Jump to content
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble
Strawberry Orange Banana Lime Leaf Slate Sky Blueberry Grape Watermelon Chocolate Marble

MSFN is made available via donations, subscriptions and advertising revenue. The use of ad-blocking software hurts the site. Please disable ad-blocking software or set an exception for MSFN. Alternatively, register and become a site sponsor/subscriber and ads will be disabled automatically. 


roytam1

My browser builds (part 2)

Recommended Posts

New regular/weekly KM-Goanna release:
https://o.rths.ml/kmeleon/KM76.2-Goanna-20191116.7z

Changelog:

Out-of-tree changes:
* update Goanna3 to git df039285e..f9e56e1ed:
- import changes from rmottola/Arctic-Fox:
 - Bug 1140767 - Build more files in security/manager in unified mode; r=dkeeler (11ab39c46)
 - Bug 1141864. Replace a bunch nsAutoPtr.h includes with nsRefPtr.h (Adapted) (ce31bfbcc)
 - Bug 1141689 - use services::GetObserverService more; r=ehsan (22e6fcf7e)
 - Bug 1140162 - IPC Proxy for TextAttributes and DefaultTextAttributes, r=tbsaunde (31bb06b0d)
 - Bug 1140499 - IPC Proxy for text/char bounds, r=tbsaunde (5441444db)
 - Bug 1140534 - IPC Proxy for offsetAtPoint, r=tbsaunde (24ca5c668)
 - Bug 1140895 - IPC Proxy for get/set/add/remove Selection, r=tbsaunde (35d3364b8)
 - Bug 1140900 - IPC Proxy for ScrollSubstringTo*, r=tbsaunde (1f7de020f)
 - Bug 1140917 - IPC Proxy for replace/insert/copy/cut/delete/paste, r=tbsaunde (b6fe2db79)
 - Bug 1140917 followup: add missing MOZ_OVERRIDE annotations on new DocAccessibleChild method-decls. r=ehsan (47cce9086)
 - Pointer style (b63b44d0a)
 - Bug 1140636 - Test CPOW function identity. r=billm (6ada9597c)
 - Bug 1134006 - Avoid IPC for domElement.QueryInterface(nsISupports) and nsIClassInfo. r=billm (753758b63)
 - Bug 1096488 - Detect and handle switching from remote to non-remote pages and back in marionette.;r=automatedtester (048279bd5)
 - Bug 1096488 - Test that switching browser remoteness leaves marionette in a usable state.;r=automatedtester (207aabadb)
 - Bug 1138650 - Update remaining callsites to use newChannel2 in toolkit/devtools (r=jryans) (a4ffc704e)
 - Bug 1138648 - Update remaining callsites to use newChannel2 in netwerk/ (r=sworkman) (cdf6612a9)
 - bug 1135160 - implement link rel=preconnect r=smaug (cfac502ce)
 - bug 1135160 - ioservice have speculative connect use proxy-resolve2() r=hurley (238b58f84)
 - Bug 1140788 - Set headers to immutable. r=bkelly,ehsan (c48c12acf)
 - Bug 1137037 - Determine the inner window ID in imgRequest::Init. r=baku (12aa73a7c)
 - Bug 1137019 (Part 1) - Get rid of unused LockImage forwarding methods on imgRequest. r=baku (6ed5c7d25)
 - Bug 1137019 (Part 2) - Replace imgRequest's image decoding methods with a single minimal method that updates an atomic. r=baku (adeb8797c) (f9e56e1ed)

* Notice: the changelog above may not always applicable to XULRunner code which K-Meleon uses.

A goanna3 source tree that has kmeleon adaption patch applied is available here: https://github.com/roytam1/palemoon27/tree/kmeleon76

Edited by roytam1
  • Like 2

Share this post


Link to post
Share on other sites

rmottola upstream pushed lots of commits of front-end side, so some merge commits issued and new test build is uploaded: http://o.rths.ml/gpc/files1.rt/arcticfox-27.9.18.win32-20191116.7z

