Changes to Pale Moon GUID and XUL Overlay Adjustments

[mattatobin]

Greetings, My name is Matt A. Tobin and I am a contributing developer and Addons site Administrator for the Pale Moon project. I am writing this bug to inform you that the Pale Moon browser (originally and historically forked from Firefox) has changed it's GUID. Though we have made modifications to allow Firefox addons to install with the Firefox GUID this change does not cover the Application XUL Overlays that allow Adblock Plus to address many Gecko based applications.

While we are preserving the Pre-Australis UI and continuing down our road of independent development this change has partly crippled your fan favorite addon. The modifications to your install.rdf and chrome manifest files are minor and it would be appreciated if you would support Pale Moon out of the box as the growing number of users in the community using Pale Moon steadily increases.

More information can be found here: ​https://forum.palemoon.org/viewtopic.php?f=1&t=5440

If you have any additional questions please contact us either via this bug, our forums, or my self by email.

Thank you for your time and consideration!

Peace,

Matt A. Tobin

Contributing Developer / Beta Tester / 3rd Party Build Contributor / Addons Site Administrator
The Pale Moon Project

[AdIsBad]

Hello and thanks for your great work. I had to register to ask if you could please fix the icon/menu problems that appeared with today's new Pale Moon version (v25), caused by the significant divergence from Firefox.

[trev]

One problem I can see immediately: Pale Moon 25 is based on Firefox 24 ESR, support for it is going to end next Tuesday (Firefox 31 is going to be the next ESR). Is the Pale Moon project going to update to the new ESR timely? Otherwise supporting it won't help much, there are some changes blocked on Firefox 24 compatibility right now - e.g. we want to use the standardized JavaScript generators syntax, and that one isn't supported by Firefox 24.

[mattatobin]

Pale Moon /USED/ to be based on Firefox ESR24.. It has evolved past it in the last year and will continue down its divergent and individualized development separate from what MozCo happens to do.

[Sajadi]

Replying to trev:

One problem I can see immediately: Pale Moon 25 is based on Firefox 24 ESR, support for it is going to end next Tuesday (Firefox 31 is going to be the next ESR). Is the Pale Moon project going to update to the new ESR timely? Otherwise supporting it won't help much, there are some changes blocked on Firefox 24 compatibility right now - e.g. we want to use the standardized JavaScript generators syntax, and that one isn't supported by Firefox 24.

Pale Moon 24 will enhance the V24 engine as it seems. Therefor so far it seems not be planned to switch to another Browser engine version.

No chance that you still add the Pale Moon GUID into adblock plus too? There may not be full compatibility, but partly is better as nothing. Or perhaps you can give Tobin or Moonchild from Pale Moon the source code so that they are able to fork adblock plus for fully compatibility with a Pale Moon based on the ESR 24 base code?

[fanboy]

Personally I see no good reason to support an older version of a forked Firefox, it will be a security issue once Mozilla stops updating ESR24 builds.

[Sajadi]

Replying to fanboy:

Personally I see no good reason to support an older version of a forked Firefox, it will be a security issue once Mozilla stops updating ESR24 builds.

Pale Moon will not be outdated, their policy is to bring manually security patches and feature improvements to the browser, it was handled that way for several years already. So, while the source engine remains more or less static, Pale Moon is still an up to date browser.

Many people are using this fork because they do not approve the idea of Australis and bringing customization features with CTR back because they see no reason why they should fix something which should not been necessary to be fixed at all.

Hopefully you consider your opinion or at least give the sourcecode towards the Pale Moon dev, so they can create their own version of adblock plus in the future.

Thanks for reading!

Just feel free to contact Moonchild of the Pale Moon projrct or MattoTobin - these guys are responsible for Pale Moon and can give you more information and explanations!

In fact you only have to add the new unique Pale Moon GUID to your add-on's code, so Pale Moon is recognized and it can be installed there. That is more or less all, possible incompatibilities will be handled by the Pale Moon guys, so all what is asked is to "pseudo support" installation for Pale Moon.

[mattatobin]

Pale Moon's backend is also updated.. many features have been backported/rewritten for us as well as our own improvements.

I realize that you want to jump onto the new jetpack sdk train and cut support away for "older" versions of the sdk.. This is understandable but could you at least add support now before you do that so we have a good known copy not modified by us. and consider perhaps keeping a fork of the current version going with minimal updates to support Pale Moon.

You already support IE (sort of) Chrome, and Firefox why not Pale Moon as well..

