File: pathmanagement.html

package info (click to toggle)
abs-guide 10-4
links: PTS, VCS
area: main
in suites: bookworm, bullseye, forky, sid, trixie
size: 6,952 kB
sloc: sh: 14,129; makefile: 81
file content (383 lines) | stat: -rw-r--r-- 11,074 bytes
parent folder | download | duplicates (4)
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<HTML
><HEAD
><TITLE
>Parsing and Managing Pathnames</TITLE
><META
NAME="GENERATOR"
CONTENT="Modular DocBook HTML Stylesheet Version 1.76b+
"><LINK
REL