"Sumozi" - A Variant to solve those age old issues

[Matt A. Tobin]

Greetings, as some may remember there was a number of misconceptions about the Pale Moon project and a number of points of contention regarding what Moonchild decided to not include in the main-line binaries.

To solve these issues I have worked with Moonchild to arrange a fully endorsed 3rd party contrib version of Pale Moon I am calling "Sumozi". This build compiles from otherwise unaltered pale moon source code and enables the following features via build options and by default.

The features re-enabled via mozconfig build options are:
Accessablity features
Parental controls
WebRTC

The features enabled via default preferences are:
Integrated PDF Reader (PDFjs)
SocialAPI

For clarification the ONLY feature that is not included is the Tab Groups which was actually removed and turned into an extension due to it's immature status and low usership. However, that can downloaded and installed from the Pale Moon website.

Also as an outgrowth of another contributed build I am preparing to release is a Windows XP x64 Edition (Server 2003) specific build of "Sumozi".

For information and downloads of this please see the related post on the Pale Moon forum @ http://forum.palemoon.org/viewtopic.php?f=40&t=4940
-AND-
The main Project page over on Binary Outcast @ http://binaryoutcast.com/software/projects/sumozi/

Remember: This variant of Pale Moon only really satisfies the edge case of needing those features that are disabled from the main-line build and if you otherwise do not have a use for them I encourage you to use the normal version of Pale Moon.

Please direct any support requests or issues to the Pale Moon forum. Mozillazine does not tend to offer (as is proper) any support for non-mozilla products.

Peace,

Matt A. Tobin
<Self important title />
Binary Outcast

[xunxun1982]

My suggest is that if you want to build a classic theme firefox at present, don't stick in 24.x ESR or the similar code, please try to revert the ux code in newer source code after 29.x, or try to update holly channel code to the present code manually, this is what the 3rd custom builders that want the native classic theme should do.
If the reverting or holly updating is successful, many people who like classic theme will enjoy it very much.

[LoudNoise]

And you will find that there isn't the slightest speed impact for compiling with the accessibility tools allowed. A number of third party builds turn this stuff off. I have never understood why or why they go out of their why to mention this. They seem to take a great deal of pride in the fact that they don't offer support to the handicapped. Creepy.

[SnoutSpout]

It's mostly confirmation bias with regards to the performance gains of such efforts with their users.

Besides making PGO more aggressive, most third-party build maintainers simply don't have the coding chops to hack the browser into a "faster" state. They would be moving into "real fork" territory. And besides, any memory savings they get are totally inconsequential compared to the effects of websites getting more and more bloated with each passing year.

[LoudNoise]

It isn't an issue of memory usage.

[Trippynet]

websites getting more and more bloated with each passing year.

LN is right, it's mainly sticking more and more Javascript onto the pages that slows things down. If I try to browse the web on my old SGI Fuel (600MHz CPU, 4GB RAM, Firefox 3.0.0.19), it's amazing just how slowly modern pages render on it due to the amount of Javascript. If I try the same thing on an older/more simple page, or I disable Javascript with "NoScript", it can render the pages quite quickly in comparison.

[Matt A. Tobin]

And you will find that there isn't the slightest speed impact for compiling with the accessibility tools allowed. A number of third party builds turn this stuff off. I have never understood why or why they go out of their why to mention this. They seem to take a great deal of pride in the fact that they don't offer support to the handicapped. Creepy.

Honestly things like Accessibility are not needed for MOST people and there are impacts on speed and efficiency when including VS excluding parts of code. HOWEVER, this and your disapproval is now moot with the advent of my "Sumozi" variant of Pale Moon.

It's mostly confirmation bias with regards to the performance gains of such efforts with their users.

Besides making PGO more aggressive, most third-party build maintainers simply don't have the coding chops to hack the browser into a "faster" state. They would be moving into "real fork" territory. And besides, any memory savings they get are totally inconsequential compared to the effects of websites getting more and more bloated with each passing year.

PGO is outdated and not really recommended anymore on VS2012+ as the benefits don't outweigh the possible issues now. Which is why Moonchild doesn't use it anymore. Even without PGO Pale Moon happens to still be faster than Firefox.

Also while my 3rd party contrib would technically be considered build maintainer status. Mainline Pale Moon has diverged considerably from it's ESR24 base. May I direct you to the main repo for Pale Moon on GitHub @ https://github.com/MoonchildProductions/Pale-Moon/commits. As always you may want to consider running a Diff between it and the current version of ESR24 from MozCo.

[xunxun1982]

And you will find that there isn't the slightest speed impact for compiling with the accessibility tools allowed. A number of third party builds turn this stuff off. I have never understood why or why they go out of their why to mention this. They seem to take a great deal of pride in the fact that they don't offer support to the handicapped. Creepy.

