fixing more crashes!

kristoff-it · kristoff-it · commit 98f3b8cfaadd · 2024-07-15T18:10:42.000+02:00
diff --git a/README.md b/README.md
@@ -1,11 +1,11 @@
 # SuperHTML
-Actually functional HTML Language Server (and Templating Language Library).
+HTML Language Server and Templating Language Library
 
-**NOTE: SuperHTML is still incomplete, some features are missing and looking of somebody to implement them :^)**  
-*And there's also bugs, waiting for people to report them as Issues :^)*
 
 ## HTML Language Server
-The Super CLI Tool offers syntax checking and autoformatting features for HTML files.
+The Super CLI Tool offers **syntax checking** and **autoformatting** features for HTML files.
+
+The tool can be used either directly (for example by running it on save), or through a LSP client implementation.
 
 ```
 $ super
@@ -36,7 +36,7 @@ As an example, HTML allows for closing some tags implicitly. For example the fol
 </ul>
 ```
 
-This will still be reported as an error by SuperHTML because otherwise the following snippet would have to be considered correct (while it's much probably a typo):
+This will still be reported as an error by SuperHTML because otherwise the following snippet would have to be considered correct (while it's most probably a typo):
 
 ```html
 <h1>Title<h1>
@@ -115,11 +115,27 @@ After:
 
 ### Editor support
 #### VSCode
-1. Download a prebuilt version of `super` from the Releases section (or build it yourself).
-2. Put `super` in your `PATH`.
+1. Download a prebuilt version of `superhtml` from the Releases section (or build it yourself).
+2. Put `superhtml` in your `PATH`.
 3. Install the [Super HTML VSCode extension](https://marketplace.visualstudio.com/items?itemName=LorisCro.super). 
 
-#### Flow
+#### Helix
+Add to your `.config/helix/languages.toml`:
+```toml
+[language-server.superhtml-lsp]
+command = "superhtml"
+args = ["lsp"]
+
+[[language]]
+name = "html"
+scope = "source.html"
+roots = []
+file-types = ["html"]
+language-servers = [ "superhtml-lsp" ]
+```
+See https://helix-editor.com for more information on how to add new language servers.
+
+#### [Flow](https://github.com/neurocyte/flow)
 Already defaults to using SuperHTML, just add the executable to your `PATH`.
 
 #### Other editors
@@ -128,7 +144,7 @@ Follow your editor specific intructions on how to define a new Language Server f
 *(Also feel free to contribute more specific intructions to this readme / add files under the `editors/` subdirectory).*
 
 ## Templating Language Library
-SuperHTML is not only an LSP but also an HTML templating language. More on that soon.
+SuperHTML is also a HTML templating language. More on that soon.
 
 ## Contributing
 SuperHTML tracks the latest Zig release (0.13.0 at the moment of writing). 
diff --git a/src/cli/lsp.zig b/src/cli/lsp.zig
@@ -1,4 +1,5 @@
 const std = @import("std");
+const builtin = @import("builtin");
 const assert = std.debug.assert;
 const lsp = @import("lsp");
 const types = lsp.types;
@@ -276,16 +277,13 @@ pub const Handler = struct {
         log.debug("format request!!", .{});
 
         const doc = self.files.getPtr(request.textDocument.uri) orelse return null;
-
         if (doc.html.errors.len != 0) {
             return null;
         }
 
         log.debug("format!!", .{});
 
-        var buf = std.ArrayList(u8).init(self.gpa);
-        errdefer buf.deinit();
-
+        var buf = std.ArrayList(u8).init(arena);
         try doc.html.render(doc.src, buf.writer());
 
         const edits = try lsp.diff.edits(
@@ -295,9 +293,11 @@ pub const Handler = struct {
             .@"utf-8",
         );
 
-        self.gpa.free(doc.src);
-        doc.src = try buf.toOwnedSlice();
-        try doc.reparse(self.gpa);
+        if (builtin.mode == .Debug) {
+            if (std.mem.eql(u8, buf.items, doc.src)) {
+                std.debug.assert(edits.items.len == 0);
+            }
+        }
 
         return edits.items;
     }
diff --git a/src/fuzz.zig b/src/fuzz.zig
@@ -5,24 +5,33 @@ pub const std_options = .{ .log_level = .err };
 
 pub fn main() !void {
     var gpa_impl: std.heap.GeneralPurposeAllocator(.{}) = .{};
+
     // this will check for leaks and crash the program if it finds any
     defer std.debug.assert(gpa_impl.deinit() == .ok);
     const gpa = gpa_impl.allocator();
 
     // Read the data from stdin
     const stdin = std.io.getStdIn();
-    const data = try stdin.readToEndAlloc(gpa, std.math.maxInt(usize));
-    defer gpa.free(data);
+    const src = try stdin.readToEndAlloc(gpa, std.math.maxInt(usize));
+    defer gpa.free(src);
 
-    const ast = try super.html.Ast.init(gpa, data, .html);
+    const ast = try super.html.Ast.init(gpa, src, .html);
     defer ast.deinit(gpa);
+
+    if (ast.errors.len == 0) {
+        try ast.render(src, std.io.null_writer);
+    }
 }
 
 test "afl++ fuzz cases" {
     const cases: []const []const u8 = &.{
         @embedFile("fuzz/2.html"),
         @embedFile("fuzz/3.html"),
         @embedFile("fuzz/12.html"),
+        @embedFile("fuzz/round2/2.html"),
+        @embedFile("fuzz/round2/3.html"),
+        @embedFile("fuzz/round3/2.html"),
+        @embedFile("fuzz/77.html"),
     };
 
     for (cases) |c| {
diff --git a/src/fuzz/77.html b/src/fuzz/77.html
@@ -0,0 +1,5 @@
+tl>�</s</>
+;sphmlK
+<pYn.
+
+
diff --git a/src/fuzz/round2/2.html b/src/fuzz/round2/2.html
diff --git a/src/fuzz/round2/3.html b/src/fuzz/round2/3.html
diff --git a/src/fuzz/round3/2.html b/src/fuzz/round3/2.html
diff --git a/src/html/Ast.zig b/src/html/Ast.zig
@@ -574,10 +574,10 @@ pub fn render(ast: Ast, src: []const u8, w: anytype) !void {
                     current,
                 });
 
-                switch (current.kind) {
+                if (pre == 0) switch (current.kind) {
                     else => {},
                     .element => indentation -= 1,
-                }
+                };
 
                 const open_was_vertical = std.ascii.isWhitespace(src[current.open.end]);
 
