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

[mikewalsh]

@baraka / @ozsouth :-

Mm. Well, I won't dispute the premise of checking for a failure, guys; obviously, further cross-checking can only help. Mind you, in all the years I've used Chrome, I've never actually known it fail.......and I tend to adhere to the K.I.S.S principle, as far as I possibly can..!!

The only boo-boo I have come across is that, just occasionally, the repo isn't updated with the new version quite as promptly as it perhaps should be, but.....that IS a rarity.

Nowhere can you find a checksum for the Chrome download. At least, I've never been able to find one...

Mike.

[ozsouth]

@mikewalsh - the only reason I included a zero size check was that during testing, I hadn't noticed
my connection had dropped earlier, & the script duly deleted my install without a replacement.
Note that a powerout during a download (considering it's usually under a minute, you'd be very unlucky),
would leave an incomplete file & the install would still be deleted, but re-running would likely succeed.

[mikewalsh]

Now then, gang.

Following Oz's (and baraka's!) mention of a download-failure check - which IS a good idea, TBH - here's what will hopefully be the final issue of the portable for a while. You can find it at the usual link in post #1, as always.

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

The updater script has had an error-check added to ensure the downloaded .deb file does in fact exist in /tmp before processing begins. If the download is not there, a GTK-dialog banner comes up warning the download has failed, and advising the user to try again later.....followed by exiting.

If it does exist, everything proceeds as normal.

A wee bit of additional research into wget has shown me how to download the .deb direct to /tmp, as opposed to the portable directory FIRST, followed by a move.... Timings have also been adjusted, so the whole thing runs somewhat quicker now.

As before, if you want to keep the thing completely portable, just run the scripts from the portable directory as normal. If you decide to add a Menu entry, you'll now also get one for the updater, beside it.....obviating the need to have to go into the portable to run it. You'll be able to do it ALL from the Menu.

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

As stated above, this will hopefully be the final, 'self-maintaining' release of Chrome-portable for the forseeable future. (Famous last words..... )

D'l; unzip; place it anywhere you like, outside the save. Click to enter, click 'LAUNCH' to, well, launch it. The 'MenuReadMe' explains what to do with the Menu scripts.

Enjoy.

Mike.

[baraka]

Launcher script variant with auto-update (of course, based on mike's).

Code: Select all

#!/bin/sh
#
# Launcher for 'portable' Google Chrome browser
#
HERE="$(dirname "$(readlink -f "$0")")"
LAST_RELEASE=`wget -qO- https://omahaproxy.appspot.com/linux`
INSTALLED_VERSION=`cat "$HERE/PROFILE/spot/google-chrome/Last Version"`

### UPDATE ###

if [ $LAST_RELEASE != $INSTALLED_VERSION ]; then

#pop-up window

 export MAIN_DIALOG='
 <vbox>
 
     <text>
      <label>Update Now!</label>
    </text>
      <hbox>
          <button>
            <label>YES</label>
            <action>'$HERE/UpdateChrome'</action>
            <action>sleep 1</action>
            <action type="exit">Exit by button</action>
          </button>

          <button cancel>
          </button>
      </hbox> 
      
 </vbox>
 '
 gtkdialog3 --program=MAIN_DIALOG 

fi 

#
mkdir "$HERE/PROFILE" 2> /dev/null
mkdir "$HERE/PROFILE/spot" 2> /dev/null
mkdir "$HERE/PROFILE/spot/google-chrome" 2> /dev/null
#
chown -R spot:spot "$HERE/PROFILE/spot" 
chown -R spot:spot "$HERE/lib"
chown -R spot:spot "$HERE/chrome"
#
if  grep -q 'tahrpup64' /etc/os-release
then
	LD_LIBRARY_PATH=$HERE/:$HERE/lib${LD_LIBRARY_PATH:+:$LD_LIBRARY_PATH}
	run-as-spot "$HERE/chrome/chrome" --user-data-dir=$HERE/PROFILE/spot/google-chrome --disable-infobars "$@"
else
	run-as-spot "$HERE/chrome/chrome" --user-data-dir=$HERE/PROFILE/spot/google-chrome --disable-infobars "$@"
fi

Works pretty well, I think. Feel free to blah, blah.

