Don't escape unicode control characters. While these characters would
break commands if present (i.e. non functional cURL) and could potentially
be sanitized to spaces, that would likely just move the goalposts again.
Following Mozilla here.
Follow-up to 15335ce39d1ea2ef4585a36a4d562a3894459a15
Trying to cross-platform sanitize causes issues escaping too much and
as a result mangling the curl commands that get spit out.
This reverts previously added Windows-specific escaping on POSIX and
vice versa.
Since we do not offer cross-platform copying as a curl command in devtools
anyway, it makes little sense to sanitize for the O.S. we're not running on
as we would not be using the command processor in such environments that
could trip over characters for the other O.S. This was previously added
as a defense-in-depth in case we would start offering this, but I see no
real reason to do so, anyway.
Also decided to in-line comments for readability instead of a bulleted
list in the function head.
Mozilla has taken similar steps because of fall-out, but they do offer
cross-platform "copy as curl" for corner cases, so they will have to
find solutions for a problem we won't have.
This removes a lot of the plumbing for having the platform embed itself
through IPC which was required for B2G running the browser as both
shell and browser application.
Remove early return in nsExternalProtocolHandler::NewChannel when handler doesn't exist.
Fix devtools expectation that newChannel throws for unsupported external handlers
This leaves getAdjustedQuads alone because it lives in its own world and its
result gets sent over IPC. That leaves things in a bit of an intermediate state,
but that should be OK for now.
Backported from Mozilla bug 1186265.
Vim control lines were re-introduced or not entirely cleaned up.
This nukes them again.
Removing from the rest of js, caps, chrome, config, devtools, docshell,
image, intl. More to come.
Windows commonly fails to work because of --compressed, and its string
escaping needed improvement because of the complexities of argument
parsing in command windows.
Some hoops to make sure the test still works if the default value for the pref
is false, requiring setting and resetting it and making sure the observers have
time to react to these changes before testing.
This works by stripping the optional chaining characters from the completion part variable, allowing the developer tools' parser to proceed as if it were a regular, non-optional expression.
Tests were partially based on: https://bugzilla.mozilla.org/show_bug.cgi?id=1594009
Tests for features that do not apply to our version of developer tools (e.g. autocomplete for integer literals, ignoring spaces between property accessors, etc) were excluded.
Moonchild
MeladJM
Andy
Francis Dominic Fajardo
FranklinDM
Martok
Brian Smith
Job Bautista