could you please consider fix for issue with fly checking non ascii clojure code in Emacs. The code is likely to be converted to iso-8859-1, translate non-ascii chars to spaces and thus greatly confusing the linter. It fixes #16.
Although Emacs is primarily setup to to communicate in ASCII to and from the subprocesses it forks on both MS-Windows and GNU/Linux, the issue is more apparent on MS-Windows. This is because Linux is most likely to have set the locale to a UTF-8 codebase with one of the relevant environment variables, such as LANG=C.UTF-8, which Emacs honours and uses as the codepage encoding to communicate data with its subprocesses.
The solution is here a very simple one. Just set the comm codepage Emacs uses to communicate to the clj-kondo executable to UTF-8.
Since this is a tricky business to test and maintain the integrity of the setup, I took the opportunity to add some basic tests (hand picked from the clj-kondo corpus) for each of the available checkers, as well as integrate them with the GitHub CI, using the almighty Elisp Development Environment. I've also included a test to check the utf-8 condition. I hope you find it useful (I can of course revert if you think is too much).
All tests pass across Ubuntu, MS-Windows and MacOS.
Hi,
could you please consider fix for issue with fly checking non ascii clojure code in Emacs. The code is likely to be converted to iso-8859-1, translate non-ascii chars to spaces and thus greatly confusing the linter. It fixes #16.
Although
Emacs
is primarily setup to to communicate in ASCII to and from the subprocesses it forks on both MS-Windows and GNU/Linux, the issue is more apparent on MS-Windows. This is because Linux is most likely to have set the locale to a UTF-8 codebase with one of the relevant environment variables, such asLANG=C.UTF-8
, which Emacs honours and uses as the codepage encoding to communicate data with its subprocesses.The solution is here a very simple one. Just set the comm codepage Emacs uses to communicate to the
clj-kondo
executable to UTF-8.Since this is a tricky business to test and maintain the integrity of the setup, I took the opportunity to add some basic tests (hand picked from the clj-kondo corpus) for each of the available checkers, as well as integrate them with the GitHub CI, using the almighty Elisp Development Environment. I've also included a test to check the utf-8 condition. I hope you find it useful (I can of course revert if you think is too much).
All tests pass across Ubuntu, MS-Windows and MacOS.
Thanks