[Date Index] [Thread Index] [Author Index]

Re: Path to *.m file

• To: mathgroup at smc.vnet.net
• Subject: [mg86594] Re: Path to *.m file
• From: David Bailey <dave at Remove_Thisdbailey.co.uk>
• Date: Fri, 14 Mar 2008 04:17:48 -0500 (EST)
• References: <200803091007.FAA15259@smc.vnet.net> <fr5e8q$odn$1@smc.vnet.net> <frclpk$rml$1@smc.vnet.net>

Vince Virgilio wrote:
> On Mar 13, 5:32=A0am, David Bailey <dave at Remove_Thisdbailey.co.uk>
> wrote:
>> Carl Woll wrote:
>>> I'm not sure why this is a `Private` variable, but this information can
>>> be found from:
>>> System`Private`$InputFileName
>> This is very useful, and I have checked, and it was there at 5.2, so I
>> guess it is one of those long-term undocumented features!
>>
>> However, it would be nice if it (or the subsequent documented feature)
>> always returned the absolute path-name. As it is, Get["xxx.m"] sets this
>> variable to ".\\xx.m", which is less convenient, and misleading if the
>> code in the package has already executed a SetDirectory call.
>>
>> David Baileyhttp://www.dbaileyconsultancy.co.uk
> 
> Some time ago, I think before 6.0, WRI technical support suggested the
> following to me (ultimately from the developers):
> 
>         System`Private`FindFile[$Input].
> 
> Vince Virgilio
> 
This seems to fail in the case where you reference a file by changing 
directory:

SetDirectory["c:\\maths"]
Get["xxx.m"]

In this case, the function you describe returns $Failed, presumably 
because $Input only returns a string - "xxx.m" in this case - and 
FindFile is looking elsewhere!

I guess it hardly needs saying (to WRI) that the whole value of a 
reliable version of a function like this would be that it would return 
its result without potential confusion of this sort!

David Bailey
http://www.dbaileyconsultancy.co.uk

• References:
  • Path to *.m file
    • From: Frank Scherbaum <Frank.Scherbaum@geo.uni-potsdam.de>