About the sanity check, it is not enough (for me ) to do an existential check, an incomplete download can potentially...

A very basic script that solves the problem is:

Code: Select all

wget https://dl.google.com/linux/direct/google-chrome-stable_current_amd64.deb --timeout=10 -o log.txt
if ! grep -Fxq "saved" log.txt
then
 rm -f google-chrome-stable_current_amd64.deb log.txt
fi

And (this is the last bit), I think that the update script should start with

Code: Select all

rm -f google-chrome-stable_current_amd64.deb

to be sure everything is clean. A rerun will not necessary repair the install. wget could create the file "google-chrome...64.deb.1" and... is easy to see what will happen after.

Excuse my paranoid self, and a good day to you all.

[mikewalsh]

@baraka :-

Hm! Interesting. I didn't think there WAS a way to check for a current version, but that seems to work a treat. I can't help but wonder how long that site will remain up, 'cos sites come & go like nobody's business these days.....although I have seen 'omahaproxy' mentioned several times in various blogs/fora over the last few years, so I have to assume it's relatively stable.

Since that's a good idea to be able to offer notification of new releases before the thing launches, it rather looks like I will have to do one more re-build. What did I say above; "Famous last words"?

So.....

-----------------<later>-----------------

Ri-i-ight. Okay. Now; I've re-worked your modifications a little, using a small YAD-powered GUI rather than using gtk-dialog (always found the latter to be too much like hard work, despite being builtin to every Pup).

Code: Select all

if [ $LAST_RELEASE != $INSTALLED_VERSION ]; then

yad --undecorated --center --text="                      An UPDATE is available.  Would you like to install it?" \
--button="YES PLEASE - update Chrome NOW":2 \
--button="NO THANKS - ask me later...":1 \
#
foo=$?
#
[[ $foo -eq 1 ]] && exit 0
#
	if [[ $foo -eq 2 ]]; then
	$HERE/UpdateChrome && exit 0
  fi
fi

I found the update wasn't re-writing the contents of the 'Last version' file, so it just got into a never ending loop of constantly checking for and re-offering the update, again & again (which could be due to the way I rewrote the wee GUI). To get round this, I've added a couple of lines at the very end of the update script, viz:-

Code: Select all

CURRENT=`wget -qO- https://omahaproxy.appspot.com/linux`
echo $CURRENT > "$HERE/PROFILE/spot/google-chrome/Last Version"

.....and that rewrites the 'Last version' file to show the version just installed.

I've also added an initial check to ensure any previous download has been removed, although since it now downloads direct to /tmp, this is only really needed within the same session:-

Code: Select all

#!/bin/sh
#
# Updater script for the Puppy 'Chrome-portable' package...
#
HERE="$(dirname "$(readlink -f "$0")")"
#
if [ -d "/tmp/chromenew" ]
then
	rm -rf /tmp/chromenew
fi
#
sleep 1.5
/usr/lib/gtkdialog/box_splash -placement top -bg "#97D94A" -fg black -timeout 6 -text "   Chrome updater is now running.  Downloading files...    "

It all seems to work quite well now.

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

For those of you already running the current Chrome-portable, there's no need to download the whole thing yet again; all you need are the re-worked 'LAUNCH' & 'UpdateChrome' scripts, which I've attached below.

For anybody else - who doesn't yet have it - a re-built package can be found at the usual link in post #1.

Enjoy!

Mike.

[xenial]

I tend not to bother with point release..chromium will last months with no update so i just go for series updates if i can.

[baraka]

@mikewalsh
From the LAUNCH script

Code: Select all

$HERE/UpdateChrome && exit 0

remove "&& exit 0" please, and last version file will be updated. (when chrome is fired up for the first time further down the script)

and remove the entire line

Code: Select all

[[ $foo -eq 1 ]] && exit 0

if you answer "no" chrome should start. Now nothing happens, the shell exits.

I'll admit, looks better.
And this is surely the very ultimate, the concluding and final update.

[mikewalsh]

@baraka :-

Yay! Nice one, mate. See, this is what I keep telling everybody....that I am absolutely crap at coding.

