Merge pull request #333 from tstibbs/bugfix/332

Fixed handling of newlines within table cells. closes #332
spencermountain · Jan 28, 2020 · ad965f4 · ad965f4
2 parents 32da0b5 + 47a2e1d
commit ad965f4
Show file tree

Hide file tree

Showing 2 changed files with 33 additions and 3 deletions.
diff --git a/src/02-section/table/parse/index.js b/src/02-section/table/parse/index.js
@@ -93,8 +93,10 @@ const firstRowHeader = function(rows) {
 
 //turn a {|...table string into an array of arrays
 const parseTable = function(wiki) {
-  let lines = wiki.replace(/\r/g, '').split(/\n/)
-  lines = lines.map(l => l.trim())
+  let lines = wiki.replace(/\r/g, '')
+    .replace(/\n(\s*[^|!{\s])/g, ' $1')//remove unecessary newlines
+    .split(/\n/)
+    .map(l => l.trim())
   let rows = findRows(lines)
   //support colspan, rowspan...
   rows = handleSpans(rows)

diff --git a/tests/table.test.js b/tests/table.test.js
@@ -486,7 +486,7 @@ test('junky-table', t => {
   t.end()
 })
 
-test('table newlines', t => {
+test('table double bar', t => {
   var str = `{| class="wikitable"
 |-
 ! h1
@@ -516,3 +516,31 @@ test('table newlines', t => {
   t.equal(data[2].h3, 'ccc', 'h3')
   t.end()
 })
+
+//testing https://github.com/spencermountain/wtf_wikipedia/issues/332
+test('table newline', t => {
+  var str = `{| class="wikitable"
+|-
+! h1
+! h2
+! h3
+|-
+| a
+| b1<br />b2
+| c
+|-
+| a
+| b1
+b2
+| c
+|}`
+  var doc = wtf(str)
+  var data = doc.tables(0).keyValue()
+  t.equal(data[0].h1, 'a', 'h1')
+  t.equal(data[0].h2, 'b1 b2', 'h2')
+  t.equal(data[0].h3, 'c', 'h3')
+  t.equal(data[0].h1, 'a', 'h1')
+  t.equal(data[0].h2, 'b1 b2', 'h2')
+  t.equal(data[0].h3, 'c', 'h3')
+  t.end()
+})