I have a chrome extension that communicates with a native messaging host to get some data.
The issue is, when I launch the Chrome browser via the shortcut or via the pinned shortcut in the taskbar, the extension is not able to connect to the host. I always get the error:
Failed to start native messaging host.
However, if I launch the chrome.exe via command prompt, everything works fine.
Things I tried with no success:
The taskbar shortcut has no extra flags. The target field has the
following value: "C:\Program Files(x86)\Google\Chrome\Application\chrome.exe"
I tried with the registry under HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Google\Chrome\NativeMessagingHosts\com.company.extension
I tried with the registry under HKEY_LOCAL_MACHINE\SOFTWARE\Google\Chrome\NativeMessagingHosts\com.company.extension
I tried with the registry entry under
HKEY_CURRENT_USER\Software\Google\Chrome\NativeMessagingHosts\com.company.extension
Tried launching the chrome.exe as administrator from the file explorer.
Update:
I added the flag --enable-logging --v=1 to the shortcut to enable logging and when I launch it I get the following output in the console:
[11036:4160:0302/113902.866:ERROR:native_process_launcher_win.cc(140)] COMSPEC is not set
[11036:11856:0302/113902.882:ERROR:native_process_launcher_win.cc(140)] COMSPEC is not set
Update
Upon investigating the chrome.exe process via Process Monitor, I found that there is no COMSPEC environment variable available to it when it is spawned via explorer.
Is there anything else that I can try or something that I am missing here?
As mentioned in the log, Chrome stumbles to start an external process, since COMSPEC, an environment variable pointing normally to cmd, is unset:
[11036:4160:0302/113902.866:ERROR:native_process_launcher_win.cc(140)] COMSPEC is not set
The behavior is different for launching Chrome from cmd itself, since it sets the variable for itself (ans spawned processes).
This can be confirmed by inspecting the Chrome process with Process Explorer.
One can run rundll32 sysdm.cpl,EditEnvironmentVariables as admin (e.g. from admin command line) to open the environment variable settings.
Alternatively, the dialog can be navigated to from Control Panel > System and Security > System > Advanced system settings > Advanced > Environment Variables...
ComSpec is usually set in System variables to
C:\WINDOWS\system32\cmd.exe
Adjust as necessary for your system install. For this setting to apply, you need to log out and log back in, or better yet restart the system.
Related
My operating system is Windows 10 64 bits. I use the latest versions of Firefox and Chrome.
I want to save the pre-master keys in order to use them with WireShark.
For this, I have found many tutorials that all recommend using the SSLKEYLOGFILE environment variable.
However, no matter the path I put in this variable, no file is being created by Firefox (normal and developer editions) or Chrome. This is true even when I restart those browsers or the operating system. This is also true when I clear the cache on the browsers.
I found an alternative way of setting the SSLKEYLOGFILE variable with Chrome by launching the browser with the following argument:
--ssl-key-log-file=PATH
And while this argument does create a log file, it is never populated. It remains completely empty.
What makes this problem even more annoying is that I don't know if there's any debug log that would at least let me know what's happening.
Anyone have ideas on how to solve this issue?
Verified in both Chrome and Firefox
Windows 10 64bit [Version 10.0.17763.379]
I would not use the --ssl-key-log-file flag with Chrome. In my testing, it does not have an effect.
Steps to get SSL keylog file
Change your directory to one that you or your programs have access to. I am using the Desktop folder.
C:\> cd $HOME\Desktop
Set the SSLKEYLOGFILE variable. This sets it for the user (HKCU). To set it for the machine (HKLM), add the /m flag to the end of the command.
PS C:\Users\rj\Desktop> SetX SSLKEYLOGFILE "$(get-location)\ssl.log"
Verify that the variable has been set in a separate powershell window (SetX does not apply to the current window).
PS C:\Users\rj\Desktop> Get-ChildItem ENV: | findstr SSLKEYLOGFILE
SSLKEYLOGFILE C:\Users\rj\Desktop\ssl.log
You can also verify that SSLKEYLOGFILE is a user variable by going to the Control Panel > System and Security > System > Advanced System Settings > Advanced tab > Environment Variables > User Variables. You should see a listing like the powershell example where the value is a directory.
Open Chrome/Firefox and go to an https website like https://stackoverflow.com.
You should see an ssl.log show up on your desktop.
Image of end result
Further Reading
2018-12-07, F5, Decrypting SSL traffic: Great article on doing the above, but cross-platform
2019-03-18, Mozilla, NSS Key Log Format
: The SSLKEYLOGFILE format
2018-08-17, Wireshark Bugzilla, Add TLS 1.3 support: Has good test pcap/SSLKEYLOGFILE contents for TLS1.3 (See Comment 83)
You mentioned that you are using Wireshark. If you using it to export a file from a TLS-encrypted stream in a capture, this article may help you.