I am not, however, too proud to accept help with things. Which is why, in many ways, these Chrome packages are something of a community effort; many individuals have contributed snippets of code for inclusion, suggested improvements, alterations, etc, over the last 2 or 3 years.

I like this, yes indeedy. I don't know if you've run the updater yet (I have an older, 'test' version of Chrome I use to try out modifications on the updater), but for me, you don't even need to manually re-launch the browser after the update. The instant the final gtk-dialog banner disappears, the new version auto-launches all by itself. And that's very neat.

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

I've re-packed/re-uploaded the tarball with these mods. I've also re-uploaded the scripts to the post above, so those are now the re-worked versions.

The new, 'current' tarball is at the usual link in post #1.

Mike.

[baraka]

@mikewalsh
Glad you like it! It works as it should.
Neah, this is kiddie stuff and both my limit. I'm sure, giving more time and attention, you'll breeze through them.
Big

[mikewalsh]

Afternoon, gang.

Now; a 'heads-up'. For anyone who's running the new version of Chrome-portable with the version-checker update script, and who has had TWO updates today, you can quit head-scratching.

I got the update to v99.0.4844.51 approx mid-morning. Around an hour ago, I switched Puppies, and launched Chrome again.....and got offered a new version for a SECOND time. I OK'd it, and let it run through once more.....to find out that I was back on the previous v98.0.4758.102....??!?

Of course, I straightaway suspected that the version-check URL was telling porkies, so I looked into the main release chart which that URL derives from. Nope, it was behaving itself all right; it appears that out of all the different channels & releases Google produce Chrome on, the Linux stable release is the only one that has been reverted BACK by one build during the course of today.

To double-check, I downloaded the .deb from the normal current releases page, extracted it, and knocked together a quick 'test' build just to check the download URL. It's definitely reverted to the previous build, even though v99 WAS there this morning.

At a guess, they've discovered something that needs fixing, so.....we'll probably get the offer of an update again at some point over the next couple of days. I've previously known the Linux stable release to be a bit late being added to the repo, but this is the first time I've ever known it be reverted like this.

With the Chrome "production line" being run by a cohort of semi-autonomous 'build bots' - no humans involved! - I wouldn't like to even hazard a guess as to what's actually happening..... Doubtless it'll resolve itself, sooner or later.

Just thought I'd let y'all know.

Mike.

[rockedge]

@mikewalsh A question about the Chrome portable. It's looking to update. It says it can not of course. I thought I could update the portable at one time but that might be wishful thinking.

Do I continue using it until the next portable and then having to redo everything using the next version?

Currently typing from Chrome-portable in KLV-Airedale-beta3 where it works outstandingly!!
(as well as the ungoogled-chromium portable, really good stuff both of these offerings)

[mikewalsh]

@rockedge :-

Right, Erik. Let's get you sorted out, shall we?

When you say "it wants to update but it of course can't", I'm going to hazard a guess that you're getting that wee red-shaded box up near the Menu 'hamburger' symbol.....saying something like an update is available, but Chrome cannot update?

See, the 'normal' internal updater for Linux Chrome, where it updates 'in-situ' (like Firefox) has never worked the way it does under Windows & Mac. Google have set the .deb package up to install a daily cron job, which operates in tandem with a URL to a dedicated Google repo that then 'adds' a repository to your package manager. Chrome updates then come down the pipe along with system updates.

This is how it's designed to work for mainstream distros, where everybody is just a 'user', with dedicated home directories. The normal system locations for Chrome installation don't work for us because of running-as-root and the need to run the browser as user 'spot'; it basically chucks a giant spanner in the works!

So; this is why the specific Puppy 'Chrome-portable updater' has been developed.

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

Initially, I built the updater to work in tandem with manual input by the user. The user checks the Chrome blog to see if a new version is available; if it is, they then run the update script manually.

Member "baraka'' was the one who first got me looking into this. A couple of weeks ago, he posted about a modified updater which checks for a new version FOR you. I don't know quite how he found it - the URL he discovered appears to be a part of Google's own internal developer network - but it gives a blank web-page with just a version number in the top left-hand corner. There's one of these web-pages for every 'channel' and platform that Chrome supports; 15-18 in total, covering 'stable', 'beta' AND 'Canary' channels.

