public/browser
1
0
Fork 0
mirror of https://github.com/lightpanda-io/browser.git synced 2025-10-29 23:23:28 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

Merge pull request #959 from lightpanda-io/html-pre

Browse Source 
handle text content type with HTML
This commit is contained in:
Pierre Tachoire
2025-08-21 09:38:36 +02:00
committed by GitHub
parent 557f8444b2 3182a47858
commit a6f37633a1
1 changed files with 58 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

68
src/browser/page.zig
View File

@@ -109,6 +109,7 @@ pub const Page = struct {
err: anyerror,
parsed: void,
html: parser.Parser,
text: parser.Parser,
raw: std.ArrayListUnmanaged(u8),
raw_done: []const u8,
};
@@ -207,6 +208,20 @@ pub const Page = struct {
return out.writeAll(buf.items);
},
.raw_done => |data| return out.writeAll(data),
.text => {
// returns the <pre> element from the HTML
const doc = parser.documentHTMLToDocument(self.window.document);
const list = try parser.documentGetElementsByTagName(doc, "pre");
const pre = try parser.nodeListItem(list, 0) orelse return error.InvalidHTML;
const walker = Walker{};
var next: ?*parser.Node = null;
while (true) {
next = try walker.get_next(pre, next) orelse break;
const v = try parser.nodeTextContent(next.?) orelse return;
try out.writeAll(v);
}
return;
},
.html => {
// maybe page.wait timed-out, print what we have
log.warn(.http, "incomplete load", .{ .mode = "html" });
@@ -284,7 +299,7 @@ pub const Page = struct {
while (true) {
SW: switch (self.mode) {
.pre, .raw => {
.pre, .raw, .text => {
if (self.request_intercepted) {
// the page request was intercepted.
@@ -627,18 +642,44 @@ pub const Page = struct {
break :blk Mime.sniff(data);
} orelse .unknown;
const is_html = mime.isHTML();
log.debug(.http, "navigate first chunk", .{ .html = is_html, .len = data.len });
log.debug(.http, "navigate first chunk", .{ .content_type = mime.content_type, .len = data.len });
if (is_html) {
self.mode = .{ .html = try parser.Parser.init(mime.charset orelse "UTF-8") };
} else {
self.mode = .{ .raw = .{} };
}
self.mode = switch (mime.content_type) {
.text_html => .{ .html = try parser.Parser.init(mime.charset orelse "UTF-8") },
.application_json,
.text_javascript,
.text_css,
.text_plain,
=> blk: {
var p = try parser.Parser.init(mime.charset orelse "UTF-8");
try p.process("<html><head><meta charset=\"utf-8\"></head><body><pre>");
break :blk .{ .text = p };
},
else => .{ .raw = .{} },
};
}
switch (self.mode) {
.html => |*p| try p.process(data),
.text => |*p| {
// we have to escape the data...
var v = data;
while (v.len > 0) {
const index = std.mem.indexOfAnyPos(u8, v, 0, &.{ '<', '>' }) orelse {
try p.process(v);
return;
};
try p.process(v[0..index]);
switch (v[index]) {
'<' => try p.process("&lt;"),
'>' => try p.process("&gt;"),
else => unreachable,
}
v = v[index + 1 ..];
}
},
.raw => |*buf| try buf.appendSlice(self.arena, data),
.pre => unreachable,
.parsed => unreachable,
@@ -658,6 +699,13 @@ pub const Page = struct {
self.mode = .{ .raw_done = buf.items };
self.documentIsComplete();
},
.text => |*p| {
try p.process("</pre></body></html>");
const html_doc = p.html_doc;
p.deinit(); // don't need the parser anymore
try self.setDocument(html_doc);
self.documentIsComplete();
},
.html => |*p| {
const html_doc = p.html_doc;
p.deinit(); // don't need the parser anymore
@@ -719,7 +767,7 @@ pub const Page = struct {
self.clearTransferArena();
switch (self.mode) {
.html => |*p| p.deinit(), // don't need the parser anymore
.html, .text => |*p| p.deinit(), // don't need the parser anymore
else => {},
}
self.mode = .{ .err = err };
@@ -1063,6 +1111,6 @@ pub export fn scriptAddedCallback(ctx: ?*anyopaque, element: ?*parser.Element) c
}
self.script_manager.addFromElement(element.?) catch |err| {
log.warn(.browser, "dynamcic script", .{ .err = err });
log.warn(.browser, "dynamic script", .{ .err = err });
};
}