Vivaldi 'portable' browser - 32-bit & 64-bit

[mikewalsh]

@gychang :-

This 'double tab' thing is nowt to do with Mochi's scripting, Greg. It's nothing specifically to do with Vivaldi OR your set-up, either; it appears to be a long-standing, endemic 'bug' in the Chromium code-base, and for some crazy reason it affects 32-bit compiles to a greater degree than it does 64-bit ones.

I noticed it quite a while back. I use 32-bit Iron as the default browser in some of my 32-bitzers. If I go to open a Thunderbird link in the 'default' browser, up pops Iron, as it should.....but with two identical tabs showing for the same link.

I haven't yet found a way to stop it from performing its party trick, and I gave up worrying about it some time ago. Sure, it's annoying.....but ya CAN live with it. Especially when it's out of your control.

Mike.

[MochiMoppel]

.I hope I haven't bored ya too much.

Just enough to let me forget about this frivolous idea

@gychang I've seen the problem in your nice video (I like how you explain what to click in Mike's Vivaldi thread).
Everything seems OK...except for your shortcut. Isn't Ctrl+U already used by Vivaldi to show the page source? I wonder why your ROX-Filer shortcut works at all. What happens if you just click on the desktop icon?

[gychang]

.I hope I haven't bored ya too much.

J
@gychang I've seen the problem in your nice video (I like how you explain what to click in Mike's Vivaldi thread).
Everything seems OK...except for your shortcut. Isn't Ctrl+U already used by Vivaldi to show the page source? I wonder why your ROX-Filer shortcut works at all. What happens if you just click on the desktop icon?

good idea, I disabled most of keyboard shortcuts in Vivaldi since I use keyboard shortcuts frequently for other tasks. When I just click on the desktop icon, I do get a message saying xclip or xset is needed. I also tried chaing the Ctrl-U to Ctrl-M and same double tab occurs. Anyway a great utility u came up with.

[MochiMoppel]

. When I just click on the desktop icon, I do get a message saying xclip or xset is needed.

That's impossible - or a miracle. After copying the URL to the clipboard it doesn't matter if you use your Ctrl+U shortcut or click on the icon. Both do the same thing, they invoke the script. If Ctrl+U works for you, so should an icon click. The script wouldn't work without an installed xclip or xsel, but xclip is installed in Xenial. So no matter how I look at it, if you get this message it makes no sense to me.
 
 

This 'double tab' thing is nowt to do with Mochi's scripting, Greg. It's nothing specifically to do with Vivaldi OR your set-up, either; it appears to be a long-standing, endemic 'bug' in the Chromium code-base, <snip> Sure, it's annoying.....but ya CAN live with it. Especially when it's out of your control.

It's in YOUR control, Mike, because it's a bug in your LAUNCH script

I tried Vivaldi32 in Xenial (I'm now posting from it). Lots of errors when starting from command line and also while running Vivaldi, but somehow Vivaldi works. I get the same double tabs. This is not surprising because the last line in LAUNCH reads

Code: Select all

exec -a "$0" "$HERE/vivaldi/vivaldi-bin" "$@" --user-data-dir=$HERE/PROFILE --no-sandbox --disable-infobars "$@"

First I stumbled on -a "$0". This gives the new process the complete path to LAUCH as a process name. Why do you need a process name at all? And why with such a weird name? Maybe no harm done, but just odd.

The double tabs occur because you let Vivaldi open the passed URLs twice. See the "$@" ? This holds the passed URLs and customarily it is the last argument.
In your case it is too, but there is another here: ... vivaldi-bin" "$@" --user-data-dir....
Remove it and you'll get rid of double tabs.

[gychang]

The double tabs occur because you let Vivaldi open the passed URLs twice. See the "$@" ? This holds the passed URLs and customarily it is the last argument.
In your case it is too, but there is another here: ... vivaldi-bin" "$@" --user-data-dir....
Remove it and you'll get rid of double tabs.

It works!, MM, u r amazing!!!

[mikewalsh]

This 'double tab' thing is nowt to do with Mochi's scripting, Greg. It's nothing specifically to do with Vivaldi OR your set-up, either; it appears to be a long-standing, endemic 'bug' in the Chromium code-base, <snip> Sure, it's annoying.....but ya CAN live with it. Especially when it's out of your control.

