[v5.2.0 now out] Conkey keyboard layout

[akam chinjir]

So for my own purposes I've learned (I think!) how to make and install an input method on OSX. I could be missing something, but I don't think it should be too hard to convert Conkey, if I had the basic mapping in a format I understood. (Conkey.klc looks like it wouldn't take too much explanation?)

[bradrn]

So for my own purposes I've learned (I think!) how to make and install an input method on OSX. I could be missing something, but I don't think it should be too hard to convert Conkey, if I had the basic mapping in a format I understood. (Conkey.klc looks like it wouldn't take too much explanation?)

Well, if you know Haskell, I’ve already written half the code for a converter to OSX — MSKLC.Parser.parse will parse an Microsoft Keyboard Layout Creator file (like Conkey.klc) into a Keyboard data structure, so you only need to write a converter from a Keyboard to whatever the OSX keyboard file format is. (Alternately, if you can give me an example of the OSX keyboard format, I could try writing a converter myself.)

[akam chinjir]

My Haskell is very weak But I'll have a look.

[bradrn]

My Haskell is very weak But I'll have a look.

Thanks! I’d love to get Conkey working on OSX as well.

[bradrn]

Version 3.1.0 is now out! Download it at the usual place.

The highlight of this release: Conkey is now available on Mac OSX! You can download it at the link above. Special thanks for this goes to akam chinjir, who wrote the program to convert Conkey to a Mac OSX keyboard layout.

Other highlights:

• Due to a lack of testing on my part, the Linux (M17N) version previously had several problems, rendering it effectively unusable; these have now been fixed.
• A new diacritic, double grave ⟨◌̏⟩, has been added.
• More click letters: ⟨ʇʖʗʞ⟩ have been added. (I know they’re archaic, but occasionally they become useful if the usual pipe symbols are hard to read.)
• Three new mathematical symbols: there exists ∃, proportional-to ∝, similar-to ∼ are all now available.

For full details, consult the release notes, at the link above.

[bradrn]

Version 4.0.0 is now out! Get it here. As usual, consult the release notes for a full description, but major changes include:

• Added various IPA letters with uppercases: ⟨ɐⱯ ɜꞫ ɪꞮ ʝꞲ ɒⱰ ɥꞍ⟩
• Added fractions: ⟨½ ⅓ ¼ ⅕ ⅙ ⅛ ⅒ ⅔ ¾ ⅟⟩ as well as fraction slash ⟨⁄⟩ (which lets you create any fraction if your font cooperates: e.g. 12345⁄67890.)
• Lots of other symbols for maths: ⟨ℜ ℑ ∂ ∇ ⅈ ⅉ⟩, and the variant Greek letters ⟨ϵ ϑ ϰ ϖ ϱ ϕ⟩
• Tironian Et ⟨⁊⟩

[bradrn]

Version 4.1.0 is now out! Get it here. Release highlights include:

• Ordinal indicators ª and º
• Mathematical and physical symbols ⟨∞ ℓ ⊗ ☉ ℵ⟩
• Small capitals ⟨ʙ ᴄ ʜ ᴋ ɴ s ᴜ⟩ and small capital Q (which has poor font support)
• Combining bridge below diacritic
• Upstep and downstep arrows

And many more!

[WarpedWartWars]

I spotted this and was curious, so I checked it out and got the latest version, 4.1.0, for Windows, and it works! I tried it by doing G-' s, and I got ś! I now have that "Yes! It works!!!" feeling that I also got yesterday upon making a very early part of clox from craftinginterpreters and it working.

[Travis B.]

I spotted this and was curious, so I checked it out and got the latest version, 4.1.0, for Windows, and it works! I tried it by doing G-' s, and I got ś! I now have that "Yes! It works!!!" feeling that I also got yesterday upon making a very early part of clox from craftinginterpreters and it working.

That is how it was when I first got zeptoforth really working, without crashing or anything, for the first time.

[bradrn]

I spotted this and was curious, so I checked it out and got the latest version, 4.1.0, for Windows, and it works! I tried it by doing G-' s, and I got ś! I now have that "Yes! It works!!!" feeling that I also got yesterday upon making a very early part of clox from craftinginterpreters and it working.

Glad to hear it ‘just works’ for you! That was the intention. (And I’m glad to know you got clox working, too… I’ve been meaning to go through that book one of those days.)

[WarpedWartWars]

I spotted this and was curious, so I checked it out and got the latest version, 4.1.0, for Windows, and it works! I tried it by doing G-' s, and I got ś! I now have that "Yes! It works!!!" feeling that I also got yesterday upon making a very early part of clox from craftinginterpreters and it working.

Glad to hear it ‘just works’ for you! That was the intention. (And I’m glad to know you got clox working, too… I’ve been meaning to go through that book one of those days.)

I currently have only gotten through the chapters before chapter 20, Hash Tables, but it does work so far.

My main use for Conkey will probably be typing IPA and various languages that use extended Latin alphabets. The first is why I was just looking all over the docs for "ɹ".

Anyway, doing some testing:

More: show

kɔŋki oɹ kɔnki?

[WarpedWartWars]

I found a bug in the docs--page 5, describing "acceptable inputs", there's a red superscript "1" indicating a footnote, but said footnote doesn't exist!

Edit: In documentation.tex, there's no \footnotetext for it:

Code: Select all

\item `Acceptable inputs' contains a list of input letters which can be used with this dead key.\footnotemark
  For clarity, every list of inputs in this column has been separated into lower- and uppercase letters (sometimes on two lines).

  This column is required because not all letter+diacritic combinations are present in Unicode;
    the missing combination thus cannot be typed in the same way.
  For instance, \out{a̋} cannot be typed using \key|G-: a|, since that particular combination is not in Unicode.

(end of the first line is the \footnotemark but there's no \footnotetext)

(the "a̋" isn't combined in code tags...)

[bradrn]

I found a bug in the docs--page 5, describing "acceptable inputs", there's a red superscript "1" indicating a footnote, but said footnote doesn't exist!

Oh, whoops, so it does! Fixed now.

(How do I do inline code on phpBB?)

Use [tt][/tt].

[WarpedWartWars]

Oh, whoops, so it does! Fixed now.

Documentation.pdf doesn't seem to be fixed, though.

[bradrn]

Oh, whoops, so it does! Fixed now.

Documentation.pdf doesn't seem to be fixed, though.

I figured it was too small of a change to bother with recompiling and modifying the release, though of course I can still do that if you want. (Besides, I have plans to move the documentation over to HTML soon, so it would become a bit irrelevant in any case.)

[WarpedWartWars]

I figured it was too small of a change to bother with recompiling and modifying the release, though of course I can still do that if you want. (Besides, I have plans to move the documentation over to HTML soon, so it would become a bit irrelevant in any case.)

Ah.

I can wait till you've ported it to HTML.

[Travis B.]

I figured it was too small of a change to bother with recompiling and modifying the release, though of course I can still do that if you want. (Besides, I have plans to move the documentation over to HTML soon, so it would become a bit irrelevant in any case.)

Ah.

I can wait till you've ported it to HTML.

I can't. LaTeX is so superior to HTML it's not funny.

[bradrn]

I figured it was too small of a change to bother with recompiling and modifying the release, though of course I can still do that if you want. (Besides, I have plans to move the documentation over to HTML soon, so it would become a bit irrelevant in any case.)

Ah.

I can wait till you've ported it to HTML.

I can't. LaTeX is so superior to HTML it's not funny.

Oh, usually I’d totally agree. But LaTeX has one issue here: its output is paginated, which makes it hard to arrange lots of huge tables as is required for Conkey’s documentation. HTML is a lot better at this specific usecase. Additionally, documentation these days tends to be hosted online for convenience, which makes HTML a natural choice.

[bradrn]

(NB. moving discussion to this thread where it’s more on-topic)

(Specifically I can't get my machine to recognize AltGr without AltGr being treated as a regular Alt key by something else, which causes a load of problems in practice, and when it doesn't treat AltGr as Alt it either ignores it or it doesn't handle dead keys right, e.g. it inserts a '-' without actually applying a diacritic to the following letter.)

Oh dear… none of this sounds good. But from your description it sounds like it’s a problem with ibus itself, rather than with this specific keyboard. Could you perhaps try out some of the other available keyboards and tell me whether these problems persist with those or not?

[Richard W]

(NB. moving discussion to this thread where it’s more on-topic)

(Specifically I can't get my machine to recognize AltGr without AltGr being treated as a regular Alt key by something else, which causes a load of problems in practice, and when it doesn't treat AltGr as Alt it either ignores it or it doesn't handle dead keys right, e.g. it inserts a '-' without actually applying a diacritic to the following letter.)

Oh dear… none of this sounds good. But from your description it sounds like it’s a problem with ibus itself, rather than with this specific keyboard. Could you perhaps try out some of the other available keyboards and tell me whether these problems persist with those or not?

There is an imperfect work round, which I only confirmed when I spoilt my physical Thai keyboard by spilling coffee on it and messing up the space key. I had to dash out and buy a new USB keyboard, which was therefore a UK keyboard. Now, on Ubuntu 20.04 (Focal) and 22.04 (Jammy), on the UK keyboard Alt-Gr is a modifier that generally gives a different result to the plain key. Consequently, if I change from a UK keyboard to an M17n keyboard under ibus, Alt-Gr will take effect, whereas if I change to the same M17n keyboard from a US keyboard (which is the Latin keyboard engraved on a Thai keyboard), the Alt-Gr is simply ignored.

The point about the US and UK keyboards is that they are defined using the basic X protocols, and the UK keyboard contains the extra bit to map the Alt-Gr key differently to the Alt key. When one uses an M17n keyboard under iBus, the incoming keyboard is mapped by the X keyboard protocol in effect, and then this character is broken into key pressed and modifier keys - a truly bizarre way round. A decision has now been made to let the previous X-keyboard to affect the key mappings - this makes sense for mnemonic keyboards, like a number of XSAMPA keyboards.

Fcitx worked differently - the modifier keys were forwarded, except for AltGr! Fortunately I was able to hack Fcitx to get round that bug / malevolent feature. When I jumped from Xenial to Focal, I tried iBus and found it was largely working. I'd already resorted to systematically using combinations with '!' as alternatives to Alt-Gr; I have to do that for the Emacs versions of the keyboards.

Ideally, Travis wants a US keyboard with a mapping for Alt-Gr modifications of every significant keypress. That then has to be inserted into the set of X keyboard mappings, which is less straightforward than it should be. (If pressed, I can dig up my notes on how to do it. I have some backup Xkeyboard mappings that I haven't reenabled for a while.)