Opened 10 years ago
Closed 10 years ago
#13228 closed bug (notabug)
Fail with notice if jQuery 2.0 is loaded on oldIE or newIE in compat mode
Reported by: | dmethvin | Owned by: | |
---|---|---|---|
Priority: | undecided | Milestone: | None |
Component: | unfiled | Version: | 2.0b1 |
Keywords: | Cc: | ||
Blocked by: | Blocking: |
Description
In most cases we try to "do no harm" when we run in an unsupported browser, for example by testing for the missing API calls and doing nothing or returning the least messed-up value.
There is so much missing in oldIE I can't imagine that approach will work. I'm thinking we need some clear way to notify the user/dev that they did something wrong. We could pop an alert()
, or perhaps change document.title
.
Change History (10)
comment:1 Changed 10 years ago by
comment:2 Changed 10 years ago by
Status: | new → open |
---|---|
Version: | git → 2.0b1 |
comment:3 Changed 10 years ago by
It might be reasonable to throw "This browser requires jQuery 1.9"... but I'm more inclined to agree with timmywil and see what happens
comment:4 Changed 10 years ago by
I think I have a base idea of where to start implementing this:
It's generally a better idea if the individual web developer is told that their user is using a browser that isn't supported, and then allow the web developer to decide what to do with this information. This would probably be done through a deferred call on a jQuery object.
We could implement this by requiring each base jQuery module to verify necessary apis exist. This could be done with dependency lists, or some other implementation.
Once one of these "dependency lists" is not met, the deferred can be called, and the developer can be notified.
comment:5 Changed 10 years ago by
There would be a lot of overhead to implementing something that general. If the list of browsers dropped ends up being just oldIE, a test as simple as if ( window.attachEvent && !window.addEventListener )
suffices.
comment:6 Changed 10 years ago by
Isn't it a better practice to filter through functionality instead of a per-browser basis?
comment:7 Changed 10 years ago by
We're not really concerned with theoretical browsers or changing implementations here. We're talking about developers loading jQuery 2.x into a known set of browsers (oldIE) and providing a warning. The chances of the proposed check providing a false positive are extremely slim.
comment:9 Changed 10 years ago by
This doesn't feel right. I'm fine with the warnings in the upgrade plugin. Not so much in core.
comment:10 Changed 10 years ago by
Resolution: | → notabug |
---|---|
Status: | open → closed |
I'm gonna close this for now, and if/when complaints come in we can reopen.
Is it out of the question to take each case as it comes and continue trying to do no harm?