In fact, because of mozilla build issues, if only using configure option, you can't remove all Accessablity features and Parental controls, so removing these functions without modifying mozilla build will not decrease the memory use and will cause some other issues.

[xunxun1982]

And you will find that there isn't the slightest speed impact for compiling with the accessibility tools allowed. A number of third party builds turn this stuff off. I have never understood why or why they go out of their why to mention this. They seem to take a great deal of pride in the fact that they don't offer support to the handicapped. Creepy.

Honestly things like Accessibility are not needed for MOST people and there are impacts on speed and efficiency when including VS excluding parts of code. HOWEVER, this and your disapproval is now moot with the advent of my "Sumozi" variant of Pale Moon.

It's mostly confirmation bias with regards to the performance gains of such efforts with their users.

Besides making PGO more aggressive, most third-party build maintainers simply don't have the coding chops to hack the browser into a "faster" state. They would be moving into "real fork" territory. And besides, any memory savings they get are totally inconsequential compared to the effects of websites getting more and more bloated with each passing year.

PGO is outdated and not really recommended anymore on VS2012+ as the benefits don't outweigh the possible issues now. Which is why Moonchild doesn't use it anymore. Even without PGO Pale Moon happens to still be faster than Firefox.

Also while my 3rd party contrib would technically be considered build maintainer status. Mainline Pale Moon has diverged considerably from it's ESR24 base. May I direct you to the main repo for Pale Moon on GitHub @ https://github.com/MoonchildProductions/Pale-Moon/commits. As always you may want to consider running a Diff between it and the current version of ESR24 from MozCo.

Try to use VC2013 update2, the edition's PGO is faster than VC2010 PGO.

And after many people's tests, Palemoon is the slowest 3rd custom build ( even much slower than official edition ) besides the same edition, which is not doubted.

[Matt A. Tobin]

We currently use MSVC2012 update 4. We also do not use PGO. While "Sumozi" is slighly slower than Mainline Pale Moon it is hardly noticeable and still remains faster than even the latest Firefox. So if you wish to contradict that claim, Please provide your data and results. xunxun

[Frank Lion]

...So if you wish to contradict that claim, Please provide your data and results. xunxun

Shouldn't that be the other way around?

Like, if you are making a claim, then it is you that provides the data and results to support that claim?

[Matt A. Tobin]

Nah, not really. I am not here to debate Pale Moon over anything. Frankly none of these reflections are in-scope with the orginal intent of this thread. I just wanted to make sure that the edge cases were covered in order to make sure any possible user of Pale Moon could have what they needed. Be it Accessibility features or WebRTC.

It is up to what your usecase requires. Some like Post-Australis Firefox or Chrome. Some like the old communicator UI of Seamonkey, and some like the open, choice filled, and complete extensibility of Pale Moon. At the end of the day no matter what synthetic benchmarks or apparent speed by compiler optimizations you use it won't matter if the product does not fit your needs or provide what you expect.

[Frank Lion]

Nah, not really. I am not here to debate Pale Moon over anything..

Trouble is, Matt, after that last post of yours, you don't really have much choice, do you?

I mean, when you wrote this, effectively, 'put up or shut up' post to xunxun, did you care if they were here to debate Palemoon or anything? -

... So if you wish to contradict that claim, Please provide your data and results. xunxun

I then just pointed out that the burden of proof does actually lie with you, as the instigator of the claim. That's not something I've just dreamt up, that's how the entire scientific community works and really, this being a technical subject and all, this falls under that.

[LoudNoise]

Question. Is "Mozilla/4.79 [en] (Win95; N)" the useragent for Sumozi? If it is you are going to spend an endless amount of time dealing with this issue and all of your users will be easily identified.

[xunxun1982]

We currently use MSVC2012 update 4. We also do not use PGO. While "Sumozi" is slighly slower than Mainline Pale Moon it is hardly noticeable and still remains faster than even the latest Firefox. So if you wish to contradict that claim, Please provide your data and results. xunxun

Don't compare it with the latest firefox, because of some reasons, latest firefox will be slower than 24.x official edition in some benchmark test ( some perf drop, especially in Peacekeeper, may be fixed some time ).

I take the same edition ( 24.x ) in the tests on Win8 OS:

Sunspider
official 171.1ms +/- 1.3%
Palemoon 191.2ms +/- 1.0%
Sumozi 194.0ms +/- 0.8%
tete009 179.6ms +/- 1.3%
pcx 171.9ms +/- 1.3%

Octane
official 15760
Palemoon 15612
Sumozi 15476
tete009 16806
pcx 16618

CanvasMark 2013
official 4623
Palemoon 4543
Sumozi 4584
tete009 4786
pcx 4836