All we are asking for is some consideration.. If at some point it becomes too much of a burdon to target our jetpack sdk and firefox's in the same version and a fork and continued development of the current codebase isn't going to work out either then that is something but as it stands TODAY abp is still compatible on a technical level and a simple inclusion of Pale Moon's guid and in the case statement would resolve our issues TODAY.

Thanks,

-Tobin

[Gingerbread Man]

Is this acceptable?
​https://addons.palemoon.org/extensions/privacy-and-security/adblock-plus/

My understanding of open source software and trademarks is that the source code can be used to build something else, but the product can't have the same name as the original, unless the trademark holder gives express permission.

Reminder: ticket #645.

[mattatobin]

The Pseudo-Static version was provided by us as a temporary stop-gap version modified to give our users a working solution now until we could get this ticket resolved. Unfortunately, a new version 2.6.5 was released and is still gecko/jetpack compatible with us.

We understood if further jetpack additions outside of the capabilities of our version were to happen that we would become incompatible on a technical level with ABP but there really was no harm to allow our users to have a version of ABP as long as you were still technologically compatible.

Obviously, this has not been given any further consideration due to misconceptions about what Pale Moon is and where it is going in the future.

So we are left with few alternatives. Obviously we can't keep providing Pasueo-Statics using the ABP brand. So this will require a fork of the code. I am rather disappointed at this type of discrimination to the Pale Moon project as ABP supports other browsers than Firefox. At least providing minimal edits to support Pale Moon for as long as the code base was technologically still compatible was a viable option but it is more than obvious it is considered too much of a burden for the developers who were informed of this issue weeks ago and continued to release versions from a compatible codebase.

Thank you for your time and consideration. If there is nothing else, feel free to close this ticket.

[trev]

After considering everything, Pale Moon isn't a platform we can realistically support right now. It isn't Firefox any more, rather an outdated Firefox version with a random set of changes on top. We cannot just expect Adblock Plus to work there (especially after we drop support for Firefox 24 soon), some extensive testing is required to support it - and even then we'll have to keep around way more hacks and work-arounds than reasonable. My understanding is that there is no plan to narrow the gap to current Firefox releases, the meager 439 commits in the repository don't look like even fixes for security issues are being backported, new functionality (sometimes requiring significant rearchitecturing efforts) is pretty unlikely to happen then.

While I was writing this a few more changes appeared in repository - WebRTC, parental controls and accessibility are being disabled by default now. I think I'll rather not comment on this...

Replying to Gingerbread Man:

Is this acceptable?
​https://addons.palemoon.org/extensions/privacy-and-security/adblock-plus/

No, it isn't. Publishing a modified version of Adblock Plus under the same name and without asking us (meaning: without giving us a chance to review the changes) is not something we can allow. If this modified version has issues (and it certainly will), people will blame us for that, not the Pale Moon developers - that's our reputation at stake here.

[Moonchild]

Hi Trev,

Even though this ticket is considered closed, I decided to register here to make a comment since it seems the status of the Pale Moon project is rather unclear to you.

• The "meager 439 commits" are from the point in time when the repo was first published in its current form (~6 months). That's almost 100 commits every month.
• All relevant security issues are addressed, because that is one of my top priorities for keeping Pale Moon users safe. I'm in close coop with the Mozilla Security team on relevant clearance for sec bugs.
• Significant rearchitecturing for the implementation of new features is also done regularly. So, it's not only likely to happen, it's actually being done.

So you can see Pale Moon is very much actively developed, up-to-date, and having a strong and growing presence in the browser world after ~5 years. The fact that I choose to stay away from Australis means that your current direction for your extension will have no future for Pale Moon, but with a growing user base and you being the single most-used extension globally, I would have expected more interest in your expansion into the independent browser segment by maintaining a compatible version active for products like Pale Moon (that will use a much more static (stable) extension front-end than the 6-week-carousel Firefox).

All that aside though: you can be pleased to know we've taken down the temporary version of ABP with minimal GUID targeting adjustments to "just make it work in the interim", and we will be moving on to our own independent alternative to serve our users.
I'd just like to once again express my sadness about your inflexibility in this respect and closing the door on community cooperation, but it's your show.