One thing I ran into is: Chrome doesn't always fully close when you close the window. Sometimes it stays open in the background. Use the Windows Process Explorer to list all the processes and make cure Chrome is closed. Also, after restarting Chrome, you can double click on it in Process Explorer and select the Environment tab to list all the variables set in it's environment. Make sure SSLKEYLOGFILE is shown in that list. Then Chrome will be sending keys to the specified file.
First off all the problem is that the browsers are protected better and that flaw is patched, thats why you cant capture no ssl. I dont know the details. But chrome with --no-sandbox(linux) option on ,when running will enable you to capture ssl/tls key without any issues from the first try, so you can play with decription of packets. Never tried on windows btw..
This worked for me...
set the SSKEYLOG env: variable as normal in Linux.
just point it to your pefererd save location of the sslkey.log
windows system/advance/env vars/user variables.
add SSLKEYLOGfile with the path to your preferred save location C:/user/sslkey.log
close and your done here.
NSS_ALLOW_SSLKEYLOG=1 is what's missing from the Booleans in Firefox's advanced options.
its been also removed from the general release but is still in the dev ops version. just not listed.
download and install the dev ops version of Firefox
open a terminal start it from its exe if on Linux.
windows just open it from the icon.
go to the about:config
create a new Boolean with NSS_ALLOW_SSLKEYLOG=1 and set it to true.
quit
now create a new symbolic/icon link to the new dev version fire fox version
when you run Firefox from the terminal in Linux the sskeylog should be created and start populating.
run it from the task bar icon in Linux and it doesn't make or update the keylog (just in case you dont want it populating every time you open the browser).
in windows, you will have to disable the env var or it will keep populating every time you open any browser that supports sslkey logging.
on that note opera and vivaldi should work with no changes to the advanced option.
lastly.. open wireshark go to
edit/preferences/protocols/tls
and put the path to the sslkey.log in the pre master secret log box.
click ok.
close everything..
start Wireshark
start your browser (from terminal in linux).
you should start seeing sslkeylog start populating and see extra options for decryption in the tabs of Wireshark.
have fun :)
1st solution:
Restart Chrome
2nd solution:
Use Firefox
I know how to enable chrome's crash reporting interactively:
Menu > Settings > Show advanced settings > Tick "Automatically send usage statistics and crash reports to Google"
However, I'm using chrome in an automated testing setup where we set a new --user-data-dir for each run (so there are no persisted settings, we just get the defaults). Chrome is crashing occasionally (running our javascript/html application) and I would like to be able to report these crashes.
I thought I could add --enable-crash-reporter (according to the List of Chromium Command Line Switches), but that doesn't seem to do anything. (We already pass --enable-logging and a couple other flags on the command line.)
I have a test that triggered an "Aw, Snap":
But if I download the user-data-dir for that run and and point chrome at it, and then visit chrome://crashes, it says crash dumps are not enabled.
(The chrome_debug.log does not say anything about the Aw-Snap either.)
We're running chrome under Xvfb on Debian. We're driving chrome via the direct webdriver (from Protractor).
According to the "crashpad with content shell" docs, the steps for using crashpad with content shell (a stripped down version of Chrome):
Create some directory for your crash dumps to go to (the docs suggest tmp/crashes on Linux and Mac, %TEMP%\crashes on Windows, and /data/local/tmp/crashes on Android)
Start Chrome with --enable-crash-reporter --crash-dumps-dir=<your_dump_dir> (looks like you got half of this)
So I'd try adding the --crash-dumps-dir flag.
In the webgui for defaultapp, you can click a button to start the selected app in the drop down.
What command is hidden under that button to start the universal app? I would like to do this from a remote console programatically.
Also, is the source code available for the defaultapp/webgui?
You can use PowerShell to set an app as startup, run in headless mode etc.
Find the details of connecting it with PowerShell here: https://ms-iot.github.io/content/en-US/win10/samples/PowerShell.htm
Also running in headless mode saves more memory, power, faster boot.
Also command line utils for enabling startup process:
https://ms-iot.github.io/content/en-US/win10/tools/CommandLineUtils.htm
-> Default App Sample: https://github.com/ms-iot/samples/tree/develop/IoTCoreDefaultApp
Let me know if you need anything else.
I am using behat+mink. I wrote some features and am now running tests.
How can I enable xdebug to to stop on breakpoints in phpstorm when running behat tests ?
