SRWare 'Iron' browser; 64-bit SFS & .pet packages

[Mike Walsh]

***New Version: Iron 60.0.3150.1***

Evening, all.

The new release of Iron, 60.0.3150.1, is now available for download from the link in post #1. I've included the current version of PepperFlash, 26.0.0.151, which installs to /usr/lib/mozilla/plugins, as usual.

----------------------------------------------------------------------------------------------

The Chrome 'upgrades' have once again carried over to Iron, since it's based very closely on it.....including the 'white on black' tooltips, which I'm still not too sure about! And the lack of highlighting in the bookmarks listings...

NetFlix is working, thanks once again to following SRWare's advice to 'poach' the libwidevinecdm.so module from the current version of Chrome itself.

-----------------------------------------------------------------------------------------------

NOTICE

I'm going to be concentrating on the 64-bit packages of SlimJet and Iron from now on; my 'run' with the Chrome64 packages is coming to an end. The newest release of Chrome, 61, now requires the very newest version of Network Security Services (libnss.so.3), which Tahr 64 doesn't support, since the new libnss has to have a newer version of the glibc.

I've tried XenialPup64 on the big Compaq, and it just will not run correctly. Unless I can find another suitable 64-bit Pup, which isn't based on Tahr, that's it. Slacko64 is out, since Chrome hasn't run properly on there for a while now. And to the best of my knowledge, all 64-bit Pups are based on one or the other. My deepest apologies to any Chrome fans.

So; enjoy Iron, instead!


Mike.

[orrin]

Hi Mike,

Iron 60.0.3150.1 does not start in Slacko 64-6.3.2.

Maybe a missing library?

It seems that anything based on Chrome 60 or above, loads but will not start!

[Mike Walsh]

Hi Mike,

Iron 60.0.3150.1 does not start in Slacko 64-6.3.2.

Maybe a missing library?

It seems that anything based on Chrome 60 or above, loads but will not start!

Hi, orrin.

Mm. Tell me about it. I don't know what they're doing over at the Chromium Project, but these last few releases have just been throwing up one problem after another. Partly, I believe it's to do with the fact that Puppy's essentially a lightweight distro, so doesn't tend to have all the stuff automatically downloaded & installed from the repos, the way it is for the big, mainstream distros...

What's the output from starting it via the terminal?

Code: Select all

/usr/bin/iron

It's a good way to check for problems; usually gives you an indication of why it won't run. NSS, by any chance?


Mike.

[orrin]

What's the output from starting it via the terminal?

Code: Select all

/usr/bin/iron

It's a good way to check for problems; usually gives you an indication of why it won't run. NSS, by any chance?

Yes, Gtk 3 is loaded and appears in /usr/lib64
Maybe the attached screen image will give you a clue as to the problem.

[battleshooter]

That looks like a clash in GTK3 versions in my unexpert opinion

[Mike Walsh]

***New Version: Iron 61.0.3200.0***

Evening, all.

The new release of Iron, 61.0.3200.0, is now available for download from the link in post #1. I've included the current version of PepperFlash, 27.0.0.130, which installs to /usr/lib/mozilla/plugins, as usual.

----------------------------------------------------------------------------------------------

The Chrome 'upgrades' have once again carried over to Iron, since it's based very closely on it.....including the 'white on black' tooltips, which I'm still not too sure about! And the lack of highlighting in the bookmarks listings...

NetFlix is working, thanks once again to following SRWare's advice to 'poach' the libwidevinecdm.so module from the current version of Chrome itself.

-----------------------------------------------------------------------------------------------

The biggest change is that of Iron having moved to Chromium 61, thus the identical problems from Chrome 61 have reared their head. I have, therefore, made use of battleshooter's expert tuition again, and incorporated the same 'fix' he figured out for t'other browser.....due to Chromium requiring the very newest Network Security Services libraries. So far, it seems to be working.....

I've also added the 'silent-launch' feature to /root/Startup, which certain individuals have been raving about..!

Enjoy it while you can...


Mike.

