External lib unittests: they're killin me!

[H. S. Teoh]

On Thu, May 23, 2013 at 06:21:59PM +0200, Kagamin wrote:
> On Thursday, 23 May 2013 at 13:56:26 UTC, Steven Schveighoffer
> wrote:
> >No I mean, I want to pass in all the files in a directory to a
> >windows command line tool.  But the tool has to "opt in" to allow me
> >to use wildcards.
> 
> What if the tool handles argv[1] and exits? Or writes to argv[2] if
> any.
> 
> >No, the shell expands wildcards, not the OS.
> 
> It was just a sarcastic comment about incredible consistency of
> wildcard expansion in linux.

Exactly!!

That's why I said that the chosen solution was wrong, even though the
idea behind it (consistent wildcards across the board) was sound.
Putting this expansion in the shell was just a poor choice. It should
either be done by the program itself (via a common library routine to
ensure consistency) or by the kernel (so exec*() will automatically
inherit it). Putting it in the shell led to a series of consequences
that eventually produced the stupid inconsistencies we have to deal with
today.


T

-- 
If lightning were to ever strike an orchestra, it'd always hit the conductor first.