Merge #52 from justinmk/fixes

fix: taglink special-cases |(| |{| …
neovim · Oct 19, 2022 · 6ffd5e3 · 6ffd5e3
2 parents 30cd470 + 8c7b8ff
commit 6ffd5e3
Show file tree

Hide file tree

Showing 7 changed files with 3,521 additions and 3,382 deletions.
diff --git a/README.md b/README.md
@@ -46,9 +46,10 @@ Known issues
   `:help lcs-tab`.
 - `url` doesn't handle _surrounding_ parens. E.g. `(https://example.com/#yay)` yields `word`
 - `url` doesn't handle _nested_ parens. E.g. `(https://example.com/(foo)#yay)`
-- `column_heading` currently only recognizes tilde "~" preceded by space (i.e.
-  "foo ~" not "foo~"). This covers 99% of :help files, but the grammar should
+- `column_heading` currently only recognizes tilde `~` preceded by space (i.e.
+  `foo ~` not `foo~`). This covers 99% of :help files, but the grammar should
   probably support "foo~" also.
+- `column_heading` children should be plaintext. Currently its children are parsed as `$._atom`.
 
 TODO
 ----

diff --git a/corpus/arguments.txt b/corpus/arguments.txt
@@ -92,11 +92,11 @@ EXTERNAL *netrw-externapp* {{{2
 (help_file
   (block
     (line
-      (word)
-      (ERROR
-        (word))
       (argument
-        (word))
+        (word)
+        (ERROR
+          (word)
+          (word)))
       (word)
       (codespan
         (word))

diff --git a/corpus/optionlink.txt b/corpus/optionlink.txt
@@ -29,12 +29,12 @@ world 'hello' world
       (word))))
 
 ================================================================================
-NOT optionlink: ' followed by whitespace
+NOT optionlink: ' or 'x
 ================================================================================
 'fillchars'
-	  stl		' ' or '^'	statusline
-	  wbr		' '		windowbar
-          tricky:	' 'yes'
+stl		' ' or '^'	statusline
+wbr		' '		windowbar
+tricky:	' 'yes'  's foo
 
 
 --------------------------------------------------------------------------------
@@ -60,12 +60,14 @@ NOT optionlink: ' followed by whitespace
       (word)
       (word)
       (optionlink
-        (word)))))
+        (word))
+      (word)
+      (word))))
 
 ================================================================================
 NOT optionlink #7 #14
 ================================================================================
-Let's see if that works.
+Let's see.
 no! ','sqlKeyword'
 single-char '-' 'g' '보'
 non-ascii: '\"' '%)' '-bang' '.*\\.log' '.gitignore' '@{${\"foo\"}}'
@@ -80,9 +82,6 @@ number: '04' 'ISO-10646-1' 'python3'
 (help_file
   (block
     (line
-      (word)
-      (word)
-      (word)
       (word)
       (word))
     (line
@@ -107,11 +106,11 @@ number: '04' 'ISO-10646-1' 'python3'
         (MISSING "*"))
       (word)
       (word)
-      (word)
-      (ERROR
-        (word))
       (argument
-        (word))
+        (word)
+        (ERROR
+          (word)
+          (word)))
       (word)
       (word))
     (line
@@ -155,8 +154,7 @@ x `after_codespan`'s
       (word)
       (codespan
         (word))
-      (word)
-      (ERROR))))
+      (word))))
 
 ================================================================================
 NOT optionlink 3

diff --git a/corpus/taglink.txt b/corpus/taglink.txt
@@ -15,6 +15,8 @@ taglink alone
 ================================================================================
 taglink in text
 ================================================================================
+|(|, |)|, |`|, |{|, |}|.
+
 Hello |world| hello
 
 |-+|	+[num]	line
@@ -26,6 +28,23 @@ Hello |world| hello
 --------------------------------------------------------------------------------
 
 (help_file
+  (block
+    (line
+      (taglink
+        (word))
+      (word)
+      (taglink
+        (word))
+      (word)
+      (taglink
+        (word))
+      (word)
+      (taglink
+        (word))
+      (word)
+      (taglink
+        (word))
+      (word)))
   (block
     (line
       (word)

diff --git a/grammar.js b/grammar.js
@@ -61,8 +61,10 @@ module.exports = grammar({
 