[trev]

Please don't get me wrong on this one, I don't mean to diminish your achievements. We aren't producing all too many commits ourselves - but we don't try to keep up with a browser vendor either. If I count it correctly, mozilla-central had 270 commits in the past 24 hours (not counting merges and various auto-generated commits). When I checked the recent backports in your repository (related to srcdoc attribute) all the corresponding bugs had target milestone set to "Firefox 25" which isn't exactly recent. Now I went further in history and there are indeed some newer bugs there - these seem to correspond to not yet published security bugs. Side-note: You might want to use less obvious commit messages for those, your repository is currently a great way to see what security bugs Mozilla hasn't fixed yet...

What about the rest of the work currently happening at Mozilla? For example, we want to switch to standardized ECMAScript generators syntax: function* foo (...) {...}. This was implemented in Firefox 26, what about Pale Moon? There is also the discussion whether we should use the array comprehension syntax that was proposed for ECMAScript 7: [for (i of numbers) i * 2] rather than [i * 2 for (i of numbers)]. The new syntax was implemented in Firefox 30, what about Pale Moon? Are changes like these on your roadmap?

I honestly don't care about the UI. Supporting a different UI isn't a big deal, we do that for a number of applications. However, we need to know that it is still Gecko underneath, and a reasonably recent version of it - it is well-documented when particular platform features were implemented or changed, and we can make informed decisions. Supporting new platforms requires an enormous testing effort, something we simply cannot do for Pale Moon right now.

As you know, Adblock Plus is open source software and you are always free to roll your own. But please make sure that you don't use our name and logo then, or ones that are confusingly similar to them. As I explained above, we aren't willing to take the blame for other people's mistakes.

[Moonchild]

You can think whatever you want about our development speed, that's your prerogative. Since pale moon doesn't try to be a swiss army knife, Phone OS, media player or skype client, and since I don't keep myself busy with busy work like renaming functions, refactoring what doesn't need to be refactored, making build system changes or messing with python, and more of that nonsense, there is no need for that kind of rapid-fire committing of code.

Yes, we've recently tackled some older outstanding bugs. Why? Because prior to our split that is the reason for your immediate v2.6 ABP incompatibility, we were tied to the MozCo versioning and core compat. So yes, we're playing catch-up for a few things. Don't think, though that we're a year behind in our development which you seem to think. Pale Moon 25 is a very far cry from Firefox 25.

So, you're going for ES6 and ES7? I hope it's worth the effort for a filtering extension :) Don't get me wrong, more power to you if it works for you and I wish you all the best in it.
A propos: JS updates are planned for Pale Moon as well and definitely on our roadmap.

As for sec bugs: No need to tell me how to describe the bugs, I'm pretty careful not to paint targets - in general, the descriptions are kept to the same level (or less descriptive) as Mozilla discloses in their tinderbox logs.

We will, indeed, be rolling our own version. Our branding (both logo and name) will be significantly different from both you and ABE so as to not cause confusion, so no worries about mistaken identity. As an aside, we'll be going back to MPL 2.0 licensing when we do.

BTW: your overzealous bayes filter is a PITA. I keep having to post in chunks.

[trev]

Replying to Moonchild:

So, you're going for ES6 and ES7? I hope it's worth the effort for a filtering extension :)

Sure. The short-term advantage is that we can use standard tools for translating to ES5 instead of JSHydra scripts. The long-term advantage is that potential contributors and developers don't need to work in a Mozilla-only silo (something that they usually don't like).

But these were only some trivial examples for changes to come. The bigger challenge is e10s - as you probably know, it was enabled by default in Firefox nightlies, so we are investigating where we need to use the message manager explicitly to get better performance. In order to avoid code duplication the message manager will be used in all applications of course. This isn't a problem for SeaMonkey or Thunderbird, they run exactly the same e10s code as Firefox even though e10s isn't enabled there. Pale Moon on the other hand is likely to still have lots of old e10s bugs.

As for sec bugs: No need to tell me how to describe the bugs, I'm pretty careful not to paint targets - in general, the descriptions are kept to the same level (or less descriptive) as Mozilla discloses in their tinderbox logs.

The problem isn't the description - it is usually pretty easy to figure out the nature of the bug from the change itself. In Mozilla's repository these changesets aren't easily found however, whereas in yours they stick out like a sore thumb.

BTW: your overzealous bayes filter is a PITA.

It certainly is :-(

[mattatobin]

Replying to trev:

As you know, Adblock Plus is open source software and you are always free to roll your own.

Done. See: ​https://addons.palemoon.org/extensions/privacy-and-security/adblock-latitude/

[philll]

Made Firefox and Firefox mobile available as seperate platforms.