Re: [boost] Re: the first version of abstract path manipulation class (interface)

From: Jesse Jones (jesjones_at_[hidden])
Date: 2002-01-25 21:30:35

At 12:22 AM +0800 1/26/02, joel de guzman wrote:
>----- Original Message -----
>From: "George A. Heintzelman" <georgeh_at_[hidden]>
>To: <boost_at_[hidden]>
>Sent: Friday, January 25, 2002 11:40 PM
>Subject: Re: [boost] Re: the first version of abstract path 
>manipulation class (interface)
>
>
>>  VMS is definitely really weird. A typical fully-specified VMS filename
>>  might look something like this:
>>
>>  DISK$SCRATCH:[GEORGE.PROJECT1.DAT]BIG_DATA_FILE.NTP;5
>
>Have anyone considered the classic Macintosh (pre X) where the extension
>does not mean anything, the directory separator is the : and with 
>max 32 chars?
>File paths manipulation is frowned upon since the system actually has
>true file descriptors with meta-file information and does not get invalidated
>when someone moves the file somewhere else while the application is
>working on it.

To expand on this: on the Mac OS (even OS X) it's considered poor 
practice to use path names. Instead FSSpecs or the new FSRef's should 
be used. I've written "path" classes for Mac and Windows before, but 
you have to get away from the notion of a path and take a higher 
level view of the problem space. You'd also have to pollute the 
classes with some platform specific code since Mac users will want to 
initialize the class with an FSRef and get the FSRef the object 
points to.

   -- Jesse

--