userdiff: adding typescript pattern

Description
===========

* Add builtin userdiff pattern for TypeScript files. Recognizes function declarations, class definitions, arrow functions, and method definitions.

* Handles common modifiers like export, async, static, etc

* Added tests for the typescript pattern in userdiff.

Dhruv Arora (2):
userdiff: adding typescript pattern
fix(userdiff): sorted pattern and tests

 t/t4018/typescript-class-method         |  7 +++++++
 t/t4018/typescript-export-default-class |  7 +++++++
 t/t4018/typescript-export-function      |  7 +++++++
 userdiff.c                              | 15 +++++++++++++++
 4 files changed, 36 insertions(+)
 create mode 100644 t/t4018/typescript-class-method
 create mode 100644 t/t4018/typescript-export-default-class
 create mode 100644 t/t4018/typescript-export-function

base-commit: ce74208c2fa13943fffa58f168ac27a76d0eb789
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-2251%2FDhruv-0-Arora%2Fuserdiff%2Ftypescript-pattern-v1
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-2251/Dhruv-0-Arora/userdiff/typescript-pattern-v1
Pull-Request: https://github.com/git/git/pull/2251

gitgitgadget

From: Dhruv Arora <[email protected]>

Add builtin userdiff pattern for TypeScript files. Recognizes function declarations, class definitions, arrow functions, and method definitions.

Handles common modifiers like export, async, static, etc..

---
 userdiff.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

index fe710a68bf..086e3fa002 100644
@@ -337,6 +337,19 @@ PATTERNS("ruby",
 	 "(@|@@|\\$)?[a-zA-Z_][a-zA-Z0-9_]*"
 	 "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+|\\?(\\\\C-)?(\\\\M-)?."
 	 "|//=?|[-+*/<>%&^|=!]=|<<=?|>>=?|===|\\.{1,3}|::|[!=]~"),
+PATTERNS("typescript",
+	/* Starting with optional whitespace */
+	"^[ \t]*"
+	/* Followed by an optional export and/or async keyword */
+	"((export[ \t]+)?(async[ \t]+)?"
+	/* Followed by either a function or class declaration */
+	"((function|class)[ \t]+[a-zA-Z_][a-zA-Z0-9_]*[^{]*)"
+	/* or */
+	"|"
+	/* a variable declaration with const, let, or var */
+	"([ \t]*(const|let|var)[ \t]+[a-zA-Z_][a-zA-Z0-9_]*[ \t]*=))",
+	/* -- */
+	"[a-zA-Z_][a-zA-Z0-9_]*"),
 PATTERNS("rust",
 	 "^[\t ]*((pub(\\([^\\)]+\\))?[\t ]+)?((async|const|unsafe|extern([\t ]+\"[^\"]+\"))[\t ]+)?(struct|enum|union|mod|trait|fn|impl|macro_rules!)[< \t]+[^;]*)$",
 	 /* -- */

gitgitgadget

From: Dhruv Arora <[email protected]>

- Typescript pattern was not in alphabetical order, causing failing tests.
- Added 3 new typescript tests.
- Fixed pattern bug - incorrectly identifying const|let|var instead of function or class.

---
 t/t4018/typescript-class-method         |  7 +++++++
 t/t4018/typescript-export-default-class |  7 +++++++
 t/t4018/typescript-export-function      |  7 +++++++
 userdiff.c                              | 28 +++++++++++++------------
 4 files changed, 36 insertions(+), 13 deletions(-)
 create mode 100644 t/t4018/typescript-class-method
 create mode 100644 t/t4018/typescript-export-default-class
 create mode 100644 t/t4018/typescript-export-function

new file mode 100644
index 0000000000..6de6eff2e7
@@ -0,0 +1,7 @@
+export class RIGHT {
+    unique = 0
+    constructor () {
+        this.doNothing()
+    }
+    function ChangeMe() { }
+}

new file mode 100644
index 0000000000..aaede0dce3
@@ -0,0 +1,7 @@
+export default class RIGHT {
+	private x = 0;
+	private y = 0;
+	private z = 0;
+	// ChangeMe
+	render() {}
+}

new file mode 100644
index 0000000000..ba5bf71e80
@@ -0,0 +1,7 @@
+export async function RIGHT(url: string): Promise<string> {
+	const a = 1;
+	const b = 2;
+	const c = 3;
+	ChangeMe
+	return url;
+}

