Latest Beta

[James]

But as least we don't have to worry about running another program on our code to check for these sorts of errors, eh, James? (Edit: Or was it Richard who made that comment? I forget)

Richard.

I'm fully pro-debuggers and stuff

[Richard Robertson]

Maybe you should write a regular expression to search the entire source for variable declarations.

See how many of them don't get initialized to something (like 0).

[Valik]

Richard is sort of on the right track. Jon, I believe the Code Analysis thing has customizable rules. You might look around when you're bored and see if the community has made any rule packages that we could use to catch some more common errors.

By the way Richard, uninitialized variables are only an issue when the type is either a pointer or non-class type. Initializing a class object without explicitly assigning it a value is not an error but a regular expression is too simple to understand that and would complain.

[Richard Robertson]

I didn't think you'd be storing any class objects on the local stack.

I always refer to classes by pointers. I've found it alleviates a number of problems people run in to.

[Valik]

You know those dumb things you say? This is one of them. I'm not really interested in giving up automatic object destruction when it goes out of scope in exchange for storing things on the heap and releasing them explicitly when the stack will do.

[GEOSoft]

And let those who would complain about no recent beta releases be silenced.

Thanks Jon.

[AdmiralAlkex]

What happened with #1573? First Jpm fixes it but it doesn't work (nor is included in the changelog) and now Jon reopened it and three others (all by Jpm).

Also I find this interesting:

AutoIt Changelog

================

3.3.8.0 (some day, 2010) (Release)

Jon has a time-machine in his backyard?

[GEOSoft]

Sounds like they weren't fixed properly. That would be my first guess anyway.

[Jon]

Short answer, I didn't like the look of the fixes at first glance and reverted them. They may be ok and reintroduced at some point.

[Jon]

Jon, I believe the Code Analysis thing has customizable rules. You might look around when you're bored and see if the community has made any rule packages that we could use to catch some more common errors.

I had a quick look. When you go into the rules they are all managed code related so it seems like when you run on a c++ project you are getting generic rules. That's why MS Minimum vs MS All Rules was giving the same output.

[Jon]

3.3.7.7 (31st May, 2011) (Beta)

Also now being compiled with VC10 profiling so should be faster than previous betas and faster than 3.3.6.1.

AutoIt:

- Fixed #1586: DllStructSetData(char, 1, string-with-NUL-character), junk data(?)

- Fixed #1633: Second AdlibRegister function starts at once when first function is too slow.

Au3Info:

- Added: Option "Use Spy++ control detection logic" toggle added to Au3Info (this was the mode always used in recent

releases, but the very old method that used to be used is useful in certain applications (for example, putty.exe)

- Fixed #1746: Option "Use Spy++ control detection logic" toggle added to Au3Info.

Misc:

- x64 installer options tweaked.

Edited May 31, 2011 by Jon

[Valik]

Since Jon didn't mention it, I will. Test @CPUArch. Anyone reading this create a quick script with MsgBox(4096, "", @CPUArch) in it and let us know if it works.

[Mat]

I'm showing X64 on a 32 bit system.

Edit: But @OSArch returns X86. Maybe I've got a x64 cpu and never knew XD

Edited May 31, 2011 by Mat

[Jon]

I'm showing X64 on a 32 bit system.

Edit: But @OSArch returns X86. Maybe I've got a x64 cpu and never knew XD

Most CPUs from the last few years are x64 capable (Test vs. 3.3.6.1 to be sure)

[water]

I get "x64" with 3.3.6.1, 3.3.7.6 and 3.3.7.7 (installed for both X86 and X64 OS support).

[Mat]

Most CPUs from the last few years are x64 capable (Test vs. 3.3.6.1 to be sure)

Same result

Wow. I've always just used @OSArch.

Edit: Just double checked systeminfo and it's right Now I'm off to install a 64 bit operating system and try it out

Edited May 31, 2011 by Mat

[jchd]

I get a correct X64 for this Intel Core Duo E8400 running x86 XP

[Zedna]

Correct value X86/X86 for @OSArch/@CPUArch on my WINXP/P4 NorthWood without EMT64

Edited May 31, 2011 by Zedna

[jaberwacky]

Correct value: X64 for both @CPUArch & @OSArch

Windows 7 Professional 64

AMD Athlon 64 X2 5000+ Black Edition

[danielkza]

I get correct results too from both @CPUArch and @OSArch on 3.3.7.7 (Both X64 on Windows 7 x64 SP1 + Core i5 2500K).

Edited May 31, 2011 by danielkza