Add typos of align, allocate, block, docker, given

[bl-ue]

With -ing and -ed form

[bl-ue]

@lurch done, PTAL.

[lurch]

You need to fix the sorting of the entries to make the CI tests happy 😉

[bl-ue]

Fixed CI issues. @lurch @larsoner PTAL.

[lurch]

Apart from the small query about dockerd appearing in the main dictionary, this LGTM now 👍

[bl-ue]

@larson #1815 (comment)

[lurch]

(@bl-ue But just to clarify - my query may be totally invalid... I can never remember how all the different dictionaries are supposed to work together 😉 )

[bl-ue]

@lurch what about macosx:

codespell/codespell_lib/data/dictionary.txt

Line 17455 in d133bef

maxosx->macosx

and Mozilla:

codespell/codespell_lib/data/dictionary.txt

Line 17459 in d133bef

mazilla->Mozilla

which are in the main dictionary?

[lurch]

It's hard to imagine a normal word that maxosx or mazilla would correct to in non-code-usage.
But IMHO it's entirely plausible that dockerd might be considered a typo for docker or docked in regular language usage? 🤷

[bl-ue]

Good point. So instead I'll add dockerd as a possible replacment for docked, etc. WDYT?

[lurch]

I've got no absolutely no idea what the "correct" action is in this situation. ping @larsoner @peternewman @sebweb3r

[sebweb3r]

There is (from my understanding) no better fitting dictionary.
One solution could be
dcokerd->dockerd, docked, docker,
in the main dictionary and
dockerd->docker, docked,
in the code dictionary (contains all checks that would be annoying with code)
However, the main dictionary is not allowed to correct to something, that is a mistake
in the code dictionary (i don't know why).

"Correct" solutions can only be provided by the other two mentioned.

[peternewman]

It's hard to imagine a normal word that maxosx or mazilla would correct to in non-code-usage.
But IMHO it's entirely plausible that dockerd might be considered a typo for docker or docked in regular language usage?

Again it doesn't matter; ALL we care about is that the typo isn't a valid word in some context, then it gets shunted out to another dictionary (see also rare e.g. chancel->chance, running that across a religious document would cause chaos, they aren't currently a big enough market so we just pop it in rare and hopefully people are careful or only turn on rare if they are). We wouldn't put e.g. macos->macros in the main dictionary as it would make a mess, it would go in code so only people not writing "code" would get that correction with hopefully no unintended consequences.

There is (from my understanding) no better fitting dictionary.
One solution could be
dcokerd->dockerd, docked, docker,
in the main dictionary and
dockerd->docker, docked,
in the code dictionary (contains all checks that would be annoying with code)

I think the only downside of this is we don't currently chain dictionaries (see also e.g. #1672), although as long as there is more than one correction in the main one (i,e. it duplicates the corrections from either side of the code one) it shouldn't have any negative side effects.

However, the main dictionary is not allowed to correct to something, that is a mistake
in the code dictionary (i don't know why).

We just need to add this commit back in (or make an equivalent edit):
fc203b4

The list of correcting to a typo in another dictionary is kept as small as practically possible because you shouldn't generally be correcting to a word that is itself a typo in another dictionary, instead you should correct directly to the final correction. The exceptions are allowed where required. The obvious case being colur->colour in the main dictionary and colour->color in GB to US is obviously acceptable and intended.

If anyone can word this in a more friendly way and put it in a readme would be great as it and similar variations do seem to keep coming up. Or better descriptions for the various dictionaries. Or do we just need a note about it generally only being the typo side of the dictionary we care about (apart from GB to US).

[peternewman]

A few more comments

[bl-ue]

@lurch @sebweb3r @peternewman done @ 307fe9e, PTAL.

[bl-ue]

@peternewman review?

[peternewman]

LGTM thanks @bl-ue