elisp completion patches v6
Main changes since v5: - replace use of notmuch-message-message-insinuate. I'm not sure if we should leave a stub function, or just break people's .emacs, and tell them not to call it anymore. - company is autoloaded, and used, if present. This can be disabled by setting notmuch-message-use-company to nil - notmuch-message-command being nil is now an explicit disabling of of completion - deduplicate=address is hardcoded Things I didn't have time to do - date timeout - make deduplication configurable - make TAB invoke company I'm still not sure about the use of notmuch-message-command as opposed to a seperate variable. This particular setup will be a smooth transition for those that have customized the variable. I'm also not sure if enabling company should turn off non-company based completion. A more conservative default would be to set notmuch-message-use-company to nil. ___ notmuch mailing list notmuch@notmuchmail.org https://notmuchmail.org/mailman/listinfo/notmuch
Re: elisp completion patches v6
David Bremner writes: > Main changes since v5: > >- replace use of notmuch-message-message-insinuate. I'm not sure if > we should leave a stub function, or just break people's .emacs, > and tell them not to call it anymore. > >- company is autoloaded, and used, if present. This can be disabled > by setting notmuch-message-use-company to nil > >- notmuch-message-command being nil is now an explicit disabling of > of completion > >- deduplicate=address is hardcoded One more important change is that this no longer special-cases notmuch-company with respect to byte-compilation. It compiles fine without company-mode (although I noticed that I missed one variable declaration). On the other hand I didn't test this with emacs23. I wouldn't expect it to work, but hopefully it still compiles. d ___ notmuch mailing list notmuch@notmuchmail.org https://notmuchmail.org/mailman/listinfo/notmuch
Re: elisp completion patches v6
Hi David, 2015ko urriak 24an, David Bremner-ek idatzi zuen: > > Main changes since v5: > >- replace use of notmuch-message-message-insinuate. I'm not sure if > we should leave a stub function, or just break people's .emacs, > and tell them not to call it anymore. Maybe you could use a combination of ‘make-obsolete’ (for byte-compiler warnings) and ‘display-warning’ (for runtime)? I understand the desire to phase the function out eventually rather than have an eternal stub, but intentionally breaking .emacs is a really drastic step. -- Aaron Ecay ___ notmuch mailing list notmuch@notmuchmail.org https://notmuchmail.org/mailman/listinfo/notmuch