Importantly, this allowed the two of us to automate the 'version-check' part of things. (It's also why I re-built the thing a few months ago, so that just one, single directory needs to be changed. Everything else remains 'as-is'.)

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

I would download the current version of the portable from my MEGA a/c (link in the first post). Swap your 'PROFILE' directory over from your existing one. Then, fire it up.

At launch, the wrapper-script now checks this URL for a new version. If no update, it just launches as normal. If an update is discovered, it then pops-up a wee GUI, asking if you want to update to the new version. If you say 'No', it launches as normal. If you say 'Yes' (which most will do), it then launches the portable updater script for you.

Once this has run through, the new version will launch automatically without the need for user input.....all ready to go.

Let me know how you get on with it. Unless this develops any major hiccups, it will remain the 'current' version in my repo, because it should henceforth update itself when YOU want it to.

Mike.

[mikeslr]

Mike, about Updating:

Before updating does one still have to have the devx.sfs loaded? --IIRC so that the 'ar' module be present. If so, there should be something to that effect in the notice of a update being available.

Tip: To avoid completely loosing GC, before updating I Right-Click the portable folder, select Duplicate and add 'backup' to the name of the duplicate; then 'update' the original. If all goes well I'll delete the backup.

p.s. Good thing I applied the above precaution. devx was needed. Running updater without it loaded just ends up deleting the chrome folder within the portable-chrome folder.

[rockedge]

@mikewalsh I downloaded from Mega and the update mechanism activated and did it's job when I started up. Update took a few moments and completed successfully.

running Version 99.0.4844.51 (Official Build) KLV-Airedale-beta3 (64-bit).

[mikewalsh]

@rockedge :-

@mikewalsh I downloaded from Mega and the update mechanism activated and did it's job when I started up. Update took a few moments and completed successfully.

running Version 99.0.4844.51 (Official Build) KLV-Airedale-beta3 (64-bit).

Excellent..! It's behaving itself. It should do so in KLV because 'ar' is installed in KLV by default; it's a part of the standard Void setup.

Nice to know the update completes as it should. I haven't yet tried it in KLV myself.

Mike.

[mikewalsh]

@mikeslr :-

Mike, about Updating:

Before updating does one still have to have the devx.sfs loaded? --IIRC so that the 'ar' module be present. If so, there should be something to that effect in the notice of a update being available.

Tip: To avoid completely loosing GC, before updating I Right-Click the portable folder, select Duplicate and add 'backup' to the name of the duplicate; then 'update' the original. If all goes well I'll delete the backup.

p.s. Good thing I applied the above precaution. devx was needed. Running updater without it loaded just ends up deleting the chrome folder within the portable-chrome folder.

Mike:-

The portable contains the necessary components for updating itself within Fossapup64, Bionicpup64, Xenialpup64 and Tahrpup64.....and derivatives thereof. For these, the updater can be run straightaway. I picked these 4 to 'build-in' components for, since they've always been the most "mainstream" of the legion of available Pups/Puplets.....and tend to get the most downloads. If nothing else, most noobs have at least heard of Ubuntu.....

There's no way in hell I could do this for every single Puppy that's ever been released..! (Just over 6700, according to ally the other day.)

For any of the Slackos, or EasyOS, or Quickpup, or ANY of the non-'buntu-based Pups/distros, you will probably need to load your devX SFS first. There's an easy way to check if you have "ar" installed, of course; just open a terminal, type in

Code: Select all

which ar

.....and hit "Enter". If it's installed, it'll return the file-system location. If not, it'll return something like

Code: Select all

bash: command not found

So that Puppians are aware of this fact, I've added additional info to the wee "update launch GUI". Re-packed it, and re-uploaded it to the usual location in post #1.

Hopefully, this additional 'refinement' will help to address your concerns.

T'other Mike.

[user1234]

(In continuation to this post.)