     // Explicit special cases: these are plaintext, not errors.
     _word_common: () => choice(
-      // NOT optionlink: single "'".
+      // NOT optionlink: '
       "'",
+      // NOT optionlink: 'x
+      seq("'", token.immediate(/[^'\n\t ]/)),
       // NOT optionlink: followed by non-lowercase char.
       seq("'", token.immediate(/[a-z]*[^'a-z\n\t ][a-z]*/), optional(token.immediate("'"))),
       // NOT optionlink: single char surrounded by "'".
@@ -199,7 +201,15 @@ module.exports = grammar({
     // Link to option: 'foo'. Lowercase non-digit ASCII, minimum 2 chars. #14
     optionlink: ($) => _word($, /[a-z][a-z]+/, "'", "'"),
     // Link to tag: |foo|
-    taglink: ($) => _word($, /[^|\n\t ]+/, '|', '|'),
+    taglink: ($) => _word($, choice(
+          token.immediate(/[^|\n\t ]+/),
+          // Special cases: |(| |{| …
+          token.immediate('{'),
+          token.immediate('}'),
+          token.immediate('('),
+          token.immediate(')'),
+          token.immediate('`'),
+    ), '|', '|'),
     // Inline code (may contain whitespace!): `foo bar`
     codespan: ($) => _word($, /[^``\n]+/, '`', '`'),
     // Argument: {arg}
@@ -208,9 +218,10 @@ module.exports = grammar({
 });
 
 // Word delimited by special chars.
-// The word_regex capture is aliased to "word" because they are semantically
-// the same: atoms of captured plain text.
-function _word($, word_regex, c1, c2, fname) {
+// `rule` can be a rule function or regex. It is aliased to "word" because they are
+// semantically the same: atoms of captured plain text.
+function _word($, rule, c1, c2, fname) {
+  rule = rule.test !== undefined ? token.immediate(rule) : rule
   fname = fname ?? 'text';
-  return seq(c1, field(fname, alias(token.immediate(word_regex), $.word)), token.immediate(c2));
+  return seq(c1, field(fname, alias(rule, $.word)), token.immediate(c2));
 }
diff --git a/src/grammar.json b/src/grammar.json
@@ -171,6 +171,22 @@
           "type": "STRING",
           "value": "'"
         },
+        {
+          "type": "SEQ",
+          "members": [
+            {
+              "type": "STRING",
+              "value": "'"
+            },
+            {
+              "type": "IMMEDIATE_TOKEN",
+              "content": {
+                "type": "PATTERN",
+                "value": "[^'\\n\\t ]"
+              }
+            }
+          ]
+        },
         {
           "type": "SEQ",
           "members": [
@@ -837,11 +853,51 @@
           "content": {
             "type": "ALIAS",
             "content": {
-              "type": "IMMEDIATE_TOKEN",
-              "content": {
-                "type": "PATTERN",
-                "value": "[^|\\n\\t ]+"
-              }
+              "type": "CHOICE",
+              "members": [
+                {
+                  "type": "IMMEDIATE_TOKEN",
+                  "content": {
+                    "type": "PATTERN",
+                    "value": "[^|\\n\\t ]+"
+                  }
+                },
+                {
+                  "type": "IMMEDIATE_TOKEN",
+                  "content": {
+                    "type": "STRING",
+                    "value": "{"
+                  }
+                },
+                {
+                  "type": "IMMEDIATE_TOKEN",
+                  "content": {
+                    "type": "STRING",
+                    "value": "}"
+                  }
+                },
+                {
+                  "type": "IMMEDIATE_TOKEN",
+                  "content": {
+                    "type": "STRING",
+                    "value": "("
+                  }
+                },
+                {
+                  "type": "IMMEDIATE_TOKEN",
+                  "content": {
+                    "type": "STRING",
+                    "value": ")"
+                  }
+                },
+                {
+                  "type": "IMMEDIATE_TOKEN",
+                  "content": {
+                    "type": "STRING",
+                    "value": "`"
+                  }
+                }
+              ]
             },
             "named": true,
             "value": "word"