Task failure loading an image from the commandline

[larsbergstrom]

./servo http://s27.postimg.org/vqbtrolyr/servo.jpg

According to @Ms2ger, this is attempting to read in the jpeg as an HTML file. We then crash in hubbub attempting to create a string with non-UTF8 characters.

task 'ScriptTask' failed at 'called `Result::unwrap()` on an `Err` value: [239, 191, 189, 23, 47, 91, 45, 239, 191, 189, 91, 239, 191, 189, 31, 239, 191, 189, 239, 191, 189, 101, 23, 198, 159, 239, 191, 189, 239, 191, 189, 246, 188, 178, 183, 239, 191, 189, 239, 191, 189]', /home/larsberg/rust/src/libcore/result.rs:548
stack backtrace:
   1:           0xfb6210 - rt::backtrace::imp::write::h31edb94522b79b60gcq
   2:           0xfb9270 - failure::on_fail::hb2b37deb41014e2frxq
   3:           0xfd7d40 - unwind::begin_unwind_inner::hb0aa0f9950827646B8d
   4:           0xfd7a30 - unwind::begin_unwind_fmt::hef9a8959da61fbba65d
   5:           0xfd79f0 - rust_begin_unwind
   6:          0x10084d0 - failure::begin_unwind::h9965619bbb55eb47c0j
   7:           0xf9f3b0 - fmt::format::h56249d3ff4157084bYp
   8:           0x7f1910 - dom::attr::Attr::set_value::ha6e6fe6170468319yAZ
   9:           0x812cf0 - dom::element::JSRef<'a, Element>.AttributeHandlers::do_set_attribute::h3a009df2ffefc702y34
  10:           0x812700 - dom::element::JSRef<'a, Element>.AttributeHandlers::set_attribute_from_parser::h473f70b24b272d89DX4
  11:           0x8a79c0 - html::hubbub_html_parser::parse_html::closure.85651
  12:           0xcb7300 - hubbub::tree_callbacks::create_element::h77b137ab60f0220fbUa
  13:           0xcbce20 - insert_element
  14:           0xcc9cd0 - handle_in_body
  15:           0xcbbff0 - hubbub_treebuilder_token_handler
  16:           0xcbd720 - hubbub_tokeniser_emit_token
  17:           0xcbd7f0 - emit_current_tag
  18:           0xcc4280 - hubbub_tokeniser_run
  19:           0xcb9880 - hubbub_parser_parse_chunk
  20:           0xcb8b50 - hubbub::Parser<'a>::parse_chunk::h1603f5c703b297c17ta
  21:           0x8959f0 - html::hubbub_html_parser::parse_html::h56843be8a1735d51SOl
  22:           0x9323f0 - script_task::ScriptTask::load::hb267e5d49c0f34359Eo
  23:           0x91f8f0 - script_task::ScriptTask::handle_msgs::hd14c335bcc1a5125Rko
  24:           0x91f8a0 - script_task::ScriptTask::start::h5f9cbd090f299ebdHko
  25:           0x484920 - script_task::ScriptTask.ScriptTaskFactory::create::closure.10135
  26:           0x47f790 - task::TaskBuilder<S>::try_future::closure.10046
  27:           0xcabb60 - task::bootstrap_green_task::closure.6716
  28:           0xfd8a20 - rust_try
  29:           0xfd58e0 - unwind::try::h0648d3e5329c30960Wd
  30:           0xfd5690 - task::Task::run::h3139914caede1196C4c
  31:           0xcab8d0 - task::bootstrap_green_task::hec3319e2cc8a64a6Pdc
  32:                0x0 - <unknown>

And:

DEBUG:hubbub::hubbub::tree_callbacks: from_hubbub_string: �/[-�[���eƟ��������

[Manishearth]

Would it make sense to detect images by MIME types and wrapping them in a tiny HTML document? Most browsers seem to do that.

[Manishearth]

Fixed in #3131