Google-Chrome 'portable' 64-bit browser - (with updater!)

[mikewalsh]

The updater script is once again fully-functional, thanks to the ever-resourceful @fredx181 , who has found a more permanent URL for the version check, and modified the 'LAUNCH' and 'UpdateChrome' scripts to suit. The 'ar' stuff n the 'DATA' directory has also been concatenated into a wee AppImage, that appears to work everywhere.

Thanks, Fred!

===============================

The permanent download URL for Chrome-portable will be found at the bottom of THIS post.

---

Afternoon, guys'n'gals.

This is the start of the regular 'portable' release threads.....and it's rather appropriate that it should be the one that really got me interested in the concept all those months ago.

Current release (after the updater has done its thing):- Google-Chrome 122.0.6261.111

Changes, alterations, etc, can be found at the Google Chrome 'blog':-

https://chromereleases.googleblog.com/2 ... sktop.html

---

Following a fair bit of head-scratching and experimenting/suck-it-and-see type stuff, Google_Chrome-portable finally is truly 'portable'.

Earlier versions insisted on putting all the config & cache stuff into the /root/spot directory. Try as I might, I couldn't see a way round this. Chrome is the only one - aside from its 'parent' Chromium, that is - that point blank refuses to run as anything other than spot. (The 'normal user' thing, y'know?) This being the case, every related thing has to have spot permissions.....even sym-links, believe it or not!

So I've carefully worked out what needs to run as 'spot', and what doesn't.....and, making use of Fred's 'readlink' trick from the Firefox portables, Chrome now basically sets its own permissions as it goes along....

And, touch wood, it seems to work.

It will create one single item in 'spot'. The 'Crash Reports' directory, in /spot/.config/google-chrome. However, in 13 years of using Chrome, I've never seen anything in here....and even if it were, it would only be bug-report-type stuff for Google themselves. And I doubt anybody particularly wants to carry that around with them.....

Chrome-portable now contains its own, dedicated, mini-'spot' directory, along with all associated permissions, etc. Instead of a 'PROFILE' directory - Chrome turns its nose up at this, seemingly - it's now the standard /spot/.config/google-chrome (except that it's all self-contained).

It works, though.....so this will be the standard layout for this particular 'portable' browser from now on.

---

Downloads, and MD5 checksums, will from now on always be available from this URL:-

https://mega.nz/folder/HDZElDQK#bbkJFR0QPA3q13JbEnAN6g

Any 'issues', y'all know where to find me. Enjoy!

Mike.

[mikewalsh]

Some news, boys & girls.

(PLEASE NOTE:- This item is no longer available, since it was in fact more trouble than it was worth. Easy enough to implement, but it was found to break many websites in the process. So I removed it.)

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

Following an interesting tidbit of information from Phil B. a bit earlier, I now know you can still run Chrome as /root.

Yep, you guessed it; all courtesy of another of those command-line "--switches". Making use of the "--no-sandbox" switch as well, the following will, it seems, run ANY Chromium-based 'clone' as root, when added to the exec line of the wrapper script:-

Code: Select all

--no-sandbox --test-type

Which is one of the best bits of news I've received for a long while, as far as these browsers are concerned!

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

For anybody who might be interested, a conventional, "run-as-root" Puppy version of Chrome 84 'portable' may be found here:-

https://drive.google.com/drive/folders/ ... sp=sharing

Feedback would be appreciated. If y'all like this version better, I may decide to build them this way in future.

Enjoy!

Mike.

[666philb]

hi mikewalsh,

i knew about the no sandbox thing but you got that horrid warning message banner. --test-type gets rid of that warning

[mikewalsh]

hi mikewalsh,

i knew about the no sandbox thing but you got that horrid warning message banner. --test-type gets rid of that warning

@ Phil:-

The --no-sandbox "switch" has been functional all the way through (although it moans like a good'un).....but unlike most of the other 'clones', Google have done summat to Chrome so if you attempt to run as root in any way, shape or form under normal circumstances, it'll turn its nose up. This --test-type, though, appears to put us back before that 2016 'watershed'. And that's welcome for me.