It's in YOUR control, Mike, because it's a bug in your LAUNCH script

I tried Vivaldi32 in Xenial (I'm now posting from it). Lots of errors when starting from command line and also while running Vivaldi, but somehow Vivaldi works. I get the same double tabs. This is not surprising because the last line in LAUNCH reads

Code: Select all

exec -a "$0" "$HERE/vivaldi/vivaldi-bin" "$@" --user-data-dir=$HERE/PROFILE --no-sandbox --disable-infobars "$@"

First I stumbled on -a "$0". This gives the new process the complete path to LAUCH as a process name. Why do you need a process name at all? And why with such a weird name? Maybe no harm done, but just odd.

The double tabs occur because you let Vivaldi open the passed URLs twice. See the "$@" ? This holds the passed URLs and customarily it is the last argument.
In your case it is too, but there is another here: ... vivaldi-bin" "$@" --user-data-dir....
Remove it and you'll get rid of double tabs.

.....which all goes to prove what I keep telling everybody. I am NOT a 'coder' or a 'script-kiddie'. 90% of what I read in scripts or in the terminal I do NOT understand.

I'm a "bumbler". More often than not, I try things out at random until summat works for me (well, OK; I'm guess I'm probably not quite that clueless, but I'm certainly no expert when it comes down to the nitty-gritty of much of what Bash does, or how you use it correctly.....and it's getting so I no longer have the enthusiasm to learn stuff that I had in the early days).

If something works for me, I'll re-use it. Not very original, I'll grant you, but hey! if it does the job, that's good enough for me.

Anyway; thank you very much for the explanation. It's much appreciated. That's something I've learnt today!

That exec line is merely a modification of the one most of the 'clone' developers use in their own launch wrapper scripts, although I don't know quite how that extra "$@" snuck in. It still worked, so it seemed as if it wasn't that important... They all use 'exec -a "$0"', and I've found that if you don't use it the browsers get all snotty about something not to their liking, and generally refuse to run. So I leave it in.....and as far as 'vivaldi-bin' being an odd process name is concerned, that's the Vivaldi devs' doing, not mine!

Like the 'zilla browsers, the bulk of the 'clones' resides in a huge 'shared' library.....but where the 'zilla browsers fire this up via a small binary, that's where the similarity ends. The Chromium clones don't do it this way; they fire the shared library up directly via a wrapper script instead.....and you've no choice but to use whatever name the devs have given this large shared library.

With regard to the seemingly endless list of errors in the terminal when you run Vivaldi - or ANY Chromium 'clone' for that matter - these browsers ARE very 'noisy' in the terminal. This is normal for these browsers, and is to be expected; Google have them set-up this way to give an ongoing, real-time debug report every second they're running.....mostly for the benefit of developers. (Very quick history lesson, if you don't already know; the Chromium Project is Google's btrowser R&D department. They sponsor it, and it's from where they grab the code for Chrome itself). There is SO much in the way of built-in 'redundancy' routines, that even if you run them in a mainstream distro where they've been specifically re-compiled for that distro by experts, you will STILL get a laundry-list of errors when running them. There is no getting away from this.

Try running Firefox from the terminal. What you will see is not so much different.

Mike.

[gychang]

@gychang :-

Do us a favour, willya? in the Vivaldi directory - same level as the 'LAUNCH' script - rt-clk->Window->Terminal here. Type in

Code: Select all

./LAUNCH

Let us have that terminal readout, please.

Mike.

Sorry for the delay here is the output of the command.

