Rename `l` property to `len`, to avoid ambiguity as
`l` could mean _left_ or _length_. Typically `l` is
to be used for _left_ (whereas `r` is to be used for
_right_).
Additionally, add CodeMirror's bracket-matching and
bracket auto-closing to _My filters_ pane and and
bracket-matching to asset viewer page.
Specifically, do not invalidate valid hostnames when
there are extraneous separators: that sort of error
will be visually highlighted but will not otherwise
prevent a filter from being properly enforced.
This ensures proper garbage collection once the parser
is no longer referenced -- this is important now that
the parser is instantiated on-demand only.
Reported internally.
Potential regex flags are passed as is to RegExp contructor,
and in case of failure the query is deemed a plain text one.
Related commit:
- 8de67d22bd (diff-3f4aa453cefa49f6431f1bba3bb53a8e)
This commit moves some of the parsing logic of static
extended filtering into the static filtering parser; this
allows better syntax highlighting and creation-time
error-catching for cosmetic, HTML, and scriptlet filters.
Specifically:
There can't be more than one `redirect` or `csp`
option.
There can be no type specified when there is a
`csp` option.
There must be one single network-related type when
there is a `redirect` option. Since `empty` and
`mp4` imply a `redirect` and a network-related
type, these can't be used along another `redirect`
or network-related type.
Related commit:
- 01b1ed9a98
Reported internally. The issue could cause the logger
to be unable to successfully reverse-lookup a filter
list for a filter which had tokens longer than 6
characters followed by wildcard.
Regression from:
- 01b1ed9a98
Reported internally.
Regression from:
- 01b1ed9a98
The regression was made apparent be the fact that
AdGuard filter lists have lines ending with the
CR character.
Computing storage-used figures can take longer than
usual sometimes when a lot of filter lists are enabled,
and this can cause the Settings pane to take longer to
be filled the first time it is opened. Fetching
settings and storage-used figures separately removes
that potential delay (they were fetched together in
a single Promise.all() call).
Regression from:
- 01b1ed9a98
The new parser needs to be able to deal with trailing
newline characters, which if present will be interpreted
as trailing spaces.
A new standalone static filtering parser is introduced,
vAPI.StaticFilteringParser. It's purpose is to parse
line of text into representation suitable for
compiling filters. It can additionally serves for
syntax highlighting purpose.
As a side effect, this solves:
- https://github.com/uBlockOrigin/uBlock-issues/issues/1038
This is a first draft, there are more work left to do
to further perfect the implementation and extend its
capabilities, especially those useful to assist filter
authors.
For the time being, this commits break line-continuation
syntax highlighting -- which was already flaky prior to
this commit anyway.
Raymond Hill