Mike.

[wiak]

Though I'm current using Chromium (and actually really like it), I don't know much at all about Chrome flags (though I do use --no-sandbox and get the nuisance infobar).

Yes, thanks, --test-type gets rid of that annoyance. My only concern would be what effect that might have in general Chrome operation. There do seem to be some effects (though I have no idea if they effect browsing in my usual practice). Time will tell whether it is wise to use --test-type on a permanent basis:

https://github.com/GoogleChrome/chrome- ... r-tools.md

--test-type
Basically the 2014 version of --enable-automation. codesearch

It avoids creating application stubs in ~/Applications on mac.
It makes exit codes slightly more correct
windows navigation jumplists arent updated https://bugs.chromium.org/p/chromium/is ... ?id=389375
doesn't start some chrome StartPageService
disables initializing chromecast service
"Component extensions with background pages are not enabled during tests because they generate a lot of background behavior that can interfere."
when quitting the browser, it disables additional checks that may stop that quitting process. (like unsaved form modifications or unhandled profile notifications..)

wiak

[wiak]

Please see me above post.

Unfortunately, whilst --no-sandbox no longer gives infobar warning message if you also use option --test-type, Gmail login (for example) will no longer work because of resulting "security" issues (i.e. Google doesn't allow that). I expect there are other side-effects so not really usable.

wiak

[TerryH]

Thanks wiak for this information, I just logged into WDL_Arch64 to try the --test-type switch, as it's the only installation I currently have Chrome. Looks like I'll stick just with the --no-sandbox. The infobar isn't a major inconvenience.

[mikewalsh]

Will:-

Agree with Terry. Thanks for the info.

I confess, I hadn't really thought things that far through, i.e., with regard to GMail, etc. For my own use-case, it makes no difference; although I've had a GMail a/c for many, many years, I never access it via the web-interface.....always via Thunderbird, so in that respect it's not affected.

However, having said that, it must be an issue specific to Chrome/Chromium's code-base. All the other 'clones'....Iron, Opera, Vivaldi, Yandex, etc., are happy to run as the "root" user.....and GMail log-in isn't an issue with them. And in all honesty, I'm no longer wedded to Chrome the way I once was; there's such a plethora of excellent alternatives out there now, that for me, at least, a lot of the 'shiny' stuff on Chrome's 'crown' is beginning to look a bit tarnished...

Even Firefox is a viable alternative now, it's such a vast improvement over the way it was ten years ago.

I have noticed that although Chrome will run as root, it sure as hell doesn't like downloading that way. Upon attempting to do so, it crashes in a huff....

I'll just carry on with the run-as-spot version, that being the case; no skin off my nose! Cheers for the feedback.


Mike.

[wiak]

Actually, I've been experimenting with an almost dumb idea of my own, which is to use xdotool mousemove to auto-place the mouse cursor over the close X symbol of the infobar warning message (having previously used xdotool getmouselocation to find the coordinates), and then to use 'xdotool click 1' to auto-click it off...

wiak

[mikewalsh]

Actually, I've been experimenting with an almost dumb idea of my own, which is to use xdotool mousemove to auto-place the mouse cursor over the close X symbol of the infobar warning message (having previously used xdotool getmouselocation to find the coordinates), and then to use 'xdotool click 1' to auto-click it off...

wiak

↑ ↑ ↑

Ooh, nice to know I'm not the only one who messes around with daft ideas like that..! This is one advantage we DO have over Windoze users. We know all these wee utilities, we have access to them, it's a minor matter to knock a script together to implement them, and, of course (most importantly).......it's FUN!


Mike.

[mikewalsh]

'Portable' Google Chrome 85.0.4183.83

Evening, guys & gals.

Google-Chrome 85 has just been released. In 'portable' format as always, it's now truly portable because the launch script & wrapper set spot ownership permissions as they go. And, with having its own, self-contained 'spot' directory, along with using the expected 'google-chrome' name for its profile, the only thing Chrome places in /home/spot is.....its crash reports. (Which are always empty, BTW).

