OpenEuphoria

Re: walk_dir: upper & mixed case problem in "equal" test

• Posted by Dan B Moyer <DANMOYER at PRODIGY.NET> Sep 10, 2000
• 402 views

<sigh> nevermind.

It's so obvious it even occurred to me (in addition to others who responded
privately), that if I just put my target in uppercase, and then "upper()"
the return, I can validly test it just fine.  <sigh>

Dan


----- Original Message -----
From: "Dan B Moyer" <DANMOYER at prodigy.net>
To: "Euphoria Mail List" <EUPHORIA at LISTSERV.MUOHIO.EDU>
Sent: Sunday, September 10, 2000 8:26 PM
Subject: walk_dir: upper & mixed case problem in "equal" test


> I'm trying to use walk_dir to find the path for a specified file, and I'm
> having trouble getting my program to *recognize* when it's found the file,
> because of difficulty with uppercase/lowercase as perceived  & reported by
> walk_dir, and then tested by "equal".
>
> It seems that sometimes when a file is created, its filename is created in
a
> format which is read by walk_dir as ALL UPPERCASE, (such as all the
> Euphoria files distributed by RDS), although they DON'T display that way
on
> all systems (ie, Win98) when shown in a directory listing (ie, they show
as
> all lowercase except for the first character); however, sometimes (perhaps
> when the file is created on Win98 system or something?), the filename *as
> reported by walk_dir* can be in a *mixture* of upper and lowercase, which
is
> like it IS displayed in a directory listing on the system.
>
> So the upshot is, if a person goes to use the routine, and types in the
file
> name to find, if they don't type it in exactly the form it is returned by
> walk_dir (all uppercase, all lowercase, or *mixture*), I can't get the
> routine to recognize that it has found it.
>
> I presume I could make some routine to take each file name as reported by
> walk_dir and create *every* possible combination of upper/lower case for
> every character in it,  to do the test against the target filename, but
this
> seems like it would make the walk_dir take a LOT longer to do.
>
> Can anyone suggest a better approach?
>
> Thanks,
> Dan Moyer
>
>
>
>
>
>