[step]

Iron-61.0.3200.0-amd64-slacko.sfs

This update appears to run in Fatdog64 after installing package gtk3 from the package manager and linking libpcre.so.3 with this command:

Code: Select all

ln -s /usr/lib64/libpcre.so.{1.2.6,3}

Then start iron from a terminal as user spot:

Code: Select all

run-as-spot iron &

A few error messages that aren't fatal:

Code: Select all

/opt/iron/nacl_helper: error while loading shared libraries: libnss3.so: cannot open shared object file: No such file or directory
[1:1:1005/213535.094326:ERROR:nacl_fork_delegate_linux.cc(316)] Bad NaCl helper startup ack (0 bytes)
Fontconfig error: Cannot load default config file

Tab bar and menus use a fixed-size font. I don't know if that's the intended iron style, but it looks good for a change.

Thanks Mike!

[Mike Walsh]

Hi, step.

Mm-hm; it's not at all a bad browser, TBH. Like all the Chromium-based browsers, it's getting more & more temperemental as its parent evolves. Don't get me wrong; I fully understand the desire of the guys over at the Chromium Project to make their browser as safe and secure as it's possible to make it.....but, boy! is it giving us problems here in Puppyland.

Oscar, peebee, myself, and those few others on the current 'roster' who develop these browsers into usable Puppy packages will, of course, try to keep doing so for as long as we can.....but if the last few releases have been anything to go by, I'm not even going to try & make any sort of guarantee. As the final line in my previous post says.....enjoy it while you can.

If it works for you.....that's great. These browsers are very 'noisy' in the terminal, and have always entailed workarounds of one sort or another to run with Pup's root model, but.....we keep managing it. (Somehow..!)

Enjoy.


Mike.

[redandwhitestripes]

Hi Mike,
Iron 61 turns the font into something resembling the old ZX Spectrum font. Now I love the old Speccy but on my widescreen TV it's not very readable. Any way to fix it? If not it's no real problem, I'll just keep using version 60. I'm just happy to have a Chromium based browser on puppy that doesn't crash one me.

[Mike Walsh]

Hi Mike,
Iron 61 turns the font into something resembling the old ZX Spectrum font. Now I love the old Speccy but on my widescreen TV it's not very readable. Any way to fix it? If not it's no real problem, I'll just keep using version 60. I'm just happy to have a Chromium based browser on puppy that doesn't crash one me.

Ah, tell me about it..!

Have you by any chance been adding/removing any fonts to Puppy? I had a spell of trying out different fonts a little while ago, and I've got the same problem in both Precise 571 and Racy 5.5. I run Chrome, Chromium, SlimJet and Iron in 571.....and Chrome, SlimJet and PaleMoon in 5.5.

In Precise, I have the 'pixellated' fonts problem in Chrome/Chromium.....but SlimJet and Iron are fine. (Slimmie's fonts aren't quite what I want, but I refuse to 'mess' with it!)

In Racy (courtesy of battleshooter's glibc-220 'workaround' he produced a few years back), the problem is there in Chrome & Chromium, too. PaleMoon's fine, though.....and QtWeb, which I also use there, is problem-free too.

I have tried till I'm dizzy to sort the problem out, but.....it persists. I really haven't got any advice to offer, I'm afraid.....but if I ever do manage to fix it, I'll let you know what the solution is. OK?


Mike.

[quirkian2new]

Thanks Mike for your tremendous effort of bundle all those gtk3 and libnss libs into iron.sfs . This saves me lots lots of trouble running iron61 on systems without gtk3.

IRON64 browser is always one of my MUST browser.

[Mike Walsh]

Thanks Mike for your tremendous effort of bundle all those gtk3 and libnss libs into iron.sfs . This saves me lots lots of trouble running iron61 on systems without gtk3.

IRON64 browser is always one of my MUST browser.

Hi, quirkian2new.

Glad to hear you're enjoying it; nice to know my efforts haven't been in vain.

That is of course the idea; it's the same concept that battleshooter (to whom the credit belongs) helped me with for Chrome itself, when it started demanding the very newest libnss. By doing it this way, it's all self-contained (rather like watchdog's PaleMoon with the built-in glibc upgrades).....no chance of the newer libs interfering with other apps/programs on your system which are happily working with older versions of the same.

