I considered using PowerShell indeed, but concluded against it in the end: PowerShell is better than CMD.exe for about everything. But, putting aside the fact that it's only recently universally available (1) it's also very slow to start, which would add a lot of latency to run-program, even though (2) we are not using any of the functionality of the shell, beside starting a program, which CMD does well enough, and (3) on many implementations, CMD is what is used anyway, and adding one more layer of indirection can only introduce discrepancy and bugs.

What exactly do you want to use from PowerShell? If you want to specifically call PowerShell, doesn't the current interface already slow you to do it?

Unless of course you mean using it for scripts that come with ASDF, but there aren't any except the shims to run asdf-tools.

On Tue, Aug 2, 2016, 10:07 Robert Goldman <rpgoldman@sift.net> wrote:
On 8/1/16 Aug 1 -7:06 PM, Jason Miller wrote:
>> My Windows box is now fixed.  Sort of.  But I really don't understand
>> Windows, and have no desire to learn.  I can run the tests, and check
>> their results, but that's about the limit.  I have no intention of
>> fussing around with cmd.exe and the like...
>>
>> Some time ago, Faré suggested that it might be worth replacing use of
>> cmd.exe with powershell in UIOP/run-program.
>>
>> I don't really know what powershell is, but I get the impression it's a
>> shell that's less horrible than cmd.exe.
> Powershell is an improved shell that ships with all versions of windows
> from 7 on, (and server 2008R2 on).  It is available going as far back as
> XP as an optional download, so it would be an external requirement for
> UIOP for Vista and older.

Do any Windows users out there have judgments about how onerous it would
be to impose this requirement?

>From what Jason says, it seems like it might be a nuisance now, but as
we move forward, this should be an acceptable change.

Best,
r