On Fri, Feb 22 2013, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
> He mentions 4 problems with the current situation:
>
>> * The first touches the disk, the second does not.
>
> Providing a wrapper that does make-temp-file doesn't solve this point,
> indeed.
>
>> * The first requires me to pick a temporary file name that does not
>>   collide with an existing file on disk, the second does not.
>> * The first requires me to cleanup a temporary file, the second does
>>   not.
>> * The first is 45 characters long, the second is 13.
>
> But the wrapper would solve all those 3 remaining points (because the
> wrapper would pay this cost, not Jameson).

Thanks again, Stefan.  This is a good analysis of the situation.
Solving these latter three issues is really the crux of things from my
perspective.  The main problem is all the book keeping and extra coding
that is currently required to get stderr into a separate buffer.  It
would obviously better if we didn't have to resort to temporary files,
but if that's the only way to get these last three issues covered then
I'm completely fine with it.

jamie.