Sounds like it's working as planned..!


Mike.

[Mike Walsh]

***New Version: Iron 62.0.3250.0***

Hallo again, boys'n'girls.

The new release of Iron, 62.0.3250.0, is now available for download from the link in post #1. I've included the current version of PepperFlash, 27.0.0.187, which installs to /usr/lib64/mozilla/plugins, as usual.

----------------------------------------------------------------------------------------------

This release of Iron not only contains battleshooter's libnss/gtk3 modifications, but it's the first 64-bit release to be using the 'run-as-spot' configuration which will, I believe, come to be the standard method for running Chromium-based browsers in Puppy. Many, many thanks, once again, to belham2 for having figured this one out originally for Chrome.

NetFlix is still working, thanks to SRWare's advice to 'poach' the libwidevinecdm.so module from the current version of Chrome itself.

-----------------------------------------------------------------------------------------------

I'm hoping, with this new launcher method - giving you operation as a 'normal' user (which Chromium now demands!), plus full sandboxing - and the inclusion of battleshooter's 'fix' for the libs'n'stuff - that these Chromium-based browsers will be 'safe' for a bit from much more in the way of major modifications. I've probably spoken too soon, but.....we shall see.

Enjoy.


Mike.

[festus]

Iron-62.0.3250.0-amd64-tahr.sfs, is working very well on xenial64-7086, is quick and responsive.

Thank you, Mike, for the great work on these browsers.

bliss,
festus

[Mike Walsh]

Iron-62.0.3250.0-amd64-tahr.sfs, is working very well on xenial64-7086, is quick and responsive.

Thank you, Mike, for the great work on these browsers.

bliss,
festus

You're very welcome. Enjoy.


Mike.

[quirkian2new]

hello Mike

Tried your iron61 on laptop with 2gb ram(Fatdog version 702) and found it freeze often. So i think it would probably due to too little Ram. Maybe 4gb ram is the minimum requirement ? I see somewhere mention the parameter of " --disk-cache-dir=/tmp", that is :

# chrome-wrpper --no-sandbox --disable-infobars --disk-cache-dir=/tmp

is this parameter of "--disk-cache-dir=/tmp" can solve the issue of insuffcient ram of 2gb or 4gb? On systems with 8gb ram, iron61 is running quite smooth.

thanks

[quirkian2new]

hello Mike
Out of curiosity, i compared file size for v61 and v62. Both chrome and chrome-sandbox are over 100MB in v61 whereas chrome-sandbox is only 18kb in v62 ?
Any idea ?

By the way, iron61 seems to kill and quit itself whenever i try to upload a file ?


Edit: v62 run-as-spot seems to kill itself when i tried install ublock origin extension.

[mikeslr]

Hi quirkian2new

hello Mike

Tried your iron61 on laptop with 2gb ram(Fatdog version 702) and found it freeze often. So i think it would probably due to too little Ram. I see somewhere mention the parameter of " --disk-cache-dir=/tmp", that is :

# chrome-wrpper --no-sandbox --disable-infobars --disk-cache-dir=/tmp"

I'm not sure it is possible and even if possible that it would help.

Most Linux Distributions --those for whom Web-browsers are designed-- are created as Multi-User operating systems. Each user has his/her own Home Folder. But that folder can expand to the extent of the available space on the partition on which it is located. Such systems do not maintain a clear distinction between what is in Random Access Memory and what is on the storage media, frequently reading and writing what is in RAM to the User's Folder in Storage. The files websites cache on a computer being so operated would, if memory were exceeded, automatically write to the user's folder on Storage.

A frugally installed Puppy, on the other hand, was designed to run as Root, having access to any storage media attached to the system, but operating only in RAM with nothing being written to Storage unless a Save was executed either manually or, if Menu>System>Puppy Event Manager, Save Session was not set to Zero, periodically. When such Save is performed, everything then in RAM (except the contents of /tmp) is copied/written to Storage. Such contents still remain in RAM until cleared by a Shutdown/Reboot.

