Page 2 of 3

Re: Google Trends

Posted: Wed Jul 29, 2015 10:18 pm
by cartel
Got some results back thanks to intofix over at Palemoon forum.

Firefox with NoScript v2.6.9.33 (all scripts allowed) and Ublock Origin turned off/on, working.
Pale Moon with NoScript v2.6.9.33 (all scripts allowed) and Ublock Origin turned off/on, not working.

Edit: with Firefox it works even with Google Analytics not allowed.


Seems to be a change from 2.6.9.31 that affects Palemoon.

I'd like to ask if the developers could please have a look for us using Palemoon...with sugar on top.

Re: Google Trends

Posted: Wed Jul 29, 2015 10:45 pm
by barbaz
https://forum.palemoon.org/viewtopic.php?f=3&t=8986
https://forum.palemoon.org/viewtopic.php?f=46&t=8987

Can you please ask them which Firefox version is closest to PaleMoon 25.6.0, since you don't seem to know?

As far as I can recall, all issues affecting PaleMoon have also affected legacy Firefox - so if we know which old Firefox version to check against, and if this issue shows up there, then Giorgio can test and fix it. If it doesn't show up in that version of Firefox... well, you first have to persuade Giorgio to support PaleMoon.

I'm sorry to have to be like this with you, because I also like PaleMoon, but Giorgio is the only developer of NoScript, and supporting another browser is a major effort as it means yet another round of testing he has to do... for basically every supported version of the browser. :|

Re: Google Trends

Posted: Wed Jul 29, 2015 11:39 pm
by cartel
For extension compatibility it would be closest to Firefox 24 ESR.

Re: Google Trends

Posted: Thu Jul 30, 2015 12:05 am
by barbaz
Thanks, I'll test Fx 24 in a few hours (no time atm).

Re: Google Trends

Posted: Thu Jul 30, 2015 3:07 am
by barbaz
Actually, that page WFM in Fx 24.8.1 w/ NoScript 2.6.9.33rc2 after allowing google-analytics (just takes ~30 sec or so to load).
Same deal with PaleMoon 25.5.0 (no 25.6.0 for Linux, sorry).

Starting from clean profile with only NoScript installed, in both cases.

So yeah, it's a surrogate issue - surrogates are broken in legacy Firefox as well as PaleMoon. Either they're not loading or they're not running, can't tell because there are no message in the Browser Console (Ctrl-Shift-J) indicating the problem.
(Turning on noscript.surrogate.debug seems to have no effect.)

Moving to NoScript Development and giving this a more descriptive title. Thanks for the report

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Thu Jul 30, 2015 1:49 pm
by cartel
Image

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Fri Jul 31, 2015 1:18 am
by imacomputer
hello thank you as well, i posted on the thread on the palemoon site and found this page as well

i encountered similar issues with google-analytics blocking on this page
https://groups.google.com/forum/#!forum/alt.games.sf2
it's just completely white

also on some other sites i had some issues, i hope this is resolved soon, i really appreciate any help!

Re: Google Trends

Posted: Sat Aug 01, 2015 4:51 am
by barbaz
barbaz wrote:(no 25.6.0 for Linux, sorry)

Apparently not true Image - actually, they're moving away from Sourceforge and putting new Linux PaleMoon builds on their own Cloudflare-based site, and I just haven't been following the announcements.

Anyway, now that I've found PaleMoon 25.6.0 builds for Linux, is it worth me testing this with PaleMoon 25.6.0?

Re: Google Trends

Posted: Wed Aug 05, 2015 3:10 am
by barbaz
cartel wrote: 2.6.9.32rc1 yes broken in this situation.

Now this is odd... the diff between (working) 2.6.9.31rc1 and (broken) 2.6.9.32rc1 is (aside from version number differences) just this:

Code: Select all

diff -r ./noscript-2.6.9.31rc1_xpi_files/components/noscriptService.js ./noscript-2.6.9.32rc1_xpi_files/components/noscriptService.js
4052c4052
<       if (!sources[0]) return 0;
---
>       if (sources.length === 0) return 0;


It makes no sense to me that that change would break anything. I think I need to re-test a few things...

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Wed Aug 05, 2015 3:53 am
by barbaz
#OK, the surrogate system is not completely dead... there's something else weird going on with the specific surrogate I was testing against (the OWASP anti-clickjacking surrogate is not working on the Adobe site in PaleMoon, but that's a different story).
What's NOT fine is that on PaleMoon, the $S() function is returning 0 and not an object.

As such I would expect that some surrogates do still work... but I'm seeing a second bug: surrogates whose sources have no prefix don't ever run if the scripts they're replacing are 1st-party (e.g. trying to replace a palemoon.org JS file while on palemoon.org).
I don't have time to track down exactly what surrogate functionality works and what doesn't atm, but it seems it's far less clear cut that I had originally thought.

Testing with NoScript 2.6.9.34rc2 in a clean profile in PaleMoon 25.6 (because my main profile has a bunch of add-ons I haven't yet fixed to work in PaleMoon 25.x and I want to not have malfunctioning other extensions in my testing).
(What a pleasant UX to use a browser with a friendly, easy-to-navigate UI and a nice set of builtin dev-tools; even better is that it's "modern".)

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Fri Aug 07, 2015 8:23 am
by cartel
Thanks for the update.
I posted the info on the PM forum to see if it rings any bells.

https://forum.palemoon.org/viewtopic.php?f=3&t=8986

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Fri Aug 07, 2015 1:52 pm
by barbaz
(direct link to the post: https://forum.palemoon.org/viewtopic.php?p=60839#p60839)

Thanks cartel, hopefully they can confirm my findings and/or give some insight into what might be causing this weirdness.
(For whatever reason I tend to run into some of the strangest computer issues which most of the time are unique to me... hopefully my testing wasn't hitting one of those.)

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Sun Aug 09, 2015 5:54 pm
by barbaz
@cartel: I think I'm somewhat in over my head with Moonchild's response to your post... I think only Giorgio can answer his question re: whether there's something they can implement in PaleMoon itself to fix this. Given that it does work in modern Gecko, naively I'd say the answer is yes but I wouldn't know where to start looking; and I don't think that the $S() function is the only thing that's broken about the surrogates system (in PaleMoon) right now...

Anyway, since he's specifically interested in fixing the $S() issue, and if he's not watching this thread himself, you can tell him that the related code is in

Code: Select all

chrome://noscript/content/ScriptSurrogate.js
in the _preamble function, and the point of it & $S is to provide a shim for __noSuchMethod__ using Proxy, on all browsers that support Proxy & the "new Proxy()" syntax. (_preamble injects the $S into the surrogate scope.) Maybe someone on the PaleMoon forums can boil it down to a test case that Moonchild can use, given this information (I'd do it myself with some guidance but I really don't have much time atm).

EDIT I can try to find the latest broken Gecko version & cross compare each w/ PaleMoon 25.6, if that will help?

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Sun Aug 09, 2015 7:18 pm
by Giorgio Maone
I'm gonna "fix" this by using the shim only on Gecko versions >= 38. Any objection?

Re: Surrogates not working in PaleMoon and legacy Fx

Posted: Sun Aug 09, 2015 7:56 pm
by barbaz
Meaning what? That you'll replace the feature detection of Proxy with an internal Gecko version check?

If so, no objections from me assuming that PaleMoon isn't going to remove __noSuchMethod__.
@cartel, can you please confirm with the PaleMoon forums that my assumption is correct? Searching there for "__noSuchMethod__" turned up 0 results...