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

Nowt startling to report with this release. 20 security 'fixes'; details of which can be found at the usual blog URL:-

https://chromereleases.googleblog.com/2 ... op_25.html

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

Thanks, as always, to everybody who's helped - over the years - to mould these packages into what they are today. Y'all know who you are; I couldn't have done it on my own.

Any issues, you know where to find me. Enjoy.


Mike.

[mikewalsh]

'Portable' Google Chrome 86.0.4240.75

Evening, kiddiwinks.

Google-Chrome 86 has just been released. In 'portable' format as always, it's now truly portable because the launch script & wrapper set spot ownership permissions as they go. And, with having its own, self-contained 'spot' directory, along with using the expected 'google-chrome' name for its profile, the only thing Chrome places in /home/spot is.....its crash reports. (Which are always empty, BTW).

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

The usual crop of small but gradual improvements, most of which are required, of course, by the Windoze version.....which accounts for something like 90%+ of their PC downloads. 35 security 'fixes'; details of which can be found at the usual blog URL:-

https://chromereleases.googleblog.com/2 ... op_25.html

Further security improvements, and more new APIs for devs. Please be aware that apparently Google wants to sooner or later force 2-factor authentication onto everyone that uses the browser; when they do, I don't quite know what will happen to the Linux release. The preferred 2FA app runs under Android, and makes the unspoken assumption that everybody has a smart phone. You can't have a web version of a 2FA app, because that defeats the entire purpose of using one.....

Download link is in post#1, as always.

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

Thanks go, as usual, to everybody who's helped - over the years - to mould these packages into what they are today. Y'all know who you are; I couldn't have done it on my own.

Any probs, give us a shout on this thread. Have fun.


Mike.

[Lekium]

This is working perfectly in BionicPup64, thank you so much

[mikewalsh]

This is working perfectly in BionicPup64, thank you so much

No worries; you're welcome. Enjoy!


Mike.

[mikewalsh]

Google_Chrome-portable v87.0.4280.66

Evening, gang.

Chrome 87 has just been released, so.....here's the 'portable' version.

One major change for this release is that startup/loading times are now around 25% faster, and page loading times around 10%. Google have tweaked the JavaScript engine so that the JavaScript Timer throttle now only allows the engine to wake-up and perform any function once a minute, instead of the 30 seconds or so it used to be set for.

This is also the first release of Chrome to have been optimised for new macOS devices running the brand-new ARM-based M1 chip we've been hearing so much about in the tech blogs recently. Apparently, Google have already had to re-release a version in the 24 hrs since this started to roll-out, due to a crashing issue on new Mac devices.

'Nuff said...!!

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

33 security issues have been addressed, along with sundry other stuff. Details at the regular Chrome blog, here:-

https://chromereleases.googleblog.com/2 ... op_17.html

You will notice that there are now two 'launchers'. The extra one is labelled 'LAUNCH_NL', and is, ATM, specifically for use with Fossapup64.....due to the fact of Fossa's 'base' dependencies wanting newer Chrome-required versions of the included 'lib' directory contents. All older Puppies, pre-Fossa, still want them, however, so doing things this way still permits using one, single portable across multiple Puppies.

(You may also need the new 'launcher' with josejp2424's DPupBuster64, as well. I seem to recall this was in fact the first Puppy I'd tried that this issue cropped up in.)

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

Download link in post #1, as always.

Thanks go, as usual, to everyone that's helped to hone these packages into what they are today. You all know who you are; I couldn't have done it without you guys!

Enjoy.

Mike.

[TiredPup]

Thank you for your work on this.

This is not working in Fossapup64. It complains that libffi.so.6 is not found. This is because libffi.so.7 is the version included in Fossapup. Creation of a symlink resolves that complaint. However it gives way to an issue with libpango.