Code: Select all

root@studio /opt/Vivaldi-portable32 # ./LAUNCH
Fontconfig warning: "/etc/fonts/fonts.conf", line 86: unknown element "blank"
[9228:9228:0729/112345.031596:ERROR:browser_dm_token_storage_linux.cc(94)] Error: /etc/machine-id contains 0 characters (32 were expected).
[9228:9324:0729/112345.227852:ERROR:bus.cc(393)] Failed to connect to the bus: Could not parse server address: Unknown address type (examples of valid types are "tcp" and on UNIX "unix")
[9228:9324:0729/112345.227938:ERROR:bus.cc(393)] Failed to connect to the bus: Could not parse server address: Unknown address type (examples of valid types are "tcp" and on UNIX "unix")
[9228:9353:0729/112345.326129:ERROR:object_proxy.cc(621)] Failed to call method: org.freedesktop.DBus.Properties.Get: object_path= /org/freedesktop/UPower: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.UPower was not provided by any .service files
[9228:9353:0729/112345.326438:ERROR:object_proxy.cc(621)] Failed to call method: org.freedesktop.UPower.GetDisplayDevice: object_path= /org/freedesktop/UPower: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.UPower was not provided by any .service files
[9228:9353:0729/112345.326893:ERROR:object_proxy.cc(621)] Failed to call method: org.freedesktop.UPower.EnumerateDevices: object_path= /org/freedesktop/UPower: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.UPower was not provided by any .service files
[9279:9279:0729/112345.420823:ERROR:angle_platform_impl.cc(43)] renderergl_utils.cpp:2152 (ClearErrors): Preexisting GL error 0x00000500 as of ../../chromium/third_party/angle/src/libANGLE/renderer/gl/TextureGL.cpp, setImageHelper:256. 
[9228:9347:0729/112345.483849:ERROR:nss_util.cc(169)] Error initializing NSS with a persistent database (sql:/root/.pki/nssdb): libsoftokn3.so: cannot open shared object file: No such file or directory
[9228:9347:0729/112345.483943:ERROR:nss_util.cc(127)] Error initializing NSS without a persistent database: NSS error code: -5925
[9228:9347:0729/112345.483967:FATAL:nss_util.cc(129)] nss_error=-5925, os_error=0
[9286:9308:0729/112345.490895:ERROR:ssl_client_socket_impl.cc(924)] handshake failed; returned -1, SSL error code 1, net_error -3
Trace/breakpoint trap
root@studio /opt/Vivaldi-portable32 # [9286:9308:0729/112345.493008:ERROR:ssl_client_socket_impl.cc(924)] handshake failed; returned -1, SSL error code 1, net_error -3
[9286:9308:0729/112345.493935:ERROR:ssl_client_socket_impl.cc(924)] handshake failed; returned -1, SSL error code 1, net_error -3
[9286:9308:0729/112345.494083:ERROR:ssl_client_socket_impl.cc(924)] handshake failed; returned -1, SSL error code 1, net_error -3
[9286:9308:0729/112345.494281:ERROR:ssl_client_socket_impl.cc(924)] handshake failed; returned -1, SSL error code 1, net_error -3
[9286:9308:0729/112345.494459:ERROR:ssl_client_socket_impl.cc(924)] handshake failed; returned -1, SSL error code 1, net_error -3

[mikewalsh]

@gychang :-

See here :- https://forum.puppylinux.com/viewtopic. ... 887#p63887

.....and let us know if that's fixed it, please. I'm not chasing this up any further, because I'm starting to get fed-up with it now.

Mike. )

[gychang]

@gychang :-

See here :- https://forum.puppylinux.com/viewtopic. ... 887#p63887

.....and let us know if that's fixed it, please. I'm not chasing this up any further, because I'm starting to get fed-up with it now.

Mike. )

@mikewalsh I really appreciate all your efforts and careful instructions. I followed best I could but after ./LAUNCH, just my left half of my display stays white. It maybe my wrong doing, running so many different 32bit pups in my 64bit laptop, not to mention my incompetence at my advanced age...

[greengeek]

Thanks for this - the 32bit version seems to be working fine ootb on VanillaDpup32. Good to have as many alternative browsers as possible.

[mikewalsh]

@greengeek :-

I've "abandoned" the Vivaldi-portable, to all intents & purposes. I'll sporadically look at it, but it's not one of my favourites - too many pointless 'features' for my liking - and there's a limit to the number of 'portables' I can keep on top of!

The 32-bit version works well, I'll grant you. It's the final, 'legacy' release; nicely stable, and released before most of the silly 'features' made it in. I use it myself in my 32-bit Pups, and am quite happy with it.

Mike.

[mikeslr]

Hi Mike,

Maybe reconsider your abandonment of Vivaldi. With pulse-audio becoming the norm, neither Iron nor Slimjet can produce sound. Vivaldi does.
I agree with you that a lot its 'features' are just gimmicks. But they can be turned off, hidden or just ignored. It might be a little faster than Brave.