index 086e3fa002..7f5cadb30b 100644
@@ -337,19 +337,6 @@ PATTERNS("ruby",
 	 "(@|@@|\\$)?[a-zA-Z_][a-zA-Z0-9_]*"
 	 "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+|\\?(\\\\C-)?(\\\\M-)?."
 	 "|//=?|[-+*/<>%&^|=!]=|<<=?|>>=?|===|\\.{1,3}|::|[!=]~"),
-PATTERNS("typescript",
-	/* Starting with optional whitespace */
-	"^[ \t]*"
-	/* Followed by an optional export and/or async keyword */
-	"((export[ \t]+)?(async[ \t]+)?"
-	/* Followed by either a function or class declaration */
-	"((function|class)[ \t]+[a-zA-Z_][a-zA-Z0-9_]*[^{]*)"
-	/* or */
-	"|"
-	/* a variable declaration with const, let, or var */
-	"([ \t]*(const|let|var)[ \t]+[a-zA-Z_][a-zA-Z0-9_]*[ \t]*=))",
-	/* -- */
-	"[a-zA-Z_][a-zA-Z0-9_]*"),
 PATTERNS("rust",
 	 "^[\t ]*((pub(\\([^\\)]+\\))?[\t ]+)?((async|const|unsafe|extern([\t ]+\"[^\"]+\"))[\t ]+)?(struct|enum|union|mod|trait|fn|impl|macro_rules!)[< \t]+[^;]*)$",
 	 /* -- */

@@ -367,6 +354,21 @@ PATTERNS("scheme",
 	 "|([^][)(}{[ \t])+"),
 PATTERNS("tex", "^(\\\\((sub)*section|chapter|part)\\*{0,1}\\{.*)$",
 	 "\\\\[a-zA-Z@]+|\\\\.|([a-zA-Z0-9]|[^\x01-\x7f])+"),
+PATTERNS("typescript",
+	/* Starting with optional whitespace */
+	"^[ \t]*"
+	"("
+	/* Followed by an optional export and/or async and/or default keyword */
+	"(export[ \t]+)?(default[ \t]+)?(async[ \t]+)?"
+	/* Followed by either a function or class declaration */
+	"((function|class)[ \t]+[a-zA-Z_][a-zA-Z0-9_]*[^{]*)"
+	/* or */
+	"|"
+	/* a variable declaration with const, let, or var */
+	"^(const|let|var)[ \\t]+[a-zA-Z_][a-zA-Z0-9_]*[ \\t]*="
+	")",
+	/* -- */
+	"[a-zA-Z_][a-zA-Z0-9_]*"),
 { .name = "default", .binary = -1 },
 };
 #undef PATTERNS

gitgitgadget

We frown upon a patch series that makes mistakes in an earlier step, only to fix them in a later step. The "git rebase -i" command helps us pretend to be more perfect developers than we actually are, whipping your patch series into a shape that builds one small step on top of another in a logical succession. Such a patch series is easier to understand than a history that faithfully records all the stumbles the developer made until they reached the final solution.

Just have a single patch that adds the right pattern at the right place and add necessary tests.

When/if your reviewers suggest further changes, the way you should work on them is the same. Pretend as if you discarded everything you did, started from scratch, and reached the ideal result without making any mistakes or taking any detours along the way.

Thanks.

Am 28.03.26 um 22:40 schrieb Dhruv Arora via GitGitGadget:

Javascript and Typescript are related languages, if I am not mistaken. It would be appreciated if the new language driver could be used for both. Then the driver's name should be the superset language and the documentation (missing, BTW) should mention that the driver can be used for both languages.

We have had a number of submissions for Javascript or Typescript drivers in the past, but none of them were followed through to be integrated.

Typescript:

https://lore.kernel.org/git/[email protected]/
https://lore.kernel.org/git/[email protected]/

Javascript:

https://lore.kernel.org/git/[email protected]/
https://lore.kernel.org/git/[email protected]/
https://lore.kernel.org/git/[email protected]/

Please review these submission and the responses that they received. Perhaps you can find inspiration for improvement from them.

-- Hannes

FWIW, I believe both languages are officially spelled with a capital S: “JavaScript”, “TypeScript”. At least that is how Wikipedia and official sites present them.

I knew what you meant of course, and will now take my pedant hat off ;)