Change single-word uses of the D macro in documentation with backticks

[andralex]

This replaces uses of $(D ...) with backticks for single-word instances. This makes the doc easier on the eyes and easier to edit. To effect this I ran:

sed -i '' -e 's/\$(D \([^ )]*\))/`\1`/g' std/**/*.d

[andralex]

cc @wilzbach can we also make this a standard check for future code?

[aG0aep6G]

Bad substitutions here.

[andralex]

Changed the sed expression to:

sed -i '' -e 's/\$(D \([^ ()]*\))/`\1`/g' std/**/*.d

[andralex]

What the heck did I do? Ugh...

[dlang-bot]

Thanks for your pull request, @andralex!

Bugzilla references

Your PR doesn't reference any Bugzilla issue.

If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog.

[dlang-bot]

Thanks for your pull request, @andralex!

Bugzilla references

Your PR doesn't reference any Bugzilla issue.

If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog.

[andralex]

OK here we go...

[andralex]

This should be merged soon, it's a bonanza of rebase conflicts...

[PetarKirov]

Looks good, as far as I could see. I'm in favor of merging and fixing possible issues (if any) in follow-up PR(s).

[andralex]

cool thx

[PetarKirov]

@andralex the ddoc build passed, so you can verify if there's anything unexpected: http://dtest.dlang.io/results/d970593fffa023018370dde1852998cd6bd83ff2/f8d3d9b2ade3031dc671481be922425e8577263a/. (I'll be able to look at it in 30-40 mins)

[andralex]

Seeing artifacts that shouldn't be there, removed auto-merge for now to investigate.

[CyberShadow]

Here is the previous attempt, may contain useful discussion: #2877

[andralex]

@CyberShadow thanks. This particular PR is more conservative so it's mainly exposed to the problem of double highlighting of focal words (those that people get rid of by using a leading underscore).

That problem is benign - it applies an anchor and a span style twice, so it increases file size but does not affect functionality. I suggest we pull this and create an issue for the backticks feature.

[PetarKirov]

I think this is one of the culprits - see here:

-  <a href="#.Base64Impl.Encoder.empty"><span class="d_inlinecode donthyphenate notranslate">empty</span></a> returns <span class="d_inlinecode donthyphenate notranslate"><code class="ddoc_keyword">true</code></span>.</div></div>
+  <a href="#.Base64Impl.Encoder.empty">`empty`</a> returns <span class="d_inlinecode donthyphenate notranslate">true</span>.</div></div>

[andralex]

OK @CyberShadow warned that there are additional _ inserted in the generated HTML, which makes this PR not viable without fixing the ddoc bug.

[andralex]

@ZombineDev good insight, I changed the macro to see how things go. BTW we already have that macro, right? No need for it.

[wilzbach]

Here is the previous attempt, may contain useful discussion: #2877

FWIW I also tried this before and failed: 😢
#5183

which makes this PR not viable without fixing the ddoc bug.

Well you do see all places where a _ needs to be inserted on DAutoTest.

@andralex let me know if I should adopt this PR and bring it to the finish line. After all, this is now at least the third attempt...

[andralex]

@wilzbach yah, that would be nice but do it only if technically easy - it's not high impact. Thx!

[quickfur]

We really really need to kill that autoexpansion habit of ddoc that leads to _ proliferation. It's clearly causing far more trouble than it's worth, and for far too long already.

[wilzbach]

@wilzbach yah, that would be nice but do it only if technically easy - it's not high impact. Thx!

Yeah, it's pretty trivial to do, but I just did it for std.algorithm as a test:
#5970

[wilzbach]

Yeah, it's pretty trivial to do, but I just did it for std.algorithm as a test:
#5970

#6391 (for all Phobos)

We really really need to kill that autoexpansion habit of ddoc that leads to _ proliferation. It's clearly causing far more trouble than it's worth, and for far too long already.

dlang/dlang.org#2307 (it never has been closer thanks to @quickfur's work on adding the opt-out macro)