Re: piping find to zip -- with spaces in path
On Tue, 11 Jan 2011 22:02:48 -0500
Doug <dmcgarrett@optonline.net> wrote:
> On 01/11/2011 08:46 PM, Robert Blair Mason Jr. wrote:
> > On Tue, 11 Jan 2011 14:53:33 -0700
> > Bob Proulx<bob@proulx.com> wrote:
> >
> >> Robert Blair Mason Jr. wrote:
> >>> Rob Owens<rowens@ptd.net> wrote:
> >>>> I tried this and it successfully creates myfile.zip:
> >>>>
> >>>> find ./ -iname "*.jpg" -print | zip myfile -@
> >>>>
> >>>> But it fails if there are spaces in the path or filename. How can I
> >>>> make it work with spaces?
> >>> I think the best way would be to quote them in the pipe:
> >>>
> >>> find ./ -iname "*.jpg" -printf "'%p'\n" | zip myfile -@
> >> But that fails when the filename contains a quote character.
> >>
> >> John's Important File
> >>
> >> Using zero terminated strings (zstrings) are best for handling
> >> arbitrary data in filenames.
> >>
> >> Real Unix(TM) users never put [^[:ascii:]] characters in file names.
> >>
> >> Bob
> > True. Underscores are _wonderful_ things. But remember, Linux is Not Unix!
> >
> > Unfortunately for the OP, i don't *think* zip accepts zstrings. Perhaps a script to just remove all of the non-ascii characters in the filename of all files in the current directory?
> >
> > Random tangent, but pascal strings are often more efficient from a programming standpoint than c-style strings...
> >
> I didn't prune anything because I can't figure out where to do it
> without losing any semblance of coherence, but anyway:
> The comment about real Unixers not using ascii characters: what about
> urls? They come from the Unix world, and are
> full of underscores and question marks and equal signs. Then there are
> emails, all of which require the @ sign. Not
> complaining, just asking.
>
> --doug
>
Well, to be technical, almost all characters *ARE* ascii. Just not the alphanumeric subset.
The underscore is often included in the set of 'ascii characters'. The C Programming Language strikes again! Generally, what is meant by non-ascii-character is any character which might have a special meaning to the users shell.
Usually, however, when you see a question mark and/or equals signs in URLs, they're generally special characters showing that whatever CGI scripting engine the site is running (usually one of the 3Ps: Perl Python PHP) should pass these variables as 'arguments' to the page. So they're not part of the file on the actual server hard disk, just the URL.
A practical example: Googling 'question mark in URL' returns (to me, i'm feeling lucky) http://www.abestweb.com/forums/showthread.php?t=55417 . Which shows us this:
- Protocol HTTP, site www.abestweb.com
- File /forums/showthread.php
- Pass paramater t=5547 to showthread.php
This way, the site doesn't need to make a separate html file to show each thread; it can have a single php file that takes the thread id (t) as a variable when you request it, and then look up this thread's postings in the database and generate HTML from that. Much more efficient.
If you request just showthread.php, you get a valid web page back, but the site says that you did not specify a thread. Handling threads with a script is also much more efficient because it fails much more graciously when you don't specify a thread or the thread is non-existent (a notice box detailing the issue instead of a page not found).
Hope this helps.
--
rbmj
Reply to: