-
Notifications
You must be signed in to change notification settings - Fork 37
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update Add-on Compatibility prefs and show warning when it gets enabled #14
Comments
On 09/02/12 13:19, Brian King wrote:
Dear Brian, I think NTT does not override compatibility checking anymore but I'm not With Compatible by Default you may or may not want to go on overriding Of course if you decide to override compatibility checking, it becomes See also http://kb.mozillazine.org/Extensions.checkCompatibility FWIW, Compatible by Default is the "false" setting of the Best regards, Tony.Brady's First Law of Problem Solving: |
I can still see the force add-on compatibility menu entry in the main menu. So I would say that we still support overriding the compatibility prefs. I also have mixed feelings and have expressed those to Brian and Dave Townsend. While I think the compatibility check removal in ACR was correct because this extension is used by even normal users, we probably should keep the feature in NTT. As Tony mentioned it would allow us to even run extensions which haven't been updated for a while, also those with binary components. |
So I looked at the code and conclude that NTT is setting a different pref than ACR was. ACR set 'extensions.checkCompatibility.nightly', and NTT sets 'nightly.disableCheckCompatibility'. To be honest, I am not sure what the latter does and how it differs from the former. Previously with 'extensions.checkCompatibility.nightly' you could install fresh from AMO add-ons not compatible. A good example (or maybe not because it has binary components) right now of one not updated for Firefox 4+ is: https://addons.mozilla.org/addon/kidzui/ But even with 'nightly.disableCheckCompatibility' set to true, Firefox doesn't allow it to be installed. Tested with Mozilla/5.0 (Windows NT 6.0; rv:11.0) Gecko/20100101 Firefox/11.0 So what I think 'nightly.disableCheckCompatibility' does is allow add-ons already installed but disabled to work again. I think it is fine to leave it then, but would like clarification first on 'nightly.disableCheckCompatibility' does as I can't find documentation on it. I'll ping Blair. |
Hi there, as You can see in nttAddonCompatibilityService.setCompatPrefs() it is a little bit outdated: IMHO if somebody would like to touch Anyway, +1 to keep this compatibility feature in NTT. |
nightly.disableCheckCompatibility isn't used by (and has never been used by) the Add-ons Manager. As with all nightly.* prefs, it's only used by NTT. From a quick look at the code, it's what NTT checks to determine whether it should set the extensions.checkCompatibility.* prefs (see nttAddonCompatibilityService). |
Doh! Obvious I am new to this extension. Maybe what we should do when we flip the prefs is a present a warning that a) old add-ons may not work and b) may cause Firefox to behave badly because of clashes with CBD. What does everyone think? |
Sounds good. I'm happy to fold this into 3.3 |
@brianking can you please let me merge pull requests the next time? The review process wasn't fully done but the merge already happened and the request has been closed. That shouldn't have happened. Also please don't commit directly as given by your last update. Any code change needs a review and the reviewer should check in or say that it's good to be checked in first. Thanks. |
Hm, actually I don't see that anything has landed on mozilla/master. What happened with the last merge and why has the pull request #48 been closed. I'm a bit confused now. |
whimboo commented
See my comment in #15:
It would be interesting how to collaspe and not to pollute: if requester rebases the request, it would became empty. |
Ah wait. Now I see. So you should never ever work on a branch you rebase to and which also exists in the original repository. Looks like a new pull request has to be opened by Brian with the changes he did on pull #48. I for myself will make sure to check the pull details exactly in the future to see such an upcoming situation early enough. Thanks for making it clear. |
Not clearly understand, but surely. :) |
Yes, I messed up. Prompted by a comment by @xabolcs somewhere yesterday, I researched a bit more about having multiple pull requests open at the same time and the way to do it is to have branches on your fork. I had already committed to my master. So what I did is deleted my fork and started fresh. github removing issue 48 was a bit confusing, sorry about that. I'm going to put together the changesets now again, this time on branches. |
Code has been merged. |
I am not sure if NTT is still doing this, or how, but with Compatible By Default (CBD) it should probably not be doing it anymore. This is why the ACR stopped doing it.
Basically, the traditional compatibility prefs and CBD do not play nice together. The dirty details are in this bug:
https://bugzilla.mozilla.org/show_bug.cgi?id=698653
ACR related bug:
https://bugzilla.mozilla.org/show_bug.cgi?id=708931
The text was updated successfully, but these errors were encountered: