'Error Retrieving Help' on rcmdcheck::rcmdcheck_process

[HenningLorenzen-ext-bayer]

System details

RStudio Edition : Desktop
RStudio Version : 2022.12.0+354 Elsbeth Geranium (desktop)
OS Version      : Windows 10 x64 (build 19044)
R Version       : R version 4.2.2 Patched (2023-01-10 r83603 ucrt)

Steps to reproduce the problem

Type rcmdcheck::rcmdcheck_ into R console

Describe the problem in detail

Error window

Error Retrieving Help
(TypeError) : Cannot read properties of null (reading 'rows')

Describe the behavior you expected

• No error window which keeps popping up and interferes with typing code.

• Nice to have: help on rcmdcheck_process R6 class.

• I have read the guide for submitting good bug reports.

• I have installed the latest version of RStudio, and confirmed that the issue still persists.

• If I am reporting an RStudio crash, I have included a diagnostics report.

• I have done my best to include a minimal, self-contained set of instructions for consistently reproducing the issue.

[HenningLorenzen-ext-bayer]

─ Session info ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
 setting  value
 version  R version 4.2.2 Patched (2023-01-10 r83603 ucrt)
 os       Windows 10 x64 (build 19044)
 system   x86_64, mingw32
 ui       RStudio
 language (EN)
 collate  German_Germany.utf8
 ctype    German_Germany.utf8
 tz       Europe/Berlin
 date     2023-01-13
 rstudio  2022.12.0+354 Elsbeth Geranium (desktop)
 pandoc   NA

─ Packages ───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
 package     * version date (UTC) lib source
 callr         3.7.3   2022-11-02 [1] RSPM (R 4.2.0)
 cli           3.6.0   2023-01-09 [1] RSPM
 crayon        1.5.2   2022-09-29 [1] RSPM (R 4.2.0)
 curl          4.3.3   2022-10-06 [1] RSPM (R 4.2.0)
 desc          1.4.2   2022-09-08 [1] RSPM (R 4.2.0)
 digest        0.6.31  2022-12-11 [1] RSPM
 pkgbuild      1.4.0   2022-11-27 [1] RSPM (R 4.2.0)
 prettyunits   1.1.1   2020-01-24 [1] RSPM (R 4.2.0)
 processx      3.8.0   2022-10-26 [1] RSPM (R 4.2.0)
 ps            1.7.2   2022-10-26 [1] RSPM (R 4.2.0)
 R6            2.5.1   2021-08-19 [1] RSPM (R 4.2.0)
 rcmdcheck     1.4.0   2021-09-27 [1] RSPM
 rprojroot     2.0.3   2022-04-02 [1] RSPM (R 4.2.0)
 rstudioapi    0.14    2022-08-22 [1] RSPM
 sessioninfo   1.2.2   2021-12-06 [1] RSPM
 withr         2.5.0   2022-03-03 [1] RSPM (R 4.2.0)
 xopen         1.0.0   2018-09-17 [1] RSPM

[dbrookeUAB]

Same problem. It happens while using autocomplete to preview package functions (e.g. ggplot2::<tab>). I uploaded an example gif that shows what it looks like. It's only for some packages, and it's only triggered when a specific function is highlighted.

[HenningLorenzen-ext-bayer]

[romainfrancois]

It looks like rcmdcheck:: rcmdcheck _process is not a function but rather an R6 class, and so the Arguments section looks different than what we'd typically get from a function:

https://github.com/r-lib/rcmdcheck/blob/a38200c997f58478c8d3e2c4d2dd5e6244dabcba/R/background.R#L28-L40

so I think the play is to make HelpInfo.java more forgiving here:

rstudio/src/gwt/src/org/rstudio/studio/client/workbench/views/help/model/HelpInfo.java

Line 90 in df7efaa

parseArguments(args, heading);

parseArguments() assumes that there is a <table> :

rstudio/src/gwt/src/org/rstudio/studio/client/workbench/views/help/model/HelpInfo.java

Lines 172 to 198 in df7efaa

	private void parseArguments(HashMap<String, String> args,
	Element heading)
	{
	Element table = findArgumentTable(heading);
	assert table != null : "Unexpected help format, no argblock table found";
	TableElement t = (TableElement) table;
	NodeList<TableRowElement> rows = t.getRows();
	for (int i = 0; i < rows.getLength(); i++)
	{
	TableRowElement row = rows.getItem(i);
	NodeList<TableCellElement> cells = row.getCells();
	TableCellElement argNameCell = cells.getItem(0);
	TableCellElement argValueCell = cells.getItem(1);
	String argNameText = argNameCell.getInnerText();
	String argValueHtml = argValueCell.getInnerHTML();
	// argNameCell may be multiple comma-delimited arguments;
	// split them up if necessary (duplicate the help across args)
	String[] argNameTextSplat = argNameText.split("\\s*,\\s*");
	for (int j = 0; j < argNameTextSplat.length; j++)
	args.put(argNameTextSplat[j], argValueHtml);
	}
	}

[romainfrancois]

In the case of limma::anova.<...> method, the Arguments section is also manually generated with a \section{Arguments} rather than a \arguments https://github.com/cran/limma/blob/1b51b3a5295f9feedb72a0df97e9c3e236592464/man/anova-method.Rd#L13-L20
and this goes against what parseArguments() expects to find.