@mikewalsh- A new idea - What about letting the people choose themselves which icon they'd prefer by creating a chrome-icon directory in DATA directory containing a chrome's icons of different years. Then people can symlink the icons themselves to DATA/chrome.png (firstly the chrome.png can be set as deafult which can be your favourite... The 2008 one).

Would attach few myself today to this post .... if you like the idea.

[mikewalsh]

@user1234 :-

(In continuation to this post.)

@mikewalsh- A new idea - What about letting the people choose themselves which icon they'd prefer by creating a chrome-icon directory in DATA directory containing a chrome's icons of different years. Then people can symlink the icons themselves to DATA/chrome.png (firstly the chrome.png can be set as deafult which can be your favourite... The 2008 one).

Would attach few myself today to this post .... if you like the idea.

Could do, could do.

What IS needed is for me to figure out where the tray icon (for the task list itself, i.e., the open application) actually comes from. Despite that I've been using different icons for years, the task-list always shows the bog-standard Chrome icon.....and that's the bit that I want to be able to change.

Either that, or else figure out how the tasklist entry is constructed, so I can call the icon from a different location. Most of the 'clones' have their icons in a separate sub-directory; with Chrome, it appears to be hard-coded into one of the multiple proprietary database files within the profile.

I suspect it would be simpler to modify the task-list code! Any ideas?

Mike.

[user1234]

Have already created icons for 2008, 2011, 2014, 2022 icons. Got all of them from internet, except that I extracted the 2008 one from chromeupdate.png (that was previously there) since I deleted it and could not get it from internet.

The size of the icons are not perfect in comparision to real chrome icons (I mean the transparency at the borders), though they look similar.

For the updater ones I just copied the arrow thing on top of chromeupdate.png and put the similar one above all of them.

Have attached a tarball containing them. Extract it in path/to/chrome-portable/DATA/.

[Grey]

Any ideas?

As I understand it, we are talking about two icons. In the "header" of the browser at the top and in the tray at the bottom. Although there is still in the browser settings.

I experimentally "drew some details" to the product_logo_48.png icon (Marmaduke). Yes, it doesn't help

Apparently the icons are "sewn" inside the binary or resources.pak. Ideally, we need to compile from scratch with our icons (it is desirable that the icon be associated with puppies and chrome at the same time)

[user1234]

Any ideas?

As I understand it, we are talking about two icons. In the "header" of the browser at the top and in the tray at the bottom. Although there is still in the browser settings.

I experimentally "drew some details" to the product_logo_48.png icon. Yes, it doesn't help

Apparently the icons are "sewn" inside the binary or resources.pak. Ideally, we need to compile from scratch with our icons (it is desirable that the icon be associated with puppies and chrome at the same time)

Well I was talking about .desktop icons.

You are right about icons being 'sewn' inside the binary. For that nobody can help.

EDIT: I have updated your LAUNCH script a bit. Please check-

Code: Select all

#!/bin/sh
#
# Launcher for 'portable' Google Chrome browser
#
TODAY_DATE=$(date +'%m/%d/%Y')
HERE="$(dirname "$(readlink -f "$0")")"
#
if [ ! -f "./LAST_UPDATE_CHECK_DATE" ] || [ $(cat "./LAST_UPDATE_CHECK_DATE") != "$TODAY_DATE" ];then
LAST_RELEASE=`wget -qO- https://omahaproxy.appspot.com/linux`
INSTALLED_VERSION=`cat "$HERE/PROFILE/spot/google-chrome/Last Version"`
#
### UPDATE ###

if [ $LAST_RELEASE != $INSTALLED_VERSION ]; then

yad --undecorated --center --form --text="                      An UPDATE is available.  Would you like to install it? 
---------------------------------------------------------------------------------
            If you are running Fossapup64, Bionicpup64, Xenialpup64 or
           Tahrpup64 - or derivatives of these - you can allow the update
        to proceed. If you are running any other Puppies or Puppy-derived
          distros - for instance, EasyOS or Quickpup, or any of the Slackos,
                       you will need to hit the 'No thanks' button now.

           To check for the prescence of the 'ar' binary, open a terminal
           and type 'which ar', followed by hitting Enter. If you receive
           a filesystem location, you are good to go. If you receive 'bash:
           command not found', then 'ar' is not present on your distro.

