use ASCII '>=' to describe version dependency

[MichaelChirico]

Closes #6931

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.59%. Comparing base (5a955df) to head (a3e01c9).
Report is 5 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #6945   +/-   ##
=======================================
  Coverage   98.59%   98.59%           
=======================================
  Files          79       79           
  Lines       14684    14684           
=======================================
  Hits        14478    14478           
  Misses        206      206           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[aitap]

Grrr. Somehow my manual testing with an empty package does something different from testing with data.table. a3e01c9 was actually tested with data.table.

I suspect that pkgdown is not enabling the Markdown extensions that would let pandoc --mathml recognise \(...\) (that pkgdown produces when it converts Rd to Markdown) as equations:

15190 execve("/usr/bin/pandoc", ["/usr/bin/pandoc", "+RTS", "-K512m", "-RTS", "/tmp/RtmpH7omXJ/file36ce1a2a663", "--to", "html4", "--from", "markdown+gfm_auto_identifiers-citations+emoji+autolink_bare_uris", "--output", "/tmp/RtmpH7omXJ/file36ce2a01fd68", "-t", "html4", "--indented-code-classes=R", "--section-divs", "--wrap=none", "--mathml"], [ <environment omitted> ] = 0

# using exactly the same arguments as pkgdown does
echo '\( 2 \ge 1 \)' | "/usr/bin/pandoc" "+RTS" "-K512m" "-RTS" "--to" "html4" "--from" "markdown+gfm_auto_identifiers-citations+emoji+autolink_bare_uris" "-t" "html4" "--indented-code-classes=R" "--section-divs" "--wrap=none" "--mathml"
# <p>( 2  )</p>

# using dollar notation instead of \(...\) that pkgdown generates from \eqn{} tags
echo '$2 \ge 1$' | "/usr/bin/pandoc" "+RTS" "-K512m" "-RTS" "--to" "html4" "--from" "markdown+gfm_auto_identifiers-citations+emoji+autolink_bare_uris" "-t" "html4" "--indented-code-classes=R" "--section-divs" "--wrap=none" "--mathml"
# <p><math display="inline" xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mn>2</mn><mo>≥</mo><mn>1</mn></mrow><annotation encoding="application/x-tex">2 \ge 1</annotation></semantics></math></p>

# enabling the additional syntax extensions
echo '\( 2 \ge 1 \)' | "/usr/bin/pandoc" "+RTS" "-K512m" "-RTS" "--to" "html4" "--from" "markdown+gfm_auto_identifiers-citations+emoji+autolink_bare_uris+tex_math_single_backslash" "-t" "html4" "--indented-code-classes=R" "--section-divs" "--wrap=none" "--mathml"
# <p><math display="inline" xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mn>2</mn><mo>≥</mo><mn>1</mn></mrow><annotation encoding="application/x-tex">2 \ge 1</annotation></semantics></math></p>

As a result, all uses of \eqn are broken. The good news is that there's only two of them and they are special-cased with \ifelse{html}; the bad news is that more are brewing in jangorecki#4.

[MichaelChirico]

FWIW, I don't see any other >=/<= usages we should be fixing. The only other I see (1.17.0) is here:

data.table/man/setattr.Rd

Line 49 in 5a955df

colnames(DF)[1] <- "A" # 4 shallow copies (R >= 3.1, was 4 deep copies before)

Ideally we'd also have a new code-quality check for these, it's somewhat annoying since grep won't work easily, we need to exclude \examples{} first, so my first instinct is to leave it as a follow-up.

[ben-schwen]

LGTM