- Bug 1140767 - Build more files in security/manager in unified mode; r=dkeeler (11ab39c46)
- Bug 1141864. Replace a bunch nsAutoPtr.h includes with nsRefPtr.h (Adapted) (ce31bfbcc)
- Bug 1141689 - use services::GetObserverService more; r=ehsan (22e6fcf7e)
- update (7d809f4f9)
- update (985adbc89)
- Bug 1140162 - IPC Proxy for TextAttributes and DefaultTextAttributes, r=tbsaunde (31bb06b0d)
- Bug 1140499 - IPC Proxy for text/char bounds, r=tbsaunde (5441444db)
- Bug 1140534 - IPC Proxy for offsetAtPoint, r=tbsaunde (24ca5c668)
- Bug 1140895 - IPC Proxy for get/set/add/remove Selection, r=tbsaunde (35d3364b8)
- Bug 1140900 - IPC Proxy for ScrollSubstringTo*, r=tbsaunde (1f7de020f)
- Bug 1140917 - IPC Proxy for replace/insert/copy/cut/delete/paste, r=tbsaunde (b6fe2db79)
- Bug 1140917 followup: add missing MOZ_OVERRIDE annotations on new DocAccessibleChild method-decls. r=ehsan (47cce9086)
- Pointer style (b63b44d0a)
- Bug 1140636 - Test CPOW function identity. r=billm (6ada9597c)
- Bug 1134006 - Avoid IPC for domElement.QueryInterface(nsISupports) and nsIClassInfo. r=billm (753758b63)
- Bug 1096488 - Detect and handle switching from remote to non-remote pages and back in marionette.;r=automatedtester (048279bd5)
- Bug 1096488 - Test that switching browser remoteness leaves marionette in a usable state.;r=automatedtester (207aabadb)
- update (f3806a904)
- Bug 1138650 - Update remaining callsites to use newChannel2 in toolkit/devtools (r=jryans) (a4ffc704e)
- Bug 1138648 - Update remaining callsites to use newChannel2 in netwerk/ (r=sworkman) (cdf6612a9)
- bug 1135160 - implement link rel=preconnect r=smaug (cfac502ce)
- bug 1135160 - ioservice have speculative connect use proxy-resolve2() r=hurley (238b58f84)
- Bug 929097 - Check window type in SessionStore.onOpen() to ignore; r=smacleod (1c3a5c781)
- Bug 1112733 - Clear tabData storage after calling restore. r=billm (6015a4a40)
- Bug 1057481 - Session restore shouldn't use contentWindow (r=ttaubert) (b585d3906)
- missing cleanup of Bug 887780 (568971045)
- update (0979339c3)
- Bug 1020831 - Make SessionStore.jsm and SessionFile.jsm share the same run state r=yoric (d48f01b12)
- Bug 847955 - Let _getMostRecentBrowserWindow utilize RecentWindow.getMostRecentBrowserWindow. r=ttaubert (58b072a89)
- update (0c56a5bba)
- Bug 1140788 - Set headers to immutable. r=bkelly,ehsan (c48c12acf)
- Bug 1137037 - Determine the inner window ID in imgRequest::Init. r=baku (12aa73a7c)
- Bug 1137019 (Part 1) - Get rid of unused LockImage forwarding methods on imgRequest. r=baku (6ed5c7d25)
- Bug 1137019 (Part 2) - Replace imgRequest's image decoding methods with a single minimal method that updates an atomic. r=baku (adeb8797c)
- update (3f7d12c98)
- Merge remote-tracking branch 'upstream/master' into fix-winbuild (9a5b558d3)

  • Like 1

Share this post


Link to post
Share on other sites

roytam1 said:


Notice: And I have no time(~5 hours more time on transportation) on porting mailnews' upstream's changes to local tree, so some components in mailnews will not be same status as upstream.



Thank you for all that incredible work you do! Even in peace times I cannot imagine how in the world you manage to keep up that enormous workload, looking at those endless long changelogs for several different browsers, even incredible if someone could work 24h per day on them. And recently even discovered, completely shocked, that even 1 single commit, 1 little changelog line, can consist of 800 changed files! Let alone now in current semi-war conditions, with a lot greater sorrows as update schedules, and much less time too. Once more, I'm stunned. But really much more important now is that you hopefully remain as safe as possible. My thoughs are with all the amazing people in your country.
  • Like 5

Share this post


Link to post
Share on other sites
3 minutes ago, siria said:

But really much more important now is that you hopefully remain as safe as possible. My thoughts are with all the amazing people in your country.

+1 :)

Share this post


Link to post
Share on other sites
10 hours ago, VistaLover said:

Sadly, wrong filenames and links, that fetch last Saturday's builds... :(

fixed

  • Like 2

Share this post


Link to post
Share on other sites
10 hours ago, siria said:

But really much more important now is that you hopefully remain as safe as possible. My thoughs are with all the amazing people in your country.

