From ba575b00a4ea38e6e6bdd1bcd5c42c87c95a97ff Mon Sep 17 00:00:00 2001
From: Ben Kettle <ben@semgrep.com>
Date: Fri, 6 Dec 2024 14:26:40 -0800
Subject: [PATCH] chore: add toml language

---
 .gitmodules                                   |  3 ++
 lang/semgrep-grammars/lang/toml               |  1 +
 .../src/semgrep-toml/Makefile                 |  1 +
 .../src/semgrep-toml/grammar.js               | 29 +++++++++++++++++++
 lang/semgrep-grammars/src/semgrep-toml/prep   |  1 +
 .../src/semgrep-toml/test/corpus/semgrep.txt  |  0
 lang/semgrep-grammars/src/tree-sitter-toml    |  1 +
 lang/toml/Makefile                            |  1 +
 lang/toml/extensions.txt                      |  6 ++++
 lang/toml/fyi.list                            |  3 ++
 lang/toml/projects.txt                        |  3 ++
 11 files changed, 49 insertions(+)
 create mode 120000 lang/semgrep-grammars/lang/toml
 create mode 120000 lang/semgrep-grammars/src/semgrep-toml/Makefile
 create mode 100644 lang/semgrep-grammars/src/semgrep-toml/grammar.js
 create mode 120000 lang/semgrep-grammars/src/semgrep-toml/prep
 create mode 100644 lang/semgrep-grammars/src/semgrep-toml/test/corpus/semgrep.txt
 create mode 160000 lang/semgrep-grammars/src/tree-sitter-toml
 create mode 120000 lang/toml/Makefile
 create mode 100644 lang/toml/extensions.txt
 create mode 100644 lang/toml/fyi.list
 create mode 100644 lang/toml/projects.txt

diff --git a/.gitmodules b/.gitmodules
index d3004c26..6265d3bf 100644
--- a/.gitmodules
+++ b/.gitmodules
@@ -127,3 +127,6 @@
 [submodule "tree-sitter-c-sharp-pro"]
 	path = lang/semgrep-grammars/src/tree-sitter-c-sharp-pro
 	url = https://github.com/tree-sitter/tree-sitter-c-sharp.git
+[submodule "lang/semgrep-grammars/src/tree-sitter-toml"]
+	path = lang/semgrep-grammars/src/tree-sitter-toml
+	url = git@github.com:tree-sitter-grammars/tree-sitter-toml.git
diff --git a/lang/semgrep-grammars/lang/toml b/lang/semgrep-grammars/lang/toml
new file mode 120000
index 00000000..d088e8f9
--- /dev/null
+++ b/lang/semgrep-grammars/lang/toml
@@ -0,0 +1 @@
+../src/semgrep-toml
\ No newline at end of file
diff --git a/lang/semgrep-grammars/src/semgrep-toml/Makefile b/lang/semgrep-grammars/src/semgrep-toml/Makefile
new file mode 120000
index 00000000..16641521
--- /dev/null
+++ b/lang/semgrep-grammars/src/semgrep-toml/Makefile
@@ -0,0 +1 @@
+../Makefile.common
\ No newline at end of file
diff --git a/lang/semgrep-grammars/src/semgrep-toml/grammar.js b/lang/semgrep-grammars/src/semgrep-toml/grammar.js
new file mode 100644
index 00000000..eea84b5b
--- /dev/null
+++ b/lang/semgrep-grammars/src/semgrep-toml/grammar.js
@@ -0,0 +1,29 @@
+/*
+  semgrep-toml
+
+  Extends the standard toml grammar with semgrep pattern constructs.
+*/
+
+const base_grammar = require('tree-sitter-toml/grammar');
+
+module.exports = grammar(base_grammar, {
+  name: 'toml',
+
+  conflicts: ($, previous) => previous.concat([
+  ]),
+
+  /*
+     Support for semgrep ellipsis ('...') and metavariables ('$FOO'),
+     if they're not already part of the base grammar.
+  */
+  rules: {
+  /*
+    semgrep_ellipsis: $ => '...',
+
+    _expression: ($, previous) => choice(
+      $.semgrep_ellipsis,
+      ...previous.members
+    ),
+  */
+  }
+});
diff --git a/lang/semgrep-grammars/src/semgrep-toml/prep b/lang/semgrep-grammars/src/semgrep-toml/prep
new file mode 120000
index 00000000..df68f30a
--- /dev/null
+++ b/lang/semgrep-grammars/src/semgrep-toml/prep
@@ -0,0 +1 @@
+../prep.common
\ No newline at end of file
diff --git a/lang/semgrep-grammars/src/semgrep-toml/test/corpus/semgrep.txt b/lang/semgrep-grammars/src/semgrep-toml/test/corpus/semgrep.txt
new file mode 100644
index 00000000..e69de29b
diff --git a/lang/semgrep-grammars/src/tree-sitter-toml b/lang/semgrep-grammars/src/tree-sitter-toml
new file mode 160000
index 00000000..64b56832
--- /dev/null
+++ b/lang/semgrep-grammars/src/tree-sitter-toml
@@ -0,0 +1 @@
+Subproject commit 64b56832c2cffe41758f28e05c756a3a98d16f41
diff --git a/lang/toml/Makefile b/lang/toml/Makefile
new file mode 120000
index 00000000..16641521
--- /dev/null
+++ b/lang/toml/Makefile
@@ -0,0 +1 @@
+../Makefile.common
\ No newline at end of file
diff --git a/lang/toml/extensions.txt b/lang/toml/extensions.txt
new file mode 100644
index 00000000..0f49d20a
--- /dev/null
+++ b/lang/toml/extensions.txt
@@ -0,0 +1,6 @@
+# File extensions for the target language, one per line. This is used for
+# collecting parsing stats from the repos specified in 'projects.txt'. e.g.:
+#
+# .h
+# .c
+#
diff --git a/lang/toml/fyi.list b/lang/toml/fyi.list
new file mode 100644
index 00000000..fe92bc41
--- /dev/null
+++ b/lang/toml/fyi.list
@@ -0,0 +1,3 @@
+semgrep-grammars/src/tree-sitter-toml/LICENSE
+semgrep-grammars/src/tree-sitter-toml/grammar.js
+semgrep-grammars/src/semgrep-toml/grammar.js
diff --git a/lang/toml/projects.txt b/lang/toml/projects.txt
new file mode 100644
index 00000000..5f8b20ca
--- /dev/null
+++ b/lang/toml/projects.txt
@@ -0,0 +1,3 @@
+# Git URLs of publicly-accessible projects to be used for parsing stats,
+# one per line.
+#