the document I'm composing.

And yes, X11 is a lot better-known than Hacker's Keyboard. But you ought not have a policy of limiting Unexpected Compose to only the characters X11 supports.

And you ought to have a way to let me insert any character for which I have looked up the UNICODE. (PS—I can't read Python, but it looks like your Compose will accommodate this.)

[Julow]

And yes, X11 is a lot better-known than Hacker's Keyboard. But you ought not have a policy of limiting Unexpected Compose to only the characters X11 supports.

As I said, Hacker's Keyboard's sequences are almost identical. Is here any sequence you miss ?

And you ought to have a way to let me insert any character for which I have looked up the UNICODE. (PS—I can't read Python, but it looks like your Compose will accommodate this.)

The unicode entry feature is tracked by this issue: #191
The compose key is not idea for this as the state machine would become too large if all the unicode characters are added.
Though it could be a good idea to add support for some common names (for example, all the HTML entities).

[Spike-from-NH]

the state machine would become too large if all the unicode characters are added.

I don't propose that the state machine enumerate every UNICODE character (nor did OP in #191), certainly not by name; only allow Compose plus four hex digits and emit that character.

(The most recent character I added to my own layout was ♪, which I use maybe once a year to identify something as a song lyric; you cover it and Hacker's didn't.)