1802313 - Debugging extension then shutdown computer uninstalls the extension

Status: NEW

(WebExtensions :: Untriaged, enhancement, P5)

Description

[mprops]

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:108.0) Gecko/20100101 Firefox/108.0

Steps to reproduce:

I'm developing an extension with a certain gecko id (so i can use sync storage). So what that means is that when I debug it, it replaces my current installed version until I stop debugging it. The other night I shutdown the computer while it was in debugging mode. Today I turned on the computer and the extension was gone. I had to install it again.

Expected results:

The extension shouldn't be uninstalled.

Comment 1

[mprops]

(This is on linux)

Comment 2

[BugBot [:suhaib / :marco/ :calixte]]

The Bugbug bot thinks this bug should belong to the 'WebExtensions::Untriaged' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 3

[Luca Greco [:rpl] [:luca] [:lgreco]]

The following is a summary of the rationale behind what we discussed and agreed on about this issue today during out triage meeting:

Despite this being

• the currently expected behavior:

temporary installed extensions are uninstalled automatically on browser startup and so when a temporary extension is installed on top of a permanent installed extension and the user shutdown the browser, the extension previously installed permanently is being recognized to be currently temporarily installed and uninstalled

• and technically something only extension developers that are installing a temporary extensions on top of a permanent installed extensions for testing purpose

It is still something that seems reasonable to be improved, e.g. by at least making it a behavior that the developers have to explicitly opting in (e.g. by flipping an about:config pref) so that it doesn't happen too much implicitly and the developers to have a chance to realize that this may have side-effect like this one.

Comment 4

[mprops]

Would it be possible to spawn a separate debugging instance even if the gecko id is the same?
That would make the developing process easier (and it would avoid the problem of this bug)
I'm not sure if that makes sense though.