Windows seems to have a tendency to occasionally update display identities, even when a display configuration has not changed, causing the user to change the specified order when invoking multibrowse. This can be quite annoying.

It might be a better approach to sort the displays by ascending x and y values, so that even if Windows assigns different identities to the displays, provided they are in the same spatial configuration, the invocation order will remain the same.

Sometimes you might want to add additional flags to the invocation of Chrome, but multibrowse has a hardcoded list of flags and currently does not allow this.

Ideally, one would invoke it as follows:

multibrowse.exe --ui-disable-zero-copy http://url1.com http://url2.com

However, the argument order shouldn't matter.

This would also make it possible to override any of the flags already present in multibrowse.

A browser that was opened by multibrowse and was left running for multiple months now shows a "upgrade chrome" popup, which is an eyesore. Would be nice if this didn't show.

It seems like it can be probably be disabled permanently with --simulate-outdated-no-au="01 Jan 2199"

Hello there,

When trying to use your multibrowse program on Ubuntu we get the traceback following :

Multibrowse v2.0.0
Traceback (most recent call last):
File "multibrowse.py", line 34, in
displays = sorted(sorted(platform.displays, key=itemgetter('x')), key=itemgetter('y'))
File "systems/linux.py", line 32, in displays
display = {k: int(v) for k, v in matches.groupdict().items()}
AttributeError: 'NoneType' object has no attribute 'groupdict'
[8330] Failed to execute script multibrowse

Do you have any idea ?

Thank a lot

Ex : base_url/mode?=1&video=1
only partially read, becomes "base_url/mode?=1"

Hi,

Let me start off with what an amazing tool!

I see that the script should support Chromimum, where should I add --ignore-certificate-errors-spki-list
Could it also be possible to support login to a dashboard, with username & password.

Thanks for creating this...much appreciated.

Basically, I have 2 issues:

1. when launching, it never remembers any previous selections we store in cookies.

2. I need all the browser windows we open to be able to share cookies (same site)

Is this possible? I'm assuming there's a way to assign a profile at launch, but I'm not seeing it.

Thanks for your help!

We should prefer the chromium-browser/chromium binary on Linux if available and use google-chrome-stable/google-chrome as a fallback.

Hi, in local machine everything works fine, but when i try to execute the program from a ssh connection i get this error.

Multibrowse v2.1.0
Can't open display
Traceback (most recent call last):
File "/home/pi/Desktop/multibrowse/multibrowse.py", line 34, in
displays = sorted(sorted(platform.displays, key=itemgetter('x')), key=itemgetter('y'))
File "/home/pi/Desktop/multibrowse/systems/linux.py", line 29, in displays
for idx, line in enumerate(check_output(['xrandr']).decode('utf8').split('\n')):
File "/usr/lib/python3.9/subprocess.py", line 424, in check_output
return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
File "/usr/lib/python3.9/subprocess.py", line 528, in run
raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['xrandr']' returned non-zero exit status 1.

There is any way to fix it?

Thanks and amazing program!

Hi,
it's possible to configure this system to open webpages that requires login and pass this credentials somehow?

According to the docs, displays should be ordered starting top left going right, then down. On linux the order seems to be determined by the operating system.

Negative monitor positions seem to cause problems due to ctypes.c_ulong being used instead of ctypes.c_long

Users are still able to exit the browser (or fullscreen) by long-pressing touchscreens on Windows platforms. This should be prevented.

On displays with smaller than 500px resolution, Chrome seems to full-screen at 500px no matter the resolution of the display. This causes overflow into adjacent displays and incorrect sizing of content.

Likely seems to be this Chromium commit which sets a minimum content size of 500px: chromium/chromium@6b0ecc1