@@ -748,7 +748,9 @@ pub fn render(ast: Ast, src: []const u8, w: anytype) !void {
                                 current = ast.nodes[current.first_child_idx];
                             }
                         },
-                        .element_void => {
+                        .element_void,
+                        .element_self_closing,
+                        => {
                             if (current.next_idx != 0) {
                                 current = ast.nodes[current.next_idx];
                             } else {
@@ -760,6 +762,7 @@ pub fn render(ast: Ast, src: []const u8, w: anytype) !void {
                 },
                 .exit => {
                     std.debug.assert(current.kind != .element_void);
+                    std.debug.assert(current.kind != .element_self_closing);
                     last_rbracket = current.close.end;
                     if (current.close.start != 0) {
                         const name = blk: {
diff --git a/src/html/Tokenizer.zig b/src/html/Tokenizer.zig
@@ -1702,7 +1702,9 @@ fn next2(self: *Tokenizer, src: []const u8) ?struct {
                     // Set the temporary buffer to the empty string. Emit a U+003C LESS-THAN SIGN character token. Reconsume in the script data double escape start state.
                     'a'...'z', 'A'...'Z' => {
                         self.idx -= 1;
-                        @panic("TODO");
+                        self.state = .{
+                            .script_data_double_escape_start = state,
+                        };
                     },
                     // Anything else
                     // Emit a U+003C LESS-THAN SIGN character token. Reconsume in the script data escaped state.
@@ -1722,7 +1724,11 @@ fn next2(self: *Tokenizer, src: []const u8) ?struct {
                     // Create a new end tag token, set its tag name to the empty string. Reconsume in the script data escaped end tag name state.
                     'a'...'z', 'A'...'Z' => {
                         self.idx -= 1;
-                        @panic("TODO");
+                        var new = state;
+                        new.name_start = self.idx;
+                        self.state = .{
+                            .script_data_escaped_end_tag_name = new,
+                        };
                     },
                     // Anything else
                     // Emit a U+003C LESS-THAN SIGN character token and a U+002F SOLIDUS character token. Reconsume in the script data escaped state.
@@ -2348,12 +2354,14 @@ fn next2(self: *Tokenizer, src: []const u8) ?struct {
                     },
                     // U+003D EQUALS SIGN (=)
                     // Switch to the before attribute value state.
-                    '=' => self.state = .{
-                        .before_attribute_value = .{
-                            .tag = state.tag,
-                            .name = state.name,
-                            .equal_sign = self.idx - 1,
-                        },
+                    '=' => {
+                        self.state = .{
+                            .before_attribute_value = .{
+                                .tag = state.tag,
+                                .name = state.name,
+                                .equal_sign = self.idx - 1,
+                            },
+                        };
                     },
                     // U+003E GREATER-THAN SIGN (>)
                     // Switch to the data state. Emit the current tag token.
@@ -2383,8 +2391,9 @@ fn next2(self: *Tokenizer, src: []const u8) ?struct {
                     // Anything else
                     // Start a new attribute in the current tag token. Set that attribute name and value to the empty string. Reconsume in the attribute name state.
                     else => {
-                        self.idx -= 1;
+                        std.debug.assert(state.name.len() != 0);
 
+                        self.idx -= 1;
                         var tag = state.tag;
                         tag.attr_count += 1;
 
@@ -3513,28 +3522,22 @@ fn next2(self: *Tokenizer, src: []const u8) ?struct {
             // https://html.spec.whatwg.org/multipage/parsing.html#comment-less-than-sign-bang-dash-state
             .comment_less_than_sign_bang_dash => |comment_start| {
                 if (!self.consume(src)) {
+                    self.idx -= 1;
                     self.state = .{ .comment_end_dash = comment_start };
                 }
                 switch (self.current) {
                     // U+002D HYPHEN-MINUS (-)
                     // Switch to the comment less-than sign bang dash dash state.
-                    '-' => switch (self.state) {
-                        else => unreachable,
-                        .comment_less_than_sign_bang => {
-                            self.state = .{
-                                .comment_less_than_sign_bang_dash = comment_start,
-                            };
-                        },
-                        .comment_less_than_sign_bang_dash => {
-                            self.state = .{
-                                .comment_less_than_sign_bang_dash_dash = comment_start,
-                            };
-                        },
+                    '-' => self.state = .{
+                        .comment_less_than_sign_bang_dash_dash = comment_start,
                     },
 
                     // Anything else
                     // Reconsume in the comment end dash state.
-                    else => self.state = .{ .comment_end_dash = comment_start },
+                    else => {
+                        self.idx -= 1;
+                        self.state = .{ .comment_end_dash = comment_start };
+                    },
                 }
             },
             // https://html.spec.whatwg.org/multipage/parsing.html#comment-less-than-sign-bang-dash-dash-state

-Original file line number
+Diff line change
@@ @@ -0,0 +1,5 @@ @@
 +tl>�</s</>
 +;sphmlK
 +<pYn.
++
++