Yeah I'm fine right now. No one knows what future will be, so "carpe diem". Just bought Nintendo Switch and Pokemon Sword and have some fun for now :)

Edited by roytam1
  • Like 1

Share this post


Link to post
Share on other sites
25 minutes ago, roytam1 said:

Yeah I'm fine right now. No one knows what future will be, so "carpe diem". Just bought Nintendo Switch and Pokemon Sword and have some fun for now :)

offtopic, but heard that some funky things happen between Sword/Shield (seems that it happens in the eShop version) and how your SD card is formatted. If it's exFAT, expect some nasty data corruption in the card, though saves should be fine since it's saved in the internal memory (worst case scenario you would need to redownload games and other content that were in the SD card)

https://www.gamesradar.com/pokemon-sword-and-shield-are-corrupting-micro-sd-cards-but-your-save-files-are-safe/

Edited by IntMD

Share this post


Link to post
Share on other sites
1 hour ago, IntMD said:

offtopic, but heard that some funky things happen between Sword/Shield (seems that it happens in the eShop version) and how your SD card is formatted. If it's exFAT, expect some nasty data corruption in the card, though saves should be fine since it's saved in the internal memory (worst case scenario you would need to redownload games and other content that were in the SD card)

https://www.gamesradar.com/pokemon-sword-and-shield-are-corrupting-micro-sd-cards-but-your-save-files-are-safe/

I bought physical game card and don't have any SD card inserted so it should be fine.

Edited by roytam1

Share this post


Link to post
Share on other sites
22 hours ago, Mathwiz said:

Wow - it turns out that privacy.resistFingerprinting does a lot more besides what canvas.poisondata does! The full list of what it's supposed to do is in the spoiler. Most of this was not implemented in Serpent 55, and I'm unsure which of these functions interferes with Flash, but something does:

  Reveal hidden contents
  • Complicated (see below)
    • Canvas image extraction is blocked
    • Absolute Screen Coordinates are obscured
    • Window Dimensions are rounded to a multiple of 200x100, and a warning is shown when maximizing
    • We only allow specific system fonts to be used, and we ship them to the user using kinto
  • Non-Trivial (see below)
    • The performance API is mostly disabled
    • Time Precision is reduced to 100ms, with up to 100ms of jitter
    • mozAddonManager may be blocked bug 1384330
    • Media Devices are spoofed bug 1372073
    • WebGL is limited bug 1217290
    • The Keyboard Layout is spoofed
    • The Locale is spoofed to en-US
    • The Date Input Field and Date Picker Panel are spoofed to en-US bug 1492587
    • If you customize the preferred language list (Accept-Language), you will be warned bug 1039069
    • System Media Queries will never match bug 1479240
    • The Pointer Event is spoofed bug 1363508 and also pointerEvent.pointerid bug 1492766
  • Trivial
    • The browser version is reported to be the most recent ESR version (but the OS is not spoofed)
    • Timezone is spoofed to 'UTC'
    • The gamepad API is disabled
    • All device sensors are disabled
    • The WebSpeech API is disabled
    • WEBGL_debug_renderer_info extension is disabled bug 1337157
    • navigator.hardwareConcurrency is spoofed to 2
    • Site-specific zoom is disabled bug 1369357
    • MediaError.message is restricted to a whitelist bug 1354633
    • The Network Information API reports an 'Unknown' connection type, and the ontypechange event is suppressed bug 1372072
    • The Media Statistics API will report calculated numbers not reflecting reality bug 1369309
    • Web Extensions are able to toggle privacy.resistFingerprinting
    • Geolocation is disabled bug 1372069 - but this will be reverted bug 1441295
    • screen.orientation.type is spoofed as 'landscape-primary' and screen.orientation.angle is spoofed to '0' bug 1281949 but also bug 1433815
    • navigator.plugins and navigator.mimeTypes are reported as empty bug 1281963 and bug 1324044
    • prefers-reduced-motion always returns false bug 1478158
    • AudioContext OutputLatency is spoofed bug 1564422

Both privacy.resistFingerprinting and canvas.poisondata are not enough to get a good browser setting for privacy.

Share this post


Link to post
Share on other sites

Totally agree. We could have a whole separate thread on setting up browsers for privacy. The topic goes far beyond those two prefs.

Remember, this topic came up because the UOC Patch sets privacy.resistFingerprinting for performance reasons, but it comes at the cost of compatibility with Flash.

