[4.x] Serious performance issues under IE 8

[MrSparks]

4.1.2 GA, The release notes don't refer to any Performance gains. Can you confirm what the status of this issue is?

4.1.2 was supposed to be a performance release, it just looks like a stack of bug fixes..?

Regards
MrSparks

[skirtle]

4.1.2 was supposed to be a performance release, it just looks like a stack of bug fixes..?

I don't believe 4.1.2 was ever intended to be a performance release. The 4.1.x line will more than likely just be bug fixes. Personally I'm expecting 4.2 to contain the next wave of performance enhancements.

[MrSparks]

I don't believe 4.1.2 was ever intended to be a performance release. The 4.1.x line will more than likely just be bug fixes. Personally I'm expecting 4.2 to contain the next wave of performance enhancements.

Ah yes 4.2, sorry my mistake. Do we have an very approx date for 4.2 ... Q4 2012 / Q1 2013 ?

Best
MrSparks

[rich02818]

I don't believe 4.1.2 was ever intended to be a performance release. The 4.1.x line will more than likely just be bug fixes. Personally I'm expecting 4.2 to contain the next wave of performance enhancements.

Well, last November Mullany wrote:

"We are working as aggressively as possible and will continue to release performance improvements in 4.1.x and 4.2 releases until we achieve our performance goals." so this disagrees with bug fixes only in 4.1.x.

[dongryphon]

Yes, 4.2 is the next significant release for performance work.

Please see http://www.sencha.com/forum/showthre...ents-in-4.next

No official date, but St. Nick might be bringing something extra in his sleigh ...

[dongryphon]

Well, last November Mullany wrote:

"We are working as aggressively as possible and will continue to release performance improvements in 4.1.x and 4.2 releases until we achieve our performance goals." so this disagrees with bug fixes only in 4.1.x.

It is not that we will reject changes that only have performance gains in 4.1.x, but we have to be sure that each release of 4.1.x is a stability improvement over the previous. So we are just going to be cautious there and hence won't be taking on the larger/riskier changes on that branch.

[rich02818]

It is not that we will reject changes that only have performance gains in 4.1.x, but we have to be sure that each release of 4.1.x is a stability improvement over the previous. So we are just going to be cautious there and hence won't be taking on the larger/riskier changes on that branch.

So the emphasis differs from what Mullany wrote. Why would your logic not also apply to 4.2?

[dongryphon]

So the emphasis differs from what Mullany wrote. Why would your logic not also apply to 4.2?

I don't see how the emphasis is different there. We are going to have performance improvements in 4.1.x and 4.2 but with different levels of risk tolerance and pro-activity.

With 4.2 we will have a beta process to help vet out changes and fix regressions. We don't have that (really) with 4.1.x releases. That allows us to take on more risk in 4.2.

[rich02818]

I don't see how the emphasis is different there. We are going to have performance improvements in 4.1.x and 4.2 but with different levels of risk tolerance and pro-activity.

With 4.2 we will have a beta process to help vet out changes and fix regressions. We don't have that (really) with 4.1.x releases. That allows us to take on more risk in 4.2.

I understand your caution and see the necessity. However, Mullany was putting the emphasis on fixing the performance problems "as aggressively as possible" ... "until we achieve our performance goals." I'm quite concerned considering that in his comments in the blog on how to achieve performance on v4, Animal stated

"In my opinion, the performance restriction now is entirely imposed by the Javascript engine being used."

There were some direct responses to this claim, but nothing further from Sencha. I'm counting on the fact that Animal's opinion is not the official Sencha position, but I get nervous when the next build released does not apparently address performance at all...

[dongryphon]

I understand your caution and see the necessity. However, Mullany was putting the emphasis on fixing the performance problems "as aggressively as possible" ... "until we achieve our performance goals." I'm quite concerned considering that in his comments in the blog on how to achieve performance on v4, Animal stated

"In my opinion, the performance restriction now is entirely imposed by the Javascript engine being used."

There were some direct responses to this claim, but nothing further from Sencha. I'm counting on the fact that Animal's opinion is not the official Sencha position, but I get nervous when the next build released does not apparently address performance at all...

We do have an aggressive timeline for 4.2 if that helps, but I certainly understand your concern.

The comment you cite is not an official statement from Sencha on what is possible in this regard. That statement is just a "tautology" - performance of JS code is always limited by the performance of the JS engine.

That does not mean the JS code in question cannot accomplish its required results faster or more efficiently. In early work on several of these areas, we are clearly seeing that further, significant improvement is possible. If any of those experiments can be "back-ported" to 4.1.x, we will certainly consider doing so.