[BusyBox] busybox "ash" doesn't report "not found"

[Friedrich Lobenstock]

*busybox mailinglist was CC, so answer goes back to the list
despite I haven't seen this mail there*

Rob Landley wrote on 06.01.2004 01:46 MET:
> On Monday 05 January 2004 08:17, Friedrich Lobenstock wrote:
> 
> 
>>>Perhaps you could provide some details at to what is broken
>>>in ash in -pre5?  It works nicely for me, and my wireless
>>>access point is very happy with it,
>>
>> From what I can read on the list the following stuff has been broken
>>in pre4/pre5 (might have been fixed but it is IMO not a very good sign
>>if things break that worked before):
> 
> 
> I take this as a sign that you've never encountered the concept of 
> "development" before, have you?  (This is why regression testing exists.  
> However, no regression test suite is perfect.  This is why -pre releases are 
> put out rather than going "we changed some stuff and we know it's perfect, 
> here's a release!")
> 
> 
>>(have not tested the following in pre3, but they've been found broken in
>>pre4/5) * sed
> 
> 
> There was a major overhaul of sed in the long pause after -pre3, and a LOT of 
> things the old sed code couldn't do are now handled by the new code.  The new 
> one actually survives usage by ./configure scripts for some packages (not all 
> yet).
> 
> The _only_ complaint I've gotten about it was due to two lines being in the 
> wrong place, and it was fixed as soon as I got a test case.  The fix is in 
> CVS now.  If anybody else has a test case where the SuS spec for sed says do 
> one thing (backed up by a sed implementation like gnu sed), and we do 
> something else, give it to me and I'll fix it.  (I have one pending test case 
> that I might still fix before release, but it's really a gnu extension we 
> don't support rather than part of the SuS description of sed.)
> 
> If you know of a real problem, gimme a test case and I'll fix it.

Forgive me if I mentioned that sed was one of the things that got
broken. This was not ment to be an offence to you or anybody.

-- 
MfG / Regards
Friedrich Lobenstock