Yeah I think hpHosts is back to being overzealous, I've seen things in it lately which should not be in there.
Hmm mine reads umatrix 52 and ublock 47. I have noticed a memory increase since i installed chrome 64. Nowhere near 54 a few people have mentioned.
The leak is easy to reproduce: keep clicking on the extension icon to open the popup UI. Memory footprint will go up (normal), but never fully reclaimed (not normal). I did also reproduce with other extensions, and any extension with non-trivial popup UI (ex. uMatrix, Ghostery) will end up leaking more severely.
Interesting, after some brief testing in a virtual machine, this memory leak seems to be greater in 64 bit Chrome. In 32 bit it is not as severe. Out of curiosity and might be completely unrelated, does umatrix have any issues deleting cookies in chrome 64 bit? After browsing to the same sites with the same settings, chrome 32 bit cookies were deleted by umatrix whereas chrome 64 cookies seem not to be deleted any more. These are my settings.
I must say that, after reading this, I have deactivated hpHosts in µMatrix and everything in Chrome works better: a lot less broken pages, Chrome starts instantly and all the extensions load right (previously there were frequently some random extensions that had to be disabled and enabled again to work).
One thing I miss from HTTPSB is the one click button for watching embedded youtube/instagram pics/disqus/etc.
Ah yes, I need to work on this, that was a good feature, but it needed work in order to support community-contributed presets. Probably many of the presets in HTTPSB are outdated now, and this is something I do not want the project to have to maintain in the long run. I just need to put in the infrastructure for people to import lists of presets.
Gorhill does umatrix protect against this with all privacy setting enabled: http://www.theregister.co.uk/2015/01/06/https_can_be_set_as_your_supercookie/
Disqus is starting to get tricky. I've run into sites that now need Gigya and some other things enabled in order to allow Disqus plus Disqus is adding scripts 'Disqus Referer' etc on some sites that have never been a part of it.
I don't know, I would have to read more, though I understood roughly the principle. > The page attempts to store this value in your web browser and read it again when you visit the page in the future. Really so far I do not see anything to be worried. If you can prevent anything from being stored long term, looks like this can't be used to fingerprint. If you clear the cache, a new id is generated in the proof of concept. Already with uMatrix you can have the cache cleared regularly, the cookies removed, the localStorage/sessionStorage cleared. And there is a new API coming to Chromium to clear all browser data on a per-origin basis, which will take care of all fingerprinting attempts as far as I can tell.
There is now an option in Dev 41 to disable hyperlink auditing, under chrome:flags. It may be in some other builds and I may have not noticed it.
I would like to be able to disable matrix filtering in a temporary usage for a scope but at same time having blacklisted hosts enabled. As an example this page: http://www.iltalehti.fi/iltvuutiset/201502240070504_v0.shtml A video about a family living in an old watertower. I don't like to make permanent rule allowing as much as is needed to see the video and underlying comments for this newspaper site. It turned out whitelisting blacklists was not necessary if I checked right. Anyways a whole lot of whitelisting. Usually for such media sites I have to disable matrix filtering temporary alltogether. My first choice if it was available would be to disable everything else except the hosts files blacklists instead. So this is a feature request wish.
Whitelisting the all cell certainly goes a long way for situations like this yes. For the comments under the video to be shown I myself have to whitelist the frame elements for the facebook.com 3rd party too too, afterwards. Depending how some ones have their basic setup rules even some more might be needed. Though one can then question the practicality of that setup. * * * block * * cookie block * * css allow * * doc allow * * frame block * * image allow In my case it is those global cookie and frame block rules. As I have the frames blocked, a GUI option to disable the matrix filtering for a scope but as an option leave hosts blacklisting effective, would be a better solution.
Well, so you need 3 clicks instead of 1: Whitelist the all cell, the frame column and - if really needed - the cookie column.
Just a question: Why is explicitely blocking ALL frames and allowing first party frames in the default? Using the logic of rule application leaving out these explicit frame rules would be included in the block all and allow first party. So I assume this is done for a reason. When somebody knows please explain. Regards Kees
If you whitelist the "all" cell, to keep blocking iframes is a good idea. I blocked iframe by default for those users who may want to use uBlock in allow-all/block-exceptionally mode.
Raymond, as far as I understand Kees' question, all frames on 3rd party domains are blocked by default while frames in the 1st party row are allowed. I don't remember if that's really the default setting as I modified it for myself. If it is, I also wonder if allowing frames in the 1st party row should be really the deafult behavior.
Now a seasoned uMatrix/uBlock user, when moving over from uBlock, I really liked the presets, not only for first use, but also getting to understand the rules 'logic'. Would it be possible to get them back (e.g. request policy mode, block third party scripts/iframes mode, block all, allow all)
Kees, I agree that presets would be useful for µMatrix newbies. But I can't imagine that you of all users need them Determining what you need/want and saving the appropriate rules in the global scope is easy enough, isn't it? Don't hide your light under a bushel!