826942 - AppInstallManager tells user app is installed once it's only just been downloaded

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Gaia::System, defect)

Description

[Myk Melez [:myk] [@mykmelez]]

AppInstallManager calls its showInstallSuccess() method when its handleDownloadSuccess() method is invoked on downloadsuccess.  But that seems too soon, since an app hasn't been installed until the download has been applied (i.e. on downloadapplied).  So AppInstallManager should wait until then to notify the user that the app has been installed.

Ben: it looks like you made this change in https://github.com/mozilla-b2g/gaia/commit/41c360b2f794b57832b3f1be72dcd9edae701bc7.  It seems like fixing this would be as simple as adding a downloadapplied handler, but I'm not very familiar with this code.  What do you think?

Comment 1

[Jason Smith [:jsmith]]

Tracker bug for this one is "app-install."

What's the risk here?

Comment 2

[Myk Melez [:myk] [@mykmelez]]

I can be confusing, but I don't think it breaks anything.  If the install would otherwise succeed (i.e. it isn't failing due to other bugs), it'll still succeed.

And, in the update case, which is what I'm testing, if the user is prompted to install an update before the update is downloaded, then downloadsuccess and downloadapplied will follow in such close succession that the user probably won't notice.

But if the user is prompted to install an update *after* the update is downloaded, which can happen if an app itself calls checkForUpdate() and then download(), then this bug will cause Gaia to tell the user that the app was installed and *then* prompt the user to install the update (or, currently, because of 826946, to redownload the update).  Which will be confusing.

Comment 3

[Julien Wajsberg [:julienw]]

In the install case, we always get downloadsuccess and downloadapplied in sequence so I don't think this could change anything. In the platform, the download for an install is applied before we get downloadsuccess.

I actually had to choose between these two when I coded this but this really makes no difference. Also, I'm not sure but IIRC I even chosed downloadsuccess for a reason (that I can't remember now).

The update case is different, and this is Bug 826946.

So I'd suggest closing this bug because I don't see what this change would change.