Spot was developed in --IMHO a misguided-- effort to emulate the security provided to one user on a Multi-user system from the activity of another user. It creates a sealed folder and limits applications so configured to only have access to it.

Under a Frugal Puppy's "Merge File System" any change occupies RAM, so a Website which has cached its files on your system has done so in RAM. As currently configured, Iron's cache folder is located at /spot/.cache, and the contents of /spot/.cache are in RAM. Reassigning Iron-running-as-spot's cache files to /tmp would not improve the system as changes to /tmp are also in RAM.

I don't think it's possible. The way the problem of cache filling up RAM is handled with a Web-browser NOT running as spot, is to move the cache folder onto storage (e.g. /mnt/home/browser-cache) and symlink it back. The system, including the Browser, follows the symlink and writes directly to Storage rather than occupying RAM. I tried to edit /spot/iron/chrome-wrapper to add such an argument to Line 54 "exec -a "$0" run-as-spot /root/spot/iron/chrome" and execute a Save to preserve that change. Save went into what appeared to be an infinite loop. Neither Menu>Exit>Restart-X nor Ctrl-Alt-Backspace functioned and I finally pulled the plug.

What you might try doing is immediately on bootup (so as not to Save any unwanted junk to your system) open Iron, Click the 3 vertical dots at the Top-Right of its Panel, Select More Tools, Click Extensions, Click Get More Extensions, search for History Eraser and install it. And then perform an immediate Save. Extensions so added will survive a reboot. I'm not sure History Eraser does anything more than merely clicking "Clear Browsing Data" also reached from "More Tools" but it works very well on Iron-Not-Running-as-Spot, and places a nice icon on Iron's Panel which may remind you from time to time to clear cache.

In case you weren't aware, you can import bookmarks, but only after you've first copied them to the /spot folder (and you may have to rename them as 'bookmarks.html' overwriting the current file of that name). However, once imported you won't be able to add or delete bookmarks and have those changes preserved without performing a Save. In my opinion, this renders Iron running as Spot rather inconvenient as one's principal web-browser. And frankly, I don't see any advantage of having one browser run as spot --to protect your system-- if you are almost always using another browser running as Root.

Perhaps I'm mistaken. But I have the impression that the idea of running a browser as Spot came into existence before the mechanism for running Puppies with Automatic Save Removed was developed. With Automatic Save Removed, whatever crap you've picked up on the internet and is currently only in RAM is deleted by a reboot/shut-down without first performing a Save.

The apparent advantage of a spot folder appears to exist without taking into consideration:
(1) Someone mentioned recently, but I didn't make a note of it, that Puppies don't access websites the way other operating systems do; but more importantly
(2) Malware of any type is an application and for an application to run it must be
(a) 'on the Path' and /mnt/home is not "on the Path"
(b) it must be recognized as an executable. Even if malware were downloaded piece-meal, some executable would have to assemble it.
(3) Malware is not written for Linux which in total has less than 10% of computer users, with Windows having approximately 80% of computer users; and Linux is not one distribution but hundreds each of which is to some extent incompatible with every other variation. "Why do I rob banks? Because that's where the money is."
(d) Using technology to protect those things which are valuable to you rather than common sense is on par with the Romans employing the German Tribes to protect the Roman Empire. With Equifax having given away personal identity information of over 143 Million people --the full count may never be known-- and that's just one example that we know about, doesn't the use of data encryption for those few sensitive files you may keep on your computer, and the steps discussed on this post, http://murga-linux.com/puppy/viewtopic. ... 321#965321 and especially babaguy's response, http://murga-linux.com/puppy/viewtopic. ... 487#973487 "Use a PAY-AS-YOU-GO Visa or MasterCard" make more sense.

mikesLr

[Mike Walsh]

@Mikeslr:-

Mike, all this boils down to one thing (and you know it as well as I do.) No matter how I set the browser up, 'you can't please all of the folks all of the time'. It will never be right for everybody.

Many folks are happy with the 'standard' Puppy behaviour. You belong to the school that prefers to control your saves manually. That's fine, 'cos those of us with the 'nous' to perform the necessary modifications for Pup to run in this way usually possess sufficient 'know-how' to re-direct the browser's cache to a different location.