          Please load your devX SFS package before attempting to update
            Chrome. If you do not, then the Chrome-portable will end up
             non-functional, due to the 'chrome' directory being deleted!
                                                                                                                         " --button="YES PLEASE - update Chrome now":2 --button="NO THANKS - ask me later...":1
#
foo=$?
#
	if [[ $foo -eq 2 ]]; then
	$HERE/UpdateChrome
  fi
fi
fi
#
mkdir "$HERE/PROFILE" 2> /dev/null
mkdir "$HERE/PROFILE/spot" 2> /dev/null
mkdir "$HERE/PROFILE/spot/google-chrome" 2> /dev/null
#
chown -R spot:spot "$HERE/PROFILE/spot"
chown -R spot:spot "$HERE/lib"
chown -R spot:spot "$HERE/chrome"
#
echo $TODAY_DATE >./LAST_UPDATE_CHECK_DATE
if  grep -q 'tahrpup64' /etc/os-release
then
	LD_LIBRARY_PATH=$HERE/:$HERE/lib${LD_LIBRARY_PATH:+:$LD_LIBRARY_PATH}
	run-as-spot "$HERE/chrome/chrome" --user-data-dir=$HERE/PROFILE/spot/google-chrome --disable-infobars "$@"
else
	run-as-spot "$HERE/chrome/chrome" --user-data-dir=$HERE/PROFILE/spot/google-chrome --disable-infobars "$@"
fi

Previously it would have checked for latest version everytime, but now it checks only once in a day. This saves chrome startup time by a lot (at least for me). Previously chrome would start in nearly 10~15 seconds everytime except for its first run every boot, which took nearly 1-2 minutes. Now it takes nearly 2-3 seconds, i.e. the normal chrome startup time.

[mikewalsh]

@user1234 :-

Mm-hm. Okay; I can see where that would make sense.

This isn't an issue for me, since Chrome starts up in the background, at boot-time, via the

Code: Select all

--silent-launch

...."switch". Thus, every launch of Chrome is instantaneous, because it's already running. Admittedly, this really only makes sense for those of us with large amounts of RAM - I have 32GB - where Chrome usage is negligible in respect of the total available.

I'll do some experiments with this mod, see what happens, like. Thanks for that, anyway.

Mike.

[xenial]

The supposed spying from google has never really bothered me like it does other people..But to give google their due they make a damn good browser.

Thanks for the portable.

[Grey]

to figure out where the tray icon

Unpacked the sources, it turned out more than 8 gigabytes (& it is recommended to compile with at least 8 GB of RAM) This is version 99.0.4844.84. There are a lot of icons for the title "what we need". Perhaps we need to redraw (replace) this /src/chrome/app/theme/chromium/linux/product_logo_32.xpm

We need to make sure first, and then compile the entire browser For example, I want to build a browser not only for the sake of the icon, but also with support for AVX2 instructions.

[Grey]

@mikewalsh We continue the fun https://forum.puppylinux.com/viewtopic.php?t=5566

[user1234]

Unpacked the sources, it turned out more than 8 gigabytes (& it is recommended to compile with at least 8 GB of RAM)

Will not be able to contribute in the fun then. Wish you best luck for the project undertaken .

[user1234]

Found a severe bug in my code which makes chrome portable not portable! Instead of using dir $HERE, I used ./ .

Here is the correct code-

Code: Select all

#!/bin/sh
#
# Launcher for 'portable' Google Chrome browser
#
TODAY_DATE=$(date +'%m/%d/%Y')
HERE="$(dirname "$(readlink -f "$0")")"
#
if [ ! -f "$HERE/LAST_UPDATE_CHECK_DATE" ] || [ $(cat "$HERE/LAST_UPDATE_CHECK_DATE") != "$TODAY_DATE" ];then
LAST_RELEASE=`wget -qO- https://omahaproxy.appspot.com/linux`
INSTALLED_VERSION=`cat "$HERE/PROFILE/spot/google-chrome/Last Version"`
#
### UPDATE ###

if [ $LAST_RELEASE != $INSTALLED_VERSION ]; then

yad --undecorated --center --form --text="                      An UPDATE is available.  Would you like to install it? 
---------------------------------------------------------------------------------
            If you are running Fossapup64, Bionicpup64, Xenialpup64 or
           Tahrpup64 - or derivatives of these - you can allow the update
        to proceed. If you are running any other Puppies or Puppy-derived
          distros - for instance, EasyOS or Quickpup, or any of the Slackos,
                       you will need to hit the 'No thanks' button now.