"chrome: symbol lookup error: /usr/lib/x86_64-linux-gnu/libpangoft2-1.0.so.0: undefined symbol: pango_font_description_get_variations
"

I am not sure what to do about that one.

[mikewalsh]

@TiredPup :-

Thank you for your work on this.

This is not working in Fossapup64. It complains that libffi.so.6 is not found. This is because libffi.so.7 is the version included in Fossapup. Creation of a symlink resolves that complaint. However it gives way to an issue with libpango.

"chrome: symbol lookup error: /usr/lib/x86_64-linux-gnu/libpangoft2-1.0.so.0: undefined symbol: pango_font_description_get_variations
"

I am not sure what to do about that one.

TiredPup;

O-kay. First things first... Which release of the Chrome-portable are you using? The newest one, I would hope...

Assuming you're using the current download, for Fossapup you launch with the 'LAUNCH_NL' script, as opposed to the usual 'LAUNCH' script. What this does is to bypass the self-contained 'lib' directory, and allows Chrome to run using Fossapup's newer versions of the dependencies in that directory. The 'lib' directory, however, remains as it is, because it's still needed for anyone running an older Puppy from Bionicpup64 backwards.

Libpangoft2 is the initial "spanner in the works", running with the original 'LAUNCH' script. From that point, it just snowballs..... Hence, the two slightly different scripts.....so the portable can still be used across multiple Pups, if so desired.

Hope that helps.

Mike.

[TiredPup]

@mikewalsh

I assume that you had already provided these instructions in a previous post. Clearly I did not pick up on them or I would not have had an issue. My apologies. Thank you for your patience. It worked fine when I followed instructions.

[mikewalsh]

@TiredPup :-

No worries, chap. We all get days like that.....I'm starting to get my own fair share of them, so I know where you're coming from.

Enjoy!

Mike.

[mikewalsh]

Evening, all.

I was starting to think it would never get here. At long last, here is the portable version of Chrome 87.0.4280.141.

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

16 security 'fixes' have been applied to this version. Details can be found here, for those of you concerned about such things:-

https://chromereleases.googleblog.com/2 ... sktop.html

And this will be the very first release of Chrome to not use PepperFlash any longer.....having reached EOL since the last version of Chrome hit the shelves.

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

Please be aware there are now TWO 'launchers'. 'LAUNCH_NL' is the one to use for Fossapup64/other very recent 64-bit Puppies; this sidesteps the built-in 'libs' directory and allows these very recent Puppies to use their own, even newer dependencies.

The original 'LAUNCH' script is still required for anything from (and including) Bionicpup64 backwards.

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

The download is available from the usual link in post #1. Any issues, drop me a line. Y'all know where to find me.

Enjoy!

Mike.

[mikewalsh]

Now then, guys'n'gals.

Chrome 88 is upon us. So; without more ado, here's the 'portable' version of Google_Chrome 88.0.4324.96.

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

A total of 36 security 'fixes' have been sorted out for this version. Details can be found here, at the usual location:-

https://chromereleases.googleblog.com/2 ... op_19.html

Remember; PepperFlash is now ancient history..!

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

Please be aware there are now TWO 'launchers'. 'LAUNCH_NL' is the one to use for Fossapup64/other very recent 64-bit Puppies; this disables the built-in 'libs' directory and allows these very recent Puppies to use their own, even newer dependencies.

The original 'LAUNCH' script is still required for anything from (and including) Bionicpup64 backwards.

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

The download is available from the usual link in post #1. Any probs, give me a shout; I don't anticipate any, in all honesty, but I'm still capable of being surprised...

Enjoy.

Mike.

[Dingo]

Now then, guys'n'gals.

Chrome 88 is upon us. So; without more ado, here's the 'portable' version of Google_Chrome 88.0.4324.96.
Mike.

It worked for me, but I found that chrome wrote a lot of data ,more than 100 MB, in hidden subdirectories of /root. I really dislike this. Is there no way to make chrome write only in its portable directory?

[s243a]

Now then, guys'n'gals.

Chrome 88 is upon us. So; without more ado, here's the 'portable' version of Google_Chrome 88.0.4324.96.

