nsis
nsis copied to clipboard
Fix typos found by codespell
Some typos are in user-visible strings and documentation, others are in code comments.
In the long term, I recommend fixing all typos, including those in code comments, because tools like codespell do not make a difference between code comments and documentation and user-visible strings.
I have left out typos in Contrib
, they should be fixed upstream.
Not sure about _miscrosoft_docs_are_wrong
→ _microsoft_docs_are_wrong
.
I believe it's OK because it's a private variable.
I have left out typos in
Contrib
, they should be fixed upstream
We own most of Contrib. If there is anything upstream I'll deal with it.
Fixed typos in Contrib
too.
I have left out Language files
by the way.
There may be a few starting comments in English to fix, for example appropiate
→ appropriate
:
https://github.com/kichik/nsis/blob/268379229b0c1ef00603779890a28d0345ab0240/Contrib/Language%20files/Romanian.nsh#L8
Should I fix these comments?
Edit: That's actually the only typo in there; I have fixed it.
And I just noticed part of the initial fixes, under Docs/src/halibut
, must actually be fixed upstream.
Edit: I've sent a mail to the upstream maintainer.
Edit: Ongoing discussion with the upstream maintainer: are both bulletted
and bulleted
(as in the Oxford dictionary) correct in UK English?
are both
bulletted
andbulleted
(as in the Oxford dictionary) correct in UK English?
Bulletted only has 35k hits on Google vs millions for Bulleted. I'll take the fix.
I see you have merged this change into the Subversion trunk. Thanks! Would it be worth adding a configuration file .codespellrc
to silence false positives and simplify future maintenance?
I suppose that depends on how many false-positives you had to deal with and what the maintenance overhead is...
I am afraid nsis has many false positives:
$ codespell | wc -l
954
$
Most of them are in translation files that should be skipped. A .codespellrc
file could drastically reduce that number by skipping these files. For example:
$ codespell --skip 'Language files' | wc -l
226
$
The can further reduce that number by ignoring some words:
$ codespell --skip 'Language files' --ignore-words-list parms | wc -l
133
$
The benefit of a .codespllrc
is that developers get consistent results when running codespell
locally.
All fixes have been merged, thank you. I have added a .codespellrc
file and fixed the last typos. Do you think it's worth merging?