Re: GNU bash, 3.00.15(1)-release, referenced cmd in cwd executes alternate cmd

[Бојан Ландекић]

On 27-Feb-07, at 2:26 PM, Matthew Woehlke wrote:

> IIRC we were talking about why e.g. 'ls' would not run 'which ls', 
> right? In that case "correct" could be said to mean that 'which <foo>' 
> ("which" is assumed to execute in a new environment, that is it is not 
> implemented as a shell alias or shell function) and 'type foo' agree. 
> This does not happen if (as I just experienced) you start bash, run 
> the command, then install a new version of that command in a 
> different, but also-in-PATH location.

That is sort of what I reported as a bug when i started this email 
thread.  Do you consider this a bug, incorrect behavior, a feature 
result or something else?  It's not expected behavior from my miniscule 
bash experience.

> No, but I maintain a small login script and a much larger toolchain of 
> (mostly-) GNU software that achieves approximately the same effect. 
> :-) But it doesn't really help with "true portability", since you then 
> depend on that particular toolchain being present. It's great for 
> personal use and build systems, though, where you can enforce its 
> availability.

Is bash truly portable (self-contained)?  Is it possible to put bash on 
some portable media and use it on a system that is bash-less without 
worrying about dependencies?