From 87254086b81c12451ccbfb36c36653e259a1c54d Mon Sep 17 00:00:00 2001
From: Pedro Faria <69123925+pedropark99@users.noreply.github.com>
Date: Tue, 24 Oct 2023 16:57:48 -0300
Subject: [PATCH] close #2226: add an error handler to improve YAML error
 message (#2294)

Co-authored-by: Yihui Xie <xie@yihui.name>
---
 NEWS.md         |  4 ++++
 R/concordance.R |  6 +++---
 R/output.R      | 23 +++++++++++++++--------
 R/parser.R      | 19 +++++++++++++++++--
 R/utils.R       |  4 +---
 5 files changed, 40 insertions(+), 16 deletions(-)

diff --git a/NEWS.md b/NEWS.md
index 3c73cd69d5..ad103bdeff 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -1,5 +1,9 @@
 # CHANGES IN knitr VERSION 1.45
 
+## NEW FEATURES
+
+- Improved the error message to contain more specific information when YAML chunk options could not be parsed (thanks, @pedropark99, #2294).
+
 ## BUG FIXES
 
 - Special characters in the chunk option `fig.alt` are properly escaped now (thanks, @jay-sf, #2290).
diff --git a/R/concordance.R b/R/concordance.R
index 6c7c6e6546..3e267cf1e0 100644
--- a/R/concordance.R
+++ b/R/concordance.R
@@ -2,7 +2,7 @@
 
 # record input/output lines numbers in Rnw/tex and filenames
 knit_concord = new_defaults(list(
-  inlines = NULL, outlines = NULL, infile = NULL, outfile = NULL
+  inlines = NULL, outlines = NULL, infile = NULL, outfile = NULL, block = NULL
 ))
 
 # do not consider child mode for concordance
@@ -10,11 +10,11 @@ concord_mode = function() {
   opts_knit$get('concordance') && !child_mode()
 }
 
-current_lines = function(i) {
+current_lines = function(i = knit_concord$get('block')) {
   # a helpr function to return line numbers for block i
   n = knit_concord$get('inlines')
   n1 = sum(head(n, i)); n0 = n1 - n[i] + 2
-  c(min(n0, n1), n1)
+  paste(c(min(n0, n1), n1), collapse = '-')
 }
 
 # generate concordance for RStudio
diff --git a/R/output.R b/R/output.R
index a0f6bb625a..1c7b5ca167 100644
--- a/R/output.R
+++ b/R/output.R
@@ -311,20 +311,18 @@ process_file = function(text, output) {
     }
     if (progress && is.function(pb$update)) pb$update(i)
     group = groups[[i]]
-    res[i] = withCallingHandlers(
+    knit_concord$set(block = i)
+    res[i] = handle_error(
       withCallingHandlers(
         if (tangle) process_tangle(group) else process_group(group),
         error = function(e) if (xfun::pkg_available('rlang', '1.0.0')) rlang::entrace(e)
       ),
-      error = function(e) {
+      function(e, loc) {
         setwd(wd)
         write_utf8(res, output %n% stdout())
-        message(
-          '\nQuitting from lines ', paste(current_lines(i), collapse = '-'),
-          if (labels[i] != '') sprintf(' [%s]', labels[i]),
-          sprintf(' (%s)', knit_concord$get('infile'))
-        )
-      }
+        paste0('\nQuitting from lines ', loc)
+      },
+      if (labels[i] != '') sprintf(' [%s]', labels[i])
     )
   }
 
@@ -337,6 +335,15 @@ process_file = function(text, output) {
   res
 }
 
+# if an expr throws an an error, message the location of the error if possible
+handle_error = function(expr, handler, label = '') {
+  withCallingHandlers(expr, error = function(e) {
+    # return a string to point out the error location
+    loc = paste0(current_lines(), label, sprintf(' (%s)', knit_concord$get('infile')))
+    message(one_string(handler(e, loc)))
+  })
+}
+
 auto_out_name = function(input, ext = tolower(file_ext(input))) {
   base = sans_ext(input)
   name = if (opts_knit$get('tangle')) c(base, '.R') else
diff --git a/R/parser.R b/R/parser.R
index 78a97ed540..dad04ba981 100644
--- a/R/parser.R
+++ b/R/parser.R
@@ -20,7 +20,9 @@ split_file = function(lines, set.preamble = TRUE, patterns = knit_patterns$get()
     knit_concord$set(inlines = sapply(groups, length)) # input line numbers for concordance
 
   # parse 'em all
-  lapply(groups, function(g) {
+  lapply(seq_along(groups), function(i) {
+    knit_concord$set(block = i)
+    g = groups[[i]]
     block = grepl(chunk.begin, g[1])
     if (!set.preamble && !parent_mode()) {
       return(if (block) '' else g) # only need to remove chunks to get pure preamble
@@ -302,7 +304,20 @@ partition_chunk = function(engine, code) {
   meta = substr(src, nchar(s1) + 1, nchar(src) - nchar(s2))
   # see if the metadata looks like YAML or CSV
   if (grepl('^[^ :]+:($|\\s)', meta[1])) {
-    meta = yaml::yaml.load(meta, handlers = list(expr = parse_only))
+    meta = handle_error(
+      yaml::yaml.load(meta, handlers = list(expr = parse_only)),
+      function(e, loc) {
+        x = e$message
+        r = 'line (\\d+), column (\\d+)'
+        m = regmatches(x, regexec(r, x, perl = TRUE))[[1]]
+        if (length(m) < 3) return()
+        m = as.integer(m[-1])  # c(row, col)
+        c(
+          sprintf('Failed to parse YAML inside code chunk at lines %s:', loc), '',
+          append(meta, paste0(strrep(' ', m[2]), '^~~~~~'), m[1]), ''
+        )
+      }
+    )
     if (!is.list(meta) || length(names(meta)) == 0) {
       warning('Invalid YAML option format in chunk: \n', one_string(meta), '\n')
       meta = list()
diff --git a/R/utils.R b/R/utils.R
index d32f63b4d1..6025b73a7f 100644
--- a/R/utils.R
+++ b/R/utils.R
@@ -1141,9 +1141,7 @@ str_split = function(x, split, ...) {
 # methods to update/close it
 txt_pb = function(total, labels) {
   if (getOption('knitr.progress.linenums', FALSE)) {
-    nums = sapply(seq_along(labels), function(i) {
-      paste(current_lines(i), collapse = '-')
-    })
+    nums = sapply(seq_along(labels), current_lines)
     labels = sprintf(
       '%s%s%s:%s', labels, ifelse(labels == '', '', ' @ '),
       knit_concord$get('infile'), nums