With hindsight, I revised decisions made earlier during
this development cycle:
Un-redirectable scriptlets have been removed from
/web_accessible_resources and instead put in the new
/assets/resources/scriptlets.js, which contains all
scriptlets used for web page injection purpose.
uBO will no longer fetch a remote version of built-in
resources.
Advanced setting `userResourcesLocation` will still be
honoured by uBO, and if set, will be fetched every
time at least one asset is updated.
This is a first step, the ultimate goal is to remove
the need for resources.txt, or at least to reduce to
only hotfixes or for trivial resources targeting very
specific websites.
Most resources will become immutable, i.e. they will
be part of uBO's code base. Advantages include easier
code maintenance (jshint, syntax highlight), and to
make scriptlets more easy to code review by external
parties (for example extension store reviewers).
TODO:
- More scriptlets need to be imported before next
release.
- Need to make legacy versions of uBO use a legacy
version of resources.txt, as all the now obsolete
scriptlets will have to be removed once uBO's
next release become widespread.
- Possibly need to add code to load binary
resources so that they can be injected as
data: URI. So far it's unclear whether this is
really needed. For example, this would be needed
if a xmlhttprequest is redirected to an image
resource.
Related issue:
- https://github.com/uBlockOrigin/uBlock-issues/issues/572
Wildcards are now allowed in the hostname part of redirect
filters. There will be an attempt to find the longest
right-hand portion of the hostname with no wildcard. If
no non-empty hostname can be extracted, `*` will be used.
Related issue:
- https://github.com/uBlockOrigin/uBlock-issues/issues/550
Related Chromium issue (I can't access it):
- https://bugs.chromium.org/p/chromium/issues/detail?id=957866
Findings so far: affects browsers based on Chromium 74.
I could not reproduce the issue with either Chromium 73 or
Google Chrome 75.
This commit is a mitigation: to prevent sites from using
uBO's internal WAR secret for tracking purpose. A secret
can be used for at most one second, after which a new secret
is generated.
The original issue related to the implementation of
secret-gated web accessible resources is:
- https://github.com/gorhill/uBlock/issues/2823
With the new PSL implementation, benchmarks do not show benefit
from caching the domain extracted from a hostname for later
reuse -- the caching seems to even add an overhead instead with
the new publicSuffixList implementation.
The motivation is to address the higher peak memory usage at launch
time with 3rd-gen HNTrie when a selfie was present.
The selfie generation prior to this change was to collect all
filtering data into a single data structure, and then to serialize
that whole structure at once into storage (using JSON.stringify).
However, HNTrie serialization requires that a large UintArray32 be
converted into a plain JS array, which itslef would be indirectly
converted into a JSON string. This was the main reason why peak
memory usage would be higher at launch from selfie, since the JSON
string would need to be wholly unserialized into JS objects, which
themselves would need to be converted into more specialized data
structures (like that Uint32Array one).
The solution to lower peak memory usage at launch is to refactor
selfie generation to allow a more piecemeal approach: each filtering
component is given the ability to serialize itself rather than to be
forced to be embedded in the master selfie. With this approach, the
HNTrie buffer can now serialize to its own storage by converting the
buffer data directly into a string which can be directly sent to
storage. This avoiding expensive intermediate steps such as
converting into a JS array and then to a JSON string.
As part of the refactoring, there was also opportunistic code
upgrade to ES6 and Promise (eventually all of uBO's code will be
proper ES6).
Additionally, the polyfill to bring getBytesInUse() to Firefox has
been revisited to replace the rather expensive previous
implementation with an implementation with virtually no overhead.
* refactoring assets management code
* finalizing refactoring of assets management
* various code review of new assets management code
* fix#2281
* fix#1961
* fix#1293
* fix#1275
* fix update scheduler timing logic
* forward compatibility (to be removed once 1.11+ is widespread)
* more codereview; give admins ability to specify own assets.json
* "assetKey" is more accurate than "path"
* fix group count update when building dom incrementally
* reorganize content (order, added URLs, etc.)
* ability to customize updater through advanced settings
* better spinner icon
