file normalize

file normalize normalizes a filename.

Synopsis

file normalize name

Documentation

official reference

Description

Returns a normalized filename for name. A normalized filename is an absolute filename that

  1. Contains no . or .. components.
  2. Contains no consecutive / characters other than the initial two / characters of a filename that begins with exactly two / characters.
  3. Contains no symbolic links, except that if name refers to a symbolic link, the final component of the normalized filename remains a symbolic link to afford operations on the link itself, such as file delete, file rename, or file copy.
  4. Is in the "standard" format for the native platform, On Windows this is the case-preserved long form of the filename.

Because file normalize transforms embedded symbolic links, it may be overkill for the purpose of just transforming a filename into an absolute filename. Instead, use file join, perhaps with pwd.

See Also

file
file nativename
file join
file readlink

Bugs

1194458
fixed in 8.4.11

Forward Compatibility

file forward compatibility
file normalize emulation code taken from critcl.tcl, and added to the wiki in the hope that more people will use normalize, and move faster to 8.4 and later.

Volume-Relative filenames

Windows has the concept of a current directory per drive, a concept it inherited from DOS. From a Command Prompt you can do:

c:
cd \winnt
d:
type c:setup.log

and see setup.log from the current directory \winnt on drive c:. It's an obscure featrue, and it doesn't work in Windows Explorer.

Therefore, if the current working directory is c:\winnt,

file normalize c:setup.log

results in

C:/WINDOWS/WindowsUpdate.log

Prior to Tcl version 8.4.5, file normalize didn't handle volume relative filenames correctly:

file normalize c:a/b

resulted in c:a/b

2023-08 JMN Has the ability to access this ever been available in Tcl? That 'file pathtype' returns 'volumerelative' only really seems useful in the context of a 'cd' command which can preserve this information when switching volumes. I couldn't find anything in twapi to retrieve this info. It would be great to have cd fixed in this regards.

File normalize and Windows folder junctions

2013-01-08: All Tcl versions before 8.5.13, 8.6.0 have bugs with folder junctions, when the access rights are restricted tcl-Bugs-3092089 , tcl-Bugs-3587096 . They are fixed in Tcl fossil Check-in 8a291bcb44 .

HaO: file normalize will resolve junctions of the Windows NTFS file system and return the filenames without the junctions, if there is a component after the junction. This is specially helpful, if the junction has less access rights than the direct way (which is the case for localized "program files" folders of Windows Vista). I took the habit, to pass any files in system folders by file normalize before accessing them.

Example

Create a folder, C:\test2, and a junction, C:\test2_junction to it.

In a dos box with administrator rights:

C:\Windows\system32> cd c:\
C:\> mkdir test2
C:\> mklink /j test2_junction test2

Now test file normalize in a wish console:

% file normalize c:/test2_junction
c:/test2_junction
% file normalize c:/test2_junction/test.txt
c:/test2/test.txt

Resolving symlinks in the last component of a path

To resolve symlinks in a path's final component (i.e., the target file or directory name itself) you can use the following trick: add /something to the path before normalizing it then strip the extra component away with file dirname.

For example,

set resolvedArgv0 [file dirname [file normalize $argv0/___]]]

dbohdan 2015-05-12: This trick was implemented by AK in Tclssg's main procedure and I thought it deserved wider exposure. The credit is all AK's.

PYK 2015-05-12: This technique is also employed in main script.

Alternative that leaves symlinks alone

mfriedrich How to normalize a path without resolving symlinks? E.g. on Windows mapped network drives where the server uses NTFS junctions pointing to server disks and not to local disks.

AMG: Try repeated [regsub]. Here's code from Wibble, sans call to the Wibble-specific [dehex] command:

regsub -all {(?:/|^)\.(?=/|$)} $path / path
while {[regsub {(?:/[^/]*/+|^[^/]*/+|^)\.\.(?=/|$)} $path "" path]} {}
regsub -all {//+} /$path / path

MHo 2016-02-12: Just realized that file normalize does not work as expected in the following special case:

% glob //?/UNC//wk101w0045/d$ -- *
base-tcl-thread-win32-ix86.dll base-tcl-thread-win32-ix86.exe base-tcl8.6-thread-win32-ix86.dll base-tcl8.6-thread-win32.....
% file normalize //?/UNC//wk101w0045/d$
D://?/UNC//wk101w0045/d$
%

In that case, a Driveletter is prepended, although it should'nt. //?/ is a valid prefix, as is //?//UNC. These are rarely used special cases, though. More important, the //?//... does NOT WORK AT ALL, if the given spec is a DFS-Link....