I have not tried this with Mink yet, but this is configuration that allows me to step through debugging of behat (with behat running on a remote server):
Configure your server with x-debug
Of note, since this is commandline, you need to edit the cli config under /etc/php5/cli/conf.d/20-xdebug.ini.
Set remote_host to the ip of the computer you're using PHPSTORM from
Set autostart = 1
Disable connect_back, you will initiate debugging from the server so there is nothign to connect back to
You can also do this without editing your ini by exporting values as env variables, just remember to do this each time you start a new shell (or add to your .bash_profile file):
export XDEBUG_CONFIG="remote_host=<YOUR IP>"
Configure PHPStorm's Debugger
It seems by default, PHPStorm doesn't understand remote-cli scripts, so we need to add a configuration that tells it to expect a CLI script to trigger xdebug
Open the Run Menu and select "Edit Configurations"
Click the Green "+" to to add a new configuration and select "PHP Remote Debug"
Name the Configuration (E.G. MyServer-Behat)
Under the Servers: menu, select your remote server.
If you haven't configured your remote server yet, then do this by clicking the "..." button on the right
Click the Green "+" to add a server configuration. Give it a name (E.G. MyServer) and fill in it's address under Host
Configure it's Path Mappings. This is important if the path to your source files is different on your PHPStorm computer from your server. You can see in my example that i'm relating my local checkout (~/Work/Symfony/) to my server install (/var/www/). I specifically added mappings for src, bin, web, app, and vendor by clicking in the space to the right under "Absolute path on the server" and typing in the path. I had issues just mapping the root's, so I had to add these paths to get my debugger to work.
Debug!
Once that is setup, select your configuration from the drop down in the debugging tool bar and click the bug icon (you can also use the Run menu) to start the debugger listening. This is similar to the default Telephone Button (circled in yellow), but it tells PHPStorm to use your new configuration.
Now simply run behat like you normally would from your server and your debugger should connect and stop on any breakpoints you've placed.
If you're having doubts about if it's working or not, try toggling the "Break on First Line" in the Run menu, as this should make the debugger break immediately when you run behat (in the bin/behat file)
Using Google Chrome 38.0.2125.111 m in a Windows 7 environment. It is launched by a batch file & set to go to a specific (internal) URL. The Windows environment is going to be locked down, permitting only a handful of options available to the user.
Chrome is being used because IE will be accessing a different site. Without going into boring detail, using both sites on a single browser allowed a user to access an account that wasn't theirs. Both browsers will be launched using one of two batch files. The Chrome batch file supplies the username & password to bypass a login screen on that application.
To keep from opening multiple windows in Chrome, the following batch file is used:
taskkill /f /t /im chrome.exe > NUL
ping -n 3 127.0.0.1 > NUL
START /max chrome "http://myinternalwebsite?uid=x&pwd=y"
Taskkill is used to close chrome (/f = force /t = terminates process & any child ones /im = image name).
If the Chrome session is closed (using the X at the upper right corner of the screen), the batch file works fine. However, if there is already a chrome instance running, we get the yellow bar with the frowning folder & a message saying:
Google Chrome didn't shut down correctly. To reopen the pages you had
open, click Restore.
I have tried several command line options when launching Google Chrome, but none will keep the message from displaying. Ones I have tried are:
--disable-restore-session-state
--disable-session-crashed-bubble
--incognito
among others. I've experimented with some of the settings in Chrome, but none have (yet) worked. The --incognito option works - sort of. It keeps blocking plugins, one of which is Silverlight - one of the chief components for the website to function. Even though it does get rid of the message, it leaves the site not functioning properly.
The intent on using Taskkill to flush any existing Chrome sessions was to prevent the user opening up a dozen different Chrome instances of the same thing.
Is there any way to keep a user from opening multiple instances of Chrome that won't leave Chrome thinking it didn't close properly? Given the user restrictions, I'm not sure that a script/batch file to edit the Preferences file in the Chrome folder will work. An extension may be an option, but a third party application will not be.
Would appreciate any thoughts or assistance.
Thanks!
Chromium writes exit status into Preferences file. When it starts it writes Preferences file with:
"exited_cleanly": false
and upon clean exit writes Preferences file with:
"exited_cleanly": true
To prevent message about unclean shut down make sure that you have "exited_cleanly": true written in your Preferences file
On my Linux kiosk running on Raspberry Pi this sed command line added to /etc/rc.local fixes issue permanently:
sed -i 's/exited_cleanly\":\ false/exited_cleanly\":\ true/g' /home/pi/.config/chromium/Default/Preferences