lintcmd: don't print \' in -list-checks and -explain

[arp242]

All the single quotes are written as \' in the descriptions. I assume that's for a good reason, but -list-checks would print:

% staticcheck -list-checks | grep "'"
S1003 Replace call to \'strings.Index\' with \'strings.Contains\'
S1004 Replace call to \'bytes.Compare\' with \'bytes.Equal\'
S1011 Use a single \'append\' to concatenate two slices
S1012 Replace \'time.Now().Sub(x)\' with \'time.Since(x)\'
...

% staticcheck -explain S1003
Replace call to \'strings.Index\' with \'strings.Contains\'

So fix that:

% staticcheck -list-checks | grep "'"
S1003 Replace call to 'strings.Index' with 'strings.Contains'
S1004 Replace call to 'bytes.Compare' with 'bytes.Equal'
S1011 Use a single 'append' to concatenate two slices
S1012 Replace 'time.Now().Sub(x)' with 'time.Since(x)'
...

% staticcheck -explain S1003
Replace call to 'strings.Index' with 'strings.Contains'

As an added bonus, make -explain s1003 (lower case) work. Seemed too small of a change to make a new PR for.

[dominikh]

FWIW, \' is how we encode backticks (for Markdown) in raw string literals.

[arp242]

As an added bonus, make -explain s1003 (lower case) work. Seemed too small of a change to make a new PR for.

I backed out this change, as per #1546. As I mentioned there, I thought all CLI flags were case-insensitive but they're not.

I also added a commit to print fewer extraneous newlines in the Before/After blocks; see the commit message.

[dominikh]

I believe this is part of a broader regression introduced by 8643e6eb787e7d3060739836c6ceea312c91ca52 where we broke the conversion of documentation to plain text and Markdown variants (cf. the use of lint.RawDocumentation and lint.Markdownify.) I'll look into fixing this myself.

[arp242]

Yeah, I also discovered that on Sunday when I wanted to build the website. Fixing the website/cmd/generate_checks program wasn't too much work, but Hugo gives some error on shortcodes, and gave up at that.

[dominikh]

Yeah, I'll have to fix website soon. Hugo has been "fun" when it comes to backwards compatibility.

[arp242]

Hugo has been "fun" when it comes to backwards compatibility.

I actually tried older versions of Hugo too, the current release from when the website was first added, and a release from a few months before that, but all gave the same error as the latest version:

% hugo
Start building sites …
hugo v0.126.2+extended linux/amd64 BuildDate=unknown

Total in 13 ms
Error: error building site: process: readAndProcessContent:
"/home/martin/src/go-tools/website/content/contact.md:9:1":
failed to extract shortcode: template for shortcode "blocks/section" not found

That was with v0.88.1 and v0.80.0, according to my shell history.

According to the Hugo docs everything looks fine too, so idk. At this point I gave up.

FYI.

[dominikh]

Make sure you've cloned the repository recursively; the website theme is pulled in via a git submodule.

[arp242]

Ah right, that furthers things! Got fresh errors though.

I just wanted to look at some of the documentation issues in the weekend, and make sure whatever I wrote appears well on the website. Well, whatever 🤷

The amount of complexity involved here for rendering a few basic HTML pages is slightly ridiculous to be honest.