And Linux being as configurable as it is, I'm sure every one of us runs our systems differently. Horses for courses, mate.....

---------------------------------------------------------------------------------

As for 'run-as-spot' being the operating mode for Chromium-based browsers from now on, I didn't voluntarily choose to run Iron this way. Blame Google. From Chromium 62 onwards, the browser code-base simply will no longer permit 'root' operation UNDER ANY CIRCUMSTANCES WHATSOEVER. Henceforth, 'root' operation is a total no-no, as far as Google are concerned. 61 was the final version where the wrapper script would work in anything approaching it's original configuration.

They've decided, in their wisdom, that we ALL need protecting from ourselves....whether we agree with that assessment or not.

Go figure.

--------------------------------------------------------------

The creation, and inclusion of Spot may perhaps (in your opinion) have been 'misguided'. But for our purposes - getting the Chromium-based browsers to keep working in Puppy - it does the job. It works. I tried manually creating a 'named' user directory, along with setting up paths, permissions, etc.....but Pup wouldn't have it. It only recognises three owners; 'root', 'spot'.....and 'fido'.

I'm just making use of existing 'tools'. Furthermore, I'm under no illusions as to my very meagre abilities; I owe a tremendous debt of gratitude to several other people on the Forum for making it possible for me to keep these 64-bit browser packages going. My scripting skills are abysmal.....so I've 'borrowed' from some of the really 'clever' people. Like Oscar and peebee. Battleshooter has helped me a lot with the latest gtk3/nss requirements. And, most recently, belham2 has taken the time, and put the effort in, to figure out the necessary modifications to the wrapper script for 'run-as-spot' operation.....

Even Iguleder and Phil B. have indirectly helped out.

Tahr64 has always been something of a 'test-bed' for me.....and originally it was Google moving Chrome to 64-bit operation which prompted its adoption. It's the only 64-bit Puppy I run; I much prefer the 32-bit Pups as 'daily drivers'.

Our software is so very often a communal effort. These Chromium-based browsers are a very good illustration of that fact.


Mike.

[mikeslr]

Hi Mike,

I hope you didn't read my post as a personal criticism. Despite your modesty, the work you've published has been greatly appreciated. That it may be derivative doesn't detract from that. Someone once said something like "I can see so far because I stand on the shoulders of giants".

As you run your Puppies pretty much the same way as I do, symlinking to external folders when possible, I figured you must have had a good reason for constructing Iron in a way that made that impossible. Seems there was:

Mike Walsh: "As for 'run-as-spot' being the operating mode for Chromium-based browsers from now on, I didn't voluntarily choose to run Iron this way. Blame Google. From Chromium 62 onwards, the browser code-base simply will no longer permit 'root' operation UNDER ANY CIRCUMSTANCES WHATSOEVER. Henceforth, 'root' operation is a total no-no, as far as Google are concerned. 61 was the final version where the wrapper script would work in anything approaching it's original configuration."

Well, while Iron was my preferred, and chrome and its clones continue in some respects to be superior to firefox and its clones, I guess in the future I'll just use a 'Chrome clone' for those operations on which it remains superior. After all, I somehow managed to get by with only firefox and seamonkey for a couple of decades.

As for the three inconveniences I discovered, there are workarounds. with the cache problem, I mentioned installing History Eraser and trying to remember to use it. Download only into /spot/Downloads can be moved to anywhere: pretty much how I handle download anyway. Leaving only the problem of keeping bookmarks up to date --organizing, adding and removing.

As I run various Puppies and various browsers, when I find a website I really think is important to occasionally access --regardless of which Puppy and which browser I'm using-- rather than bookmark it, one thing I already do is copy its URL and a short description into a LibreOffice Writer file. Writer recognizes it as a url and Ctrl-Clicking the text is supposed to open the url in one's default browser. [I've never gotten that to work, but copy and paste does.]

But perhaps there's some application for recording urls which would be better for keeping them organized. I'll have to look around and will post if I find anything useful. Perhaps others have suggestions.

mikesLr