Anyone concerned with privacy is probably very reluctant to use Flash to start with, so that incompatibility may not bother them. But if they're installing the UOC Patch for performance, they may still want to use Flash, and need to know about the incompatibility.

I don't know if it helps performance, but canvas.poisondata doesn't cause Flash compatibility issues. I only mentioned it because it does one of the many things that privacy.resistFingerprinting does.

  • Like 1

Share this post


Link to post
Share on other sites
On 11/13/2019 at 12:26 PM, VistaLover said:

Mozilla had intentionally removed the default display of AVN (addon version number) in the Australis[later WebExAM] AOM back in Firefox v40, when Bugzilla bug1161183 landed:

https://bugzilla.mozilla.org/show_bug.cgi?id=1161183

This was again an unnecessary move, a type of "chop head to get rid of headache" approach, since the bug originally wasn't about AVN per se...

So, I finally got around to looking at this, with an eye toward bringing Serpent's AM display back in line with official Basilisk (without disabling WE compatibility). Shouldn't be hard (as long as I don't make any mistakes!), but I have to say Mozilla really made a mess of this.

The screen shot at the link tells the story: Mozilla redesigned the AM display after FF 38, I guess as part of the whole Australis thing. They chose a more attractive color scheme but also a layout that wastes lots more space on the screen! I guess the devs didn't understand that the PC universe hadn't moved entirely to big 24-inch (61-cm) monitors back in 2014! (Devs are so spoiled sometimes.)

Of course they decided the problem couldn't be with their new layout; instead, it must be that the user was just getting too much information, so they moved the version number to a "tooltip;" with the "fix" above, the version appears when you "hover" over the add-on's name with the mouse pointer (as well as on the "More" page).

Then, to compound the error, rather than simply conditioning the version number display on a new pref, they ripped out most of the code needed to display it, meaning you couldn't bring it back without an extension like CTR.

BTW, CTR has several options to help the version number fit on a smaller screen: there's "alternative appearance," which brings back the pre-Australis look (unfortunately it also brings back the old color scheme, but nothing's perfect, I guess); there's "replace button labels with icons," which makes the right-side buttons less enormous, and there's "compact view," which makes everything smaller, although that one's a mixed bag (looks pretty good on extensions, but on the plug-ins page, where the version numbers are typically short but the names and descriptions are long, you really need to stick with a 2-line format). Anyway, the point is there were plenty of ways to fix this without taking info off the display!

  • Like 1

Share this post


Link to post
Share on other sites
3 hours ago, Mathwiz said:

So, I finally got around to looking at this, with an eye toward bringing Serpent's AM display back in line with official Basilisk (without disabling WE compatibility). Shouldn't be hard (as long as I don't make any mistakes!), but I have to say Mozilla really made a mess of this.

merged, thanks.

Share this post


Link to post
Share on other sites
On 11/17/2019 at 1:18 AM, roytam1 said:

merged, thanks.

@roytam1, @Mathwiz : How is that guaranteed to only touch Serpent 52's AOM (WebExtAM) but NOT NM28's AOM (TychoAM) ? On GitHub, in the PR's description, it says

Quote

New pref will be for both Serpent and New Moon.

But NM does not need that pref, OR the functionality of that PR :no: ... Also,

Quote

This is a more complex change than usual; please review my work for errors before merging!

Was that actually carefully examined (... I mean, the impact it has on the rest of the UXP-based apps) ? :dubbio:

If I am to voice my personal humble opinion, this whole issue was rushed... Also, I'm a firm believer in the adage: "Don't fix it if it ain't broken!" and, as of this writing, NOTHING IS BROKEN (yet) ...

To re-emphasise what I wrote in my previous posts:

1. Serpent 52.9.0/UXP and Serpent 55.0/Moebius both have a WebExAM type of AOM (St52 one derived from FxESR52, St55 one derived from Fx53.0a1 - probably not identical, but with very few differences between them...)
2. CTR v1.7.8 serves both above browsers perfectly, and AVN is offered as an option (OFF by default); also, CTR is incompatible with NM, so this leaves NM totally out of this discussion. 
3. CTR v1.7.8 won't break in St52 in the future, unless @roytam1 explicitly touches Serpent's AOM code (not likely)
4. The upstream team have implemented the TychoAM on both Pale Moon and Basilisk; I'd speculate that they are unlikely to change TychoAM any further, but you can never really be sure with them... In the event they do change it, it's probable (from my viewpoint) that those changes won't be applicable to St52's WebExAM and would have to be reverted by Roy (but somehow applied selectively to NM28?)...

