Last modified: 2012-04-19 21:10:41 UTC
As summary said. According to http://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm MSIE versions below 6 have an usage of aprox. 0.13%, which I think is not worthy to support with aditional CSS/JS hacks. Maybe not right now but something to do in future!
Would kill off quite a lot of bugs too...
IE6 is still 12%, so changing subject into "MSIE < 6"
See r61083, r61085, r61128: Dropping IE 5.0 and 5.5 fixes.
I tested the IEMacFixes file against IE 5.2 on OSX. Doesn't seem to make a huge difference in page rendering with or without the stylesheet (the failures are there with and without the file). I wouldn't be opposed to dropping this too.
Yes, IEMac also applies here. But it could be wise to keep support for them until 1.16 be released. After the branch, we could kill all these fixes, and if market share keeps dropping MSIE6 fixes could also be deleted. Somewhere I have seen hacks for old versions of Firefox too. Are they worth?
It was reverted in r62882 and backported to 1.16 in r62928 by simetrical. IE < 6 should be redeleted now in trunk.
(In reply to comment #6) > IE < 6 should be redeleted now in trunk. No, the point of it was exactly the opposite, that IE 5 still has a share although small but still worth at least retaining what we presently have for them. Certainy, this does not mean that IE 5 should e.g. be supported by UsabilityInitiative. Closing for now, suggest revisiting in 1 - 1.5 years.
This bug is listed as fixed in the 1.16 release notes. Should probably remove it since this was reverted.
According to SquidReports MSIE 6 still has 3.46% but MSIE 5.5 has 0.08% and older versions doesn't even appear.
Is having the current IE5 support causing any direct bugs at the moment? I see no reason just to kill of a "feature" because we no longer directly support it if it isn't causing any issues.
(In reply to comment #10) > Is having the current IE5 support causing any direct bugs at the moment? I see > no reason just to kill of a "feature" because we no longer directly support it > if it isn't causing any issues. That's sort of the point When we stop "supporting" a browser, it generally means no new fixes for that browser alone. But if having the support for the older browser was causing issues, then looking at removing it would be a way forward
Removed again in r99934, r100150. They were @todos to be ported to the new format. Dropped instead to save work+bandwidth by not converting them to css hacks in the main stylesheet.
