Closed Half-Shot closed 4 months ago
The --cookies
option was added by a PR from a user that was using that functionality (they were using an extension to manually pull their cookies from their browser).
One thing I've seen is that the --browser
option will fail to find the cookie if the browser is installed through a flatpack or a snap, because the directories for those browsers gets placed into a different set of directories in the home dir. A workaround I've been doing where needed for that is to sym link the dir to where it is normally expected to be.
In regards to this PR, there is one issue: --browser
has a default value, so it is always specified, meaning that you'll never get the case where it isn't set but the --cookies
is. I'll think on this and either add a commit to fix it, or you can if you get to it before I do.
@Half-Shot Figured it out. This change will first try to act as the old behavior, and if that fails it'll then try and use the --cookies
path as the store for your given browser.
For whatever reason on my setup, browser_cookie3 was unable to locate my browser's cookie jar when just using
--browser
. I've added the option to use the--cookies
parameter when specifying --browser to remedy this.I'm aware though that this is a breaking change, as anyone specifying both currently will have the browser parameter ignored.
As an aside, do we still think
http.cookiejar.MozillaCookieJar
works? I did try it with my Firefox setup but it was unable to decode my cookie jar, so unless we've got some very old FF users or Netscape users (:grin:), it might not be worth keeping that functionality.