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

A total of 36 security 'fixes' have been sorted out for this version. Details can be found here, at the usual location:-

https://chromereleases.googleblog.com/2 ... op_19.html

Remember; PepperFlash is now ancient history..!

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

Please be aware there are now TWO 'launchers'. 'LAUNCH_NL' is the one to use for Fossapup64/other very recent 64-bit Puppies; this disables the built-in 'libs' directory and allows these very recent Puppies to use their own, even newer dependencies.

The original 'LAUNCH' script is still required for anything from (and including) Bionicpup64 backwards.

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

The download is available from the usual link in post #1. Any probs, give me a shout; I don't anticipate any, in all honesty, but I'm still capable of being surprised...

Enjoy.

Mike.

It worked for me, but I found that chrome wrote a lot of data ,more than 100 MB, in hidden subdirectories of /root. I really dislike this. Is there no way to make chrome write only in its portable directory?

Maybe create a new user called chrome and have the home directory where the portable is located.

[Dingo]

@mikewalsh
uh, oh, I discovered why starting the launcher for portable chrome it written a lot of megabytes in my /root/spot directory
you left a broken symlink in your portable chrome archive /spot/.config/
pointing to /mnt/sda3/SYSTEM/BROWSERS/PROFILES/Chrome
removing the symlink and the upper directory, now your chrome portable don't write anything in /root
If you meanwhile already deleted this, please ignore my report
thanks for chrome portable!

[Dingo]

do you notice blurred images in browsing webpages using chrome on your puppy? this is the same image I posted in the last message as appears in chrome:

I taken image at 100% of zoom in chrome without resixing windows. I wonder if it is fault of my monitor with TN panel (1366x768) of Thinkpad L440, because the same image in Palemoon looks ok, not blurred

Maybe it depends from chrome antialiasing?

[mikewalsh]

@Dingo :-

Hi, Dingo.

I owe you another apology, it seems... Same excuse as for the Vivaldi-portable, I'm afraid; self-inflicted, and too many late nights..!

Cheers for the "catch" again. I've re-packed & re-uploaded. I'm really going to have to plan my 'work-sessions' out a bit better; after a certain time in the evening, say "That's enough! Leave it till the morning..."

Trouble is, I never take a blind bit of notice of myself..! Why put off till tomorrow what you can do today (or 'tonight'?)

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

With regard to any 'blurring', mmm.....that's not an issue I've experienced myself. The 22" monitor that came with this new Pavilion desktop is an LED one; first one I've used, and also the first time I've been running an Nvidia card, AND the official driver for it. To say it's 'pin sharp' would be an apt description, because I've never had such a clear image before.

It could be the LCD panel in your Thinkpad, that's true, but I've found this can vary from Puppy to Puppy, even on the same machine. On the old Dell lappie, for some reason, Lucid always seemed sharper, and clearer, than any other Pup; whether this was because Lucid used the fbdev 'framebuffer' thing, I don't know.....but it always seemed to work better with the Dell's awkward Intel GPU than any other option did.

Mike.

[Dingo]

@mikewalsh
Very strange since text is pretty clear and sharp, it is a thing I should investigate. I already planned to buy an IPS panel with resolution of 1920x1080. I should make a test with vivaldi to check if also vivaldi has this strange issue. Is your pavillion desktop pricey? What model? I used very often raster and vector programs in my puppy 3.01 rock solid for 11 years, so I'm very interested in high quality monitor graphics

[mikewalsh]

@Dingo :-

Mm. Right; well-l-l....

This is the one I've got (I've switched it to Euros to give you a better idea, price-wise):-

https://www.amazon.co.uk/HP-Pavilion-59 ... ers&sr=1-3

Comes with the Pentium "Gold" G5400 - dual-core, but has H/T, so is effectively a quad-core - running @ 3.7 GHz. It's upgradeable to some of the 'Coffee Lake'-gen i3/i5 'Core' CPUs, if you wanted to.....though for many folks, this G5400 is not a bad chip as CPUs go. Copes well with most modern software, and multitasks with aplomb. 4 GB DDR4 RAM is standard, though I immediately bought another 4 GB stick and turned that into 8 GB. That was in January last year.

