From: Charles Roper <reachme@...>
Date: 2010-10-05T17:29:29+09:00
Subject: Re: Strip illegal characters from a file name

I've got a script that strips out invalid characters from TextMate
bundles so that they can be used on Windows. It has worked in all
cases so far and is reversible because it uses URI::escape to encode
the characters. It might be helpful as a starting point for you.

http://gist.github.com/166433

Charles

On 5 October 2010 09:03, Jim Burgess <jack.zelig@gmail.com> wrote:
> Hi,
>
> I have a program which takes a users details and saves them in a pdf.
> The name of the pdf file should consist of the users id, an underscore
> and the users surname, e.g. 01_Smith.pdf
>
> Thing is, how do I check for illegal file name characters in the
> surname?
> On Windows for example, the characters \/:*?"<>| are not allowed in a
> file name, so obviously they shouldn't occur in the surname.
>
> As there seem to be multiple combinations of illegal file name
> characters on different operating systems, I thought about stripping
> everything except letters from the surname and came up with:
>
> surname.gsub(/[^a-zA-Z]/,"")
>
> But as we have users from around the world, it is common for them to
> have special characters in their surname (e.g. äöüß in Germany). The
> above reg exp strips these out too, which isn't perfect.
>
> Is there a reg exp which would strip out everything but letters and
> leave special characters in tact or am I going about this the wrong
> way??
>
> Any tips would be greatly appreciated.
> --
> Posted via http://www.ruby-forum.com/.
>
>