           To check for the prescence of the 'ar' binary, open a terminal
           and type 'which ar', followed by hitting Enter. If you receive
           a filesystem location, you are good to go. If you receive 'bash:
           command not found', then 'ar' is not present on your distro.

          Please load your devX SFS package before attempting to update
            Chrome. If you do not, then the Chrome-portable will end up
             non-functional, due to the 'chrome' directory being deleted!
                                                                                                                         " --button="YES PLEASE - update Chrome now":2 --button="NO THANKS - ask me later...":1
#
foo=$?
#
	if [[ $foo -eq 2 ]]; then
	$HERE/UpdateChrome
  fi
fi
fi
#
mkdir "$HERE/PROFILE" 2> /dev/null
mkdir "$HERE/PROFILE/spot" 2> /dev/null
mkdir "$HERE/PROFILE/spot/google-chrome" 2> /dev/null
#
chown -R spot:spot "$HERE/PROFILE/spot"
chown -R spot:spot "$HERE/lib"
chown -R spot:spot "$HERE/chrome"
#
echo $TODAY_DATE >$HERE/LAST_UPDATE_CHECK_DATE
if  grep -q 'tahrpup64' /etc/os-release
then
	LD_LIBRARY_PATH=$HERE/:$HERE/lib${LD_LIBRARY_PATH:+:$LD_LIBRARY_PATH}
	run-as-spot "$HERE/chrome/chrome" --user-data-dir=$HERE/PROFILE/spot/google-chrome --disable-infobars "$@"
else
	run-as-spot "$HERE/chrome/chrome" --user-data-dir=$HERE/PROFILE/spot/google-chrome --disable-infobars "$@"
fi

[rockedge]

@mikewalsh I have a bug!

Now every time I run Google-Chrome portable first pops up the update routine which works it seems then nothing.

I looked it over and now every time I attempt to start, the updater runs and finishes then nothing happens.
Then I noticed the /chrome directory is completely gone!

I replaced it then went to start Chrome and the same thing. Updater runs and removes /chrome directory completely ending up totally broken.

I have tried the code fix @user1234 shared and that seems to work.

Odd that the /mnt/home/Google_Chrome-portable/chrome directory is disappearing entirely, is any one else experienced this?

Just started to occur yesterday.

[mikewalsh]

@rockedge :-

Mm. Odd....

I will say this; I had a spell a few days ago when the updater was going potty. Coming up every time I launched the browser. Mine launches at boot with the "--silent-launch" parameter, so it's just idling in the background until I want it, then....bang! it's there immediately when I open it.

Which is essentially what happens with a Chromebook, except of course, the browser is a core part of the OS with them.

As of about 3 days ago it's settled down and is behaving itself now. But at no point did the /chrome directory actually disappear....

Keep me posted, will you?

Mike.

[user1234]

@mikewalsh I have a bug!

Now every time I run Google-Chrome portable first pops up the update routine which works it seems then nothing.

I looked it over and now every time I attempt to start, the updater runs and finishes then nothing happens.
Then I noticed the /chrome directory is completely gone!

I replaced it then went to start Chrome and the same thing. Updater runs and removes /chrome directory completely ending up totally broken.

I have tried the code fix @user1234 shared and that seems to work.

Odd that the /mnt/home/Google_Chrome-portable/chrome directory is disappearing entirely, is any one else experienced this?

Just started to occur yesterday.

Which LAUNCH file are you having problem with? You say that my fix fixes the problem. I think the one already given by mike must not create problem, since my fix just fixed a bug which I found in my code. But I don't think the LAUNCH file without my fix would have been the cause of the problem either.