\section{Arguments}{
\describe{
  \item{\code{object}}{object of class \code{MAList}. Missing values in the M-values are not allowed.}
  \item{\code{design}}{numeric vector or single-column matrix containing the design matrix for linear model. The length of the vector or the number of rows of the matrix should agree with the number of columns of M.}
  \item{\code{ndups}}{number of duplicate spots. Each gene is printed ndups times in adjacent spots on each array.}
  \item{\code{...}}{other arguments are not used}
}
}

which gives this html:

<!-- > writeLines(.rs.getHelp("anova.MAList", "limma")$html) -->
<div class="container">

<table style="width: 100%;"><tr><td>anova.MAList-method {limma}</td><td style="text-align: right;">R Documentation</td></tr></table>

<h2>ANOVA Table - method</h2>

<h3>Description</h3>

<p>Analysis of variance method for objects of class <code>MAList</code>.
Produces an ANOVA table useful for quality assessment by decomposing between and within gene sums of squares for a series of replicate arrays.
This method produces a single ANOVA Table rather than one for each gene and is not used to identify differentially expressed genes.
</p>


<h3>Usage</h3>

<p><code>anova(object,design=NULL,ndups=2,...)</code>
</p>


<h3>Arguments</h3>


<dl>
<dt><code>object</code></dt><dd><p>object of class <code>MAList</code>. Missing values in the M-values are not allowed.</p>
</dd>
<dt><code>design</code></dt><dd><p>numeric vector or single-column matrix containing the design matrix for linear model. The length of the vector or the number of rows of the matrix should agree with the number of columns of M.</p>
</dd>
<dt><code>ndups</code></dt><dd><p>number of duplicate spots. Each gene is printed ndups times in adjacent spots on each array.</p>
</dd>
<dt><code>...</code></dt><dd><p>other arguments are not used</p>
</dd>
</dl>

[romainfrancois]

The rendered rcmdcheck::rcmdcheck_process() help gives this, which also drifts from what HelpInfo.java expected to find.

<!-- [...] -->
<h3>Arguments</h3>


<ul>
<li> <p><code>cp</code>: A new rcmdcheck_process object.
</p>
</li>
<li> <p><code>path</code>: Path to a package tree or a package archive file. This is the
package to check.
</p>
</li>
<li> <p><code>args</code>: Command line arguments to <code style="white-space: pre;">&#8288;R CMD check&#8288;</code>.
</p>
</li>
<li> <p><code>build_args</code>: Command line arguments to <code style="white-space: pre;">&#8288;R CMD build&#8288;</code>.
</p>
</li>
<li> <p><code>check_dir</code>: Directory for the results.
</p>
</li>
<li> <p><code>libpath</code>: The library path to set for the check.
</p>
</li>
<li> <p><code>repos</code>: The <code>repos</code> option to set for the check.
This is needed for cyclic dependency checks if you use the
<code>--as-cran</code> argument. The default uses the current value.
</p>
</li>
<li> <p><code>env</code>: A named character vector, extra environment variables to
set in the check process.
</p>
</li></ul>

[romainfrancois]

Related to #10844

[ronblum]

Note: This occurs in both the Console and Editor Panes. Also, this was a problem in previous releases, such as 2022.07.2+576.

[ronblum]

Verified in RStudio Desktop 2023.03.0-daily+301 on MacOS 13.2. Tested in both Console and Source Editor, both by typing out the _ and using autocomplete.

[klwilson23]

@ronblum I see you closed this, but what was the fix? I'm getting the same error now from autocomplete in RStudio 2023.03.0 Build 386 "Cherry Blossom" Release (3c53477, 2023-03-09) for Windows

Error in if (nchar(current) > nchars) { : missing value where TRUE/FALSE needed

[kevinushey]

@klwilson23 the issue you're describing is different and is tracked in #12918.

[klwilson23]

@klwilson23 the issue you're describing is different and is tracked in #12918.

Fair enough - effectively my error looks like the .gif posted by @HenningLorenzen-ext-bayer associated with autocomplete and the object preview generating an error