Hadi Hariri - All Comments

re: Profiling Apps 1 of N: The MVC ActionLink

Tim Barcz — Thu, 11 Mar 2010 04:57:38 GMT

Hadi,

This was a great read, thanks for sharing.

One thought for future performance topics (one that's been on my mind) is the RenderAction versus RenderPartial in the MVC framework since the two operate differently.

Thanks again,

Tim

Profiling Apps 1 of N: The MVC ActionLink - Hadi Hariri - Devlicio.us - Just the Tasty Bits

9eFish — Wed, 10 Mar 2010 15:43:23 GMT

9efish.感谢你的文章 - Trackback from 9eFish

re: Profiling Apps 1 of N: The MVC ActionLink

Hadi Hariri — Wed, 10 Mar 2010 06:46:27 GMT

@Paco,

The point with covering ActionLinks and comparing the two were to see what the real difference is between them. Many times one has been discarded in favour of other for performance reasons, whereas that performance difference might not even be a real issue.

Obviously context is important.

re: Profiling Apps 1 of N: The MVC ActionLink

Paco — Tue, 09 Mar 2010 22:16:32 GMT

You don't have a performance problem when method a is slower than method b, but you have a performance problem when your application is too slow to agree with your SLA. When you have a webpage with a lot of links (count the links on this page), I can imagine situations where 100ms is significant. When your SLA says that 100ms is fine, you don't have to worry, even when 99 of that 100ms are involved in generating links. It is a reason to be curious to start the profiler, but it is not a performance problem.

When you want to compare two approaches of doing something, you have to realize you just compare it in the context where you use it. The two ways of generating links are almost equally fast in your situation, but that doesn't mean that there is no significant difference in every situation. When you really want to compare the two approaches, you should create a chart with all involved parameters, and show which approach is the fastest in which situation.

I once had a performance problem with generating links in mvc. I opened reflector, and saw that the largest difference of generating the links with the 2 approaches in my situation was caused by RouteCollection.GetVirualPath. With some link generating methods, all routes would be enumerated here, and with some there would be a look up on routename. The application I was working had a few hundreds of views and about 100-200 routes. The difference was much larger than the difference in your situation, something like 2 whole seconds vs some milliseconds for 1000 links. I solved the problem by creating my own strongly typed version of the actionlink method.

There is no performance problem when the app is not to slow.

There is no context in a benchmark to compare two methods, but there is a context in profiling an application.

re: Dynamic types and ASP.NET MVC

Peter — Sun, 07 Mar 2010 13:19:28 GMT

Looks great. I think it's usable for scenario's where you have dynamic formcontent and using this to create a dynamic model could be very handy.

re: ASP.NET Support in ReSharper 5

renantech — Mon, 08 Feb 2010 01:00:16 GMT

asp.net 2010 version is great because of the improve features and easy to use.

re: ASP.NET Support in ReSharper 5

Hadi Hariri — Thu, 21 Jan 2010 11:27:32 GMT

@David,

It was a joke :). I appreciate that a lot of people are used to the IntelliJ setup. In fact I think it was Howard that mentioned that it would be a good idea to have a global key to change schemes. Would help pairing up with devs using different layouts.

re: ASP.NET Support in ReSharper 5

David Kemp — Thu, 21 Jan 2010 11:09:36 GMT

re: If you’re using IntelliJ (why?)

I use this because I've been using R# forever, and originally there wasn't a choice over keyboard mappings, so I learnt the Intelli-j ones and now CBA to relearn them just because VS started playing catch-up

re: ASP.NET Support in ReSharper 5

sergiopereira — Wed, 20 Jan 2010 17:43:16 GMT

Good stuff. I shall upgrade ASAP.

re: Writing plug-ins for ReSharper: Part 1 of Undefined

Pascal — Sat, 16 Jan 2010 16:08:02 GMT

Great post. For a while now I've been wanting to create a R# plugin but it always seemed too complicated and undocumented. I think your posts on the topic will help a lot. Please keep up with the series. Thanks.

re: Writing plug-ins for ReSharper: Part 1 of Undefined

Andrew Kazyrevich — Fri, 15 Jan 2010 17:08:42 GMT

Hadi, great start.

R# OpenAPI definitely lacks documentation - hopefully your blog posts would shed some light!

--Andrew

re: Writing plug-ins for ReSharper: Part 1 of Undefined

Gamlor — Wed, 13 Jan 2010 12:29:34 GMT

Hey, a cool blog post. I'll certainly follow the next post on this topic.

And adding a context-action is easier than I imagined.

Is there a good documentation of the Resharper-API somewhere. I haven't found anything on Resharpers website.

re: Writing plug-ins for ReSharper: Part 1 of Undefined

web design — Wed, 13 Jan 2010 10:33:04 GMT

what a nice blog dear. I never seen such an attractive and useful blog after seeing it. I will pray for your success and also appreciate you at your great step that you have taken toward your bright future..

Thanks a lot again.

re: Writing plug-ins for ReSharper: Part 1 of Undefined

Kenny — Wed, 13 Jan 2010 08:00:52 GMT

Cool stuff! :)

re: Writing plug-ins for ReSharper: Part 1 of Undefined

Hadi Hariri — Tue, 12 Jan 2010 20:23:35 GMT

@Sergio,

Unforunately Live Writer didn't play well with the Attachment plugin. The source code had all the correct references. I've uploaded the project to hadihariri.com/.../MakeVirtualPlugin.zip. Let me know if you have any issues.