Around July, I splashed-out and bought a 16 GB kit for it, from Crucial's web-site. Then in November, I found a really good half-price deal on Amazon - again, Crucial RAM - for a 32 GB kit (maximum the board supports), which actually cost me slightly less than the 16 GB kit. Like many of us during this last crazy 12 months, with being stuck at home & not going anywhere, I had a small but steady build-up in the bank that under normal circumstances wouldn't have been there, so I decided to go for it, and max it out (something I'd always wanted to do).

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

The Pentium comes with the Intel UHD 610 built-in GPU. For many people, this will work fine, and gives a good clear picture. The only reason I went for a separate GPU - an Nvidia GT710, in my case - is because 'onboard' graphics, whether a separate chip OR built-in to the CPU, always 'reserve' system RAM for use as VRAM instead. Specs are pretty much identical to the Intel GPU, but I wanted to make sure that whatever RAM I had remained for mine & Puppy's use; a 'discrete' GPU will always use its own pool of on-board VRAM instead (2 GB in this case, which is plenty for my needs).

If you use the onboard Intel GPU, you don't need to do anything; the in-kernel Intel driver works well. Even with an Nvidia card, you don't HAVE to install the official driver, 'cos 'nouveau' works well these days, too......it's come on a LONG way this last few years, it seems. I just installed the official driver 'cos it was so easy to do with Shinobar's 'Get-Nvidia' utility, that's all; I'd never had any reason to before!

This runs with a 1920x1080 display; you'll need an HDMI lead for this, as there IS no VGA socket. The Pavilion comes with keyboard and mouse, but the monitor is usually extra. My HP 22W came with this one, 'cos it was a 'deal' bundle from Argos here in the UK.....ready-to-go, OOTB. The keyboard & mouse have never been unwrapped, 'cos I prefer my own Logitech items, but the keyboard looks to be quite a reasonable item, with additional 'multimedia' control keys; these work in some Puppies, not in others, though the volume up/down/mute keys will work with most. The mouse is nothing special; it's just 'a mouse'!

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

I wouldn't be surprised if you could find this a bit cheaper by shopping around. Its only 'downside', I guess, is that it doesn't use a 'standard'-sized PSU; it's a custom, small-form-factor one made specially for HP, and it's only 180W. That's OK, 'cos one reason I chose the GT710 is it only draws 19W total; the PCI-e slot supplies what it needs. To date, I've had no problems with it at all. The other 'downside' for some folks would be the fact that being a 'mini-tower', it also uses a custom-sized, smaller-than-usual motherboard, so a replacement might be awkward if one was needed.

Those two 'niggles' aside, however, I'd thoroughly recommend this as a Puppy box. It came with Windoze 10 (of course), but the very first thing I did was to fire-up a Slacko 570 LiveCD and remove Win 10 with extreme prejudice; this was going to be a Puppy-only box from the word go, and I didn't want Windoze digging its smelly claws into my nice new hardware..!!

Mike.

[Dingo]

@mikewalsh

Many thanks Mike, meanwhile I booted Winblows (this notebook has winblows10 installed that I use only if forced to finish some works) and with Google chrome on Windows images in browser are as crisp as in Palemoon in Xenialpup

It seems an issue with chrome (and Vivaldi, since in Vivaldi i noted the same blurring) in xenialpup. Maybe it is worth to start a thread on xenialpup forum to ask to xenialpup users if they noticed the same issue

[mikewalsh]

@Dingo :-

I just prefer desktops over laptops because of the fact you can upgrade bits'n'pieces as & when you want to; there aren't the space limitations you get with a laptop, and with few exceptions - storage, perhaps - you can't usually upgrade very much on a lappie anyway. More so these days, with more & more manufacturers soldering everything to the board.....even CPU & RAM.

Mike.