add conversions for demon* from enGB to enUS

[sebweb3r]

No description provided.

[sebweb3r]

how to proceed with words, that are "wrong" in regular english but used in coding?

[peternewman]

how to proceed with words, that are "wrong" in regular english but used in coding?

As per my comment, I'm not convinced they are, certainly one isn't, so I suspect the others probably shouldn't be either if you're using the Greek sense.

[sebweb3r]

ok. I reduced it to the enGB->enUS

[peternewman]

LGTM. Do you want to add equivalents for daemonize etc? Up to you?

This looks like it exists!
https://linux.die.net/man/1/daemonize

[sebweb3r]

I wanted to add them first, but they are not in aspell. Thus, the tests fail.
I've put them in #1626, which is blocked by the costum dictionary in #1485.

[lurch]

Whilst I agree that daemon can be considered a "proper word", I'm not convinced that daemonize, daemonisation, etc. should ever be considered "proper words", although I accept that they may appear in computing-only contexts?
I.e. in natural-language I think that codespell should suggest daemonize->demonize, and only "allow" daemonize when checking code. But perhaps that's a level of subtlety that codespell doesn't have yet? 🤷

[peternewman]

Whilst I agree that daemon can be considered a "proper word", I'm not convinced that daemonize, daemonisation, etc. should ever be considered "proper words", although I accept that they may appear in computing-only contexts?
I.e. in natural-language I think that codespell should suggest daemonize->demonize, and only "allow" daemonize when checking code. But perhaps that's a level of subtlety that codespell doesn't have yet? shrug

They were already in the code dictionary before @sebweb3r deleted them, which would mean if you chose to enable that dictionary (typically not in code contexts), it would see them as typos and convert them as you suggested. It does mean you can't have a typo which suggests them if they are corrected to something else elsewhere.

[sebweb3r]

There were not in the dictionary. I proposed to add them in this MR (see b3e9127), but after your remarks I deleted the addition. 😄

[lurch]

it would see them as typos and convert them as you suggested. It does mean you can't have a typo which suggests them if they are corrected to something else elsewhere.

I'm afraid you've lost me again 🙃 😆

[sebweb3r]

to simplify it:

You are not allowed to have the same word as a right word in one dictionary and as wrong in a different dictionary (except of enGB_to_enUS).

[peternewman]

You are not allowed to have the same word as a right word in one dictionary and as wrong in a different dictionary (except of enGB_to_enUS).

And the others currently defined here (which does need a comment clarifying direction if any):
https://github.com/codespell-project/codespell/blob/master/codespell_lib/tests/test_dictionary.py#L180-L186

There were not in the dictionary. I proposed to add them in this MR (see b3e9127), but after your remarks I deleted the addition. smile

Sorry, I meant you'd added them to achieve what @lurch was suggesting. I only meant and said that daemon is never a typo, not the other variants. I haven't done any searching, but I suspect @lurch could be correct about daemonize only being valid in computing concepts (and so should be corrected in the code dictionary). Someone just needs to check a proper dictionary to see if they have both spellings.

[lurch]

I can't find it being used outside of a computing context? https://www.google.com/search?q=daemonize+-computer+-process+-command+-program+-library

[peternewman]

I can't find it being used outside of a computing context? https://www.google.com/search?q=daemonize+-computer+-process+-command+-program+-library

https://www.google.com/search?q=daemonize+-computer+-process+-command+-program+-library+-application+-server+-software+-github+-tool+-bitbucket+-ruby+-java+-php+-perl+-node+-js+-c+-c%2B%2B+-django+-rust+-api+-python+-ubuntu+-debian+-redhat+-ceph+-xen

Finds:
https://www.reddit.com/r/AskLiteraryStudies/comments/8o4zvf/what_is_daemonization/
Also:
https://books.google.co.uk/books?id=ybSsZzPgtgEC&pg=PA153&lpg=PA153&dq=daemonize#v=onepage&q=daemonize&f=false

So it's not exclusively a computing term, but quite rare. But this is rather irrelevant given we don't currently have a computing only dictionary, just the opposite.

[lurch]

Wow, you clearly have more google-persistence than me!
Given that you've found it used in a real printed book I'll withdraw my objections. 📘

[peternewman]

@sebweb3r do you want to add back the other dea- to dae- words?

[sebweb3r]

Which additions exactly do you mean?
For the code dictionary I had those:

daemonisation->demonisation, demonization,
daemonise->demonise, demonize,
daemonised->demonised, demonized,
daemonising->demonising, demonizing,
daemonization->demonization, demonisation,
daemonize->demonize, demonise,
daemonized->demonized, demonised,
daemonizing->demonizing, demonising,

But you said, that they are real words, that are rarely used outside of the code context.
I would be fine to add them to the code dictionary again (but it is maybe a false positive)

dea- to dae are alreay in.
daemonis*->daemoniz* don't work, due to aspell

[peternewman]

Sorry I lost track the other stuff had been merged elsewhere.

daemonis*->daemoniz* don't work, due to aspell

Okay. They just aren't in aspell at all, so we'd need to have a second dictionary for non-aspell entries or something, so lets give up on them for now.

I'm happy if @lurch is happy with the latest state too.

[lurch]

The title of the PR needs updating, but otherwise LGTM!