Using undefined variables in JScript

re: Using undefined variables in JScript

Eric Lippert — Sat, 06 May 2006 09:05:01 GMT

1) Not to my knowledge, but I am on neither the IE nor the scripting teams anymore, so I am not privy to their strategic or tactical planning.

2) No, it is not deprecated to my knowledge, and I am hoping that we will be making it a higher priority to invest in JScript .NET in several areas -- I would also like to see JS.NET on the IE client, as well as support for e4x features and new CLR features such as generics. However, note that this is my personal hope, and not based on any information one way or the other, see #1.

re: Using undefined variables in JScript

bmm6o — Sat, 06 May 2006 03:22:26 GMT

Ah, of course. Once you expanded it to "evaluate the LHS" the light bulb went off (or on, rather). I was stuck on thinking of doing just a global variable lookup.

re: Using undefined variables in JScript

Pete T — Fri, 05 May 2006 21:53:48 GMT

Slightly off-topic, but could you clarify:
1. Is IE 7 due to include JScript.Net (i.e. client access to the frameworks from page scripts)?
2. Is JScript.Net an officially deprecated language on the server, or is it just not a priority for Microsoft?

Personally I think the ability to use the same syntax on client and server is great for UI Coders who need to code in both environments to produce a modern, efficient UI.

re: Using undefined variables in JScript

Eric Lippert — Fri, 05 May 2006 01:54:11 GMT

There are three things that have to happen in the right order:

1) Evaluate the LHS. (Where's it going?)
2) Evaluate the RHS. (What's going there?)
3) Put it there, creating "there" if necessary.

I was asking the question "why does #3 have to happen after #2?"

Because if #2 produces an exception, then #3, which potentially has the side effect of altering the global name table, must never happen.

You are asking the question "why does #2 have to happen after #1?", and the answer is pretty much the same == because evaluating the left hand side can produce arbitrary errors and those errors must happen BEFORE any side effects produced by evaluating the RHS. When we ask the host "do you have y?" the host is free to say "throw an exception" rather than simply saying "nope", and if the host does that, then we need to make sure that any exception catcher runs without the side effects of the right hand side having happened.

re: Using undefined variables in JScript

bmm6o — Fri, 05 May 2006 01:32:44 GMT

I understand the point you make at the end, but why do you have to do the lookup before you have a value to store? Is there a reason you can't evaluate the right side first, and then look up y?