The CTR v1.7.8.2019 branch was created by Aris with only classic Waterfox (Firefox 56 derived, but heavily forked...) and official Basilisk in mind, but after official Basilisk had changed to the TychoAM; if upstream implement changes to TychoAM, then CTR v1.7.8.2019 will have to accommodate, of course, but with code not relevant to Serpent's AOM... There's absolutely no compelling reason to install CTR v1.7.8.2019 on Serpent 52 currently (... just because you can is not a valid argument in my book ;) )...
5. If, despite all the above, one insists on installing CTR v1.7.8.2019 in St52, then, as posted, to get the AVN feature back (present in CTR 1.7.8) just co-install either of

Version Number in Add-ons Manager 1.10 (by magicp)
caa:addon/addonvernumber/versions?page=1#version-1.10

Add-ons Manager - Version Number 1.4 (by Aris)
caa:addon/amversionnumber/versions?page=1#version-1.4
NB: Latest v1.5 not compatible with Serpent (it installs, but BREAKS the AOM!)

Of course, you can install either one without any version of CTR, if you JUST want the feature of AVN in St52's AOM... Thus, no need at present to touch the UXP platform code...
In all fairness, MCP might in the future do something more radical in Basilisk's GUI, outside its TychoAM; if that big change lands in Serpent 52, then CTR 1.7.8 might get (partially) broken, and an update to a future version of 1.7.8.20xx might be needed; even then, CTR 1.7.8.20xx + one of the above extensions should suffice (but let's cross that bridge when we get to it...).

In closing, I thought I'd detail my personal, lukewarm, view on that PR, since it was I that instigated this chain of events (after pointing out to @Mathwiz the existence of CTR versions past 1.7.7.2 and providing a link to the Wf PR that implemented AVN natively in its - Fx56 based - AOM) ; and to clear out  any trace of a possible misunderstanding, this has nothing to do with the author of the PR; I highly appreciate all his efforts (coding and otherwise) and contributions offered to this great community here, not to mention the outstanding quality of his posts; plus, I sort of think of him as a good friend (especially since we've conversed in the past over PMs); so, @Mathwiz, nothing personal here :)

Best regards!

Edited by VistaLover

Share this post


Link to post
Share on other sites

Newmoon 27 SSE backed up prefs only AFTER patching.... was in Newmoon Productions-folder not Newmoon.
FF 45ESR SSE
UOC-Patch and Enforcer installed. Cannot install Browser- apps. running from folder.
Newmoon and FB working much smoother than with Firefox 45ESR. all advised addons installed. More later.
KM Goanna 76.2
Cannot patch /enforce in advised location since in dont see these: have prefs.js at different location. see further down.
"For K-Meleon users, check out Mathwiz's post below: .
One little correction: the UOC Patch for FF 38-based browsers also applies to the K-Meleon browser, but you have to extract it into a different folder. Instead of <browser folder>\defaults\pref, for K-Meleon you should extract the patch into <browser folder>\browser\defaults\preferences."
True for me: ....\browser folder>\browser\profiles\xyz.default\ 

-patching/enforcing done. see image at the end.
user.js backed up will be replaced. prefs.js sitting beside UOC-patch.
Browserinstaller not working. Only palemoon-loader is working. will leave older version alone if stable since have to adjust this or that. in about:config and settings.

1a). Firefox 45.9.17 Nightly: cannot find german languagefile. Installation possible?

1b) KMeleon: Already asked but no answer: How to find out which version is running? From which date?
2. I am about to install all these advised Addons:
NoScript (I recommend v2.6.9.32 or v2.6.9.27 for New Moon, and v2.9.0.9 for Firefox 45 ESR SSE) Using 5.1.8.4. maybe better using the advised ones or checking how they behave? how stable system is?
    Decentraleyes 1.4.2
    Bluhell Firewall 2.5.3
    UAControl 0.1.3.1.1 (To change the user agent on the fly, per website)
    uBlock Origin 1.10.0
    Vacuum Places Improved 1.2.1

- Bluhell Firewall 2.5.3 cannot be found only newer and older.
Can anyone provide it here pls?

will then add direkt-links of the other advised addons.
-------
3. Browserinstallation never worked here: maybe not enough space only 6gb left in systempartition. see image

 

BROWSERINSTALLER NOT WORKING.png

kmgonna_uoc_patch_enforcer.png

Edited by 3dreal
FF 45ESR -45.9.17 Nightly confusion. Which KMeleon version?

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...