gnu-arch-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnu-arch-users] Re: tla-update-ids lacks support for filenames with


From: chth
Subject: Re: [Gnu-arch-users] Re: tla-update-ids lacks support for filenames with spaces
Date: Mon, 24 May 2004 16:33:30 +0200

> On Sat, May 22, 2004 at 05:24:35PM -0500, Charles Duffy wrote:
> > Per subject. Right now tla-update-ids doesn't add tags for filenames
> > with spaces, but rather adds a tag for a (nonexistant) file with the
> > post-escaping name.
> > 
> > Perhaps the simplest course of action would be to pass --unescaped
> > when using a version of tla which supports long file names (and
> > thus, this option)?
> 
> I must admit, I'm confused about how all this escaping stuff is
> supposed to work -- I _thought_ that the escaped forms were supposed
> to be an internal-only thing, restricted to being used in various tla
> files and rel-tables etc.
> 
> Is this not the case?  I remember that the original escaping patch was
> pretty screwed up in this area, but I assumed that had been fixed.
> 
> Are you saying that one must use --unescaped or else output will be
> escaped?

Yes.

tla outputs syntactically unambigous escaped forms by default (slap tom
if you rather want the opposite way arround and have a --escaped). in
general tools need to be aware of that. --escaped vs. --unescaped is
only matter of choice. Parsing tla output in unescaped form where it
might put more than one filename on a line with spaces will suck too and
having the command which only output one file per line unescaped while
the other being escaped would be inconsistent. So no matter how you turn
the thing, tools which use tla as backend need to be aware about spaces
& escapeing in one or another way. While the current implementation the
has the nice sideeffect that it will cause relative easy to spot bugs if
tools are still not escaping aware. 

The escaping stuff is technically unicode aware and can handle things
like \(U+FFFFFFFF) but until tla has a solution for *proper* unicode
handling it is hardwired to 7bit ascii. Further we will not need 2 forms
of escaping and escaped forms of filenames should not go into the
filesystem when they do, there is a bug somewhere.

The 'tla escape' comand is right there for handling such things within
tools and I can only urge to use it.

        Christian




reply via email to

[Prev in Thread] Current Thread [Next in Thread]