First, the ipa.scm I suggested had one error. I'll provide the fix.
Secondly, I'm not sure how to do this, but whenever a character that is part of
a sequence is used but followed by something not in any sequence, the characters
get mixed up.
* type 'd'
* 'd' gets highlighted because it can be part of the 'd_<' sequence
* type 'a'
* the string is now "ad" instead of the expected "da"
The same problem occurs with space after a highlighted character.
It would be nice if a charcter not in a sequence or a space could just commit
what is in preedit and get printed too.
Created attachment 1834 [details]
ipa.scm with proper character for F
((("F" )) ("m"))
((("F" )) ("ɱ"))
Thanks for your patch. I applied part of your patch in revision 570.
What I didn't understand is the reason for existence of ipa-key-press-handler.
(It doesn't works with latest code because we introduced lazy-loading feature,
you can't call hangul-* functions from ipa.scm.)
If the reason is to fix your second problem, I'll provide another fix. If not,
please tell me the reason.
oh I'm really sorry I got mixed up between the fixed ipa.scm and the one i was
testing things with. I'm really sorry about that.
But,yes it would be really nice to fix the preedit being cancelled. I can't
figured it out right now, so if you have the knowledge and time I'd really
Another thing, we should probably change the label for the input method to
"International Phonetic Alphabet (X-SAMPA)" so the user has an idea of what the
sequences are. We could even rename the file ipa-xsampa.scm or something similar.
I am currently thinking of implementing another IPA input method, one developed
by SIL (Summer Institute of Linguistics),
The sequences are more intuitive than the X-SAMPA sequences and many linguists
already use the SIL IPA Keyman layout on windows.
That would also need a conservation of the preedit.
I only changed ipa.scm's description about the input method, because unfortunately
I don't have enough time to fix the problem until next Tuesday...
BTW, do you think we should change our input method name from ipa to x-sampa or
for the name of the method i would suggest ipa-x-sampa or ipa-xsampa, so later
on once i make a SIL based IPA input method we can call that one ipa-sil.
Ok, I'll change the input method name as 'ipa-x-sampa' later.
I tried, but I can't changed input method name. It seems awkward problem, I need
more time to investigate.
Mark as blocking bug #2304. See following page.
I fixed that "ad" instead of the expected "da" in revision 716, but space key
still erases preedit strings. Fixing this problem is a bit difficult, so
I'll release 0.4.6 final without fix. Sorry for inconvenience.
Thanks a lot for the fix I'll download the updated file and test as soon as
I finally got around to testing it.
As you mentioned the space erases the preedit string.
But there's another bug now, a major one. Keys that do not deal with preedit
don't do anything. For example pressing the 'm' key does not print anything.
This does not happen with shift combinations, for example Shift+'m' works as it
should, passing 'ʍ'.
Again, thanks for you work.
Created attachment 2189 [details] [review]
patch to change ipa to ipa-x-sampa in test/test-im.scm
Here's a patch to follow the change of ipa to ipa-x-sampa in the test file
Sorry for long time absence, yesterday I committed your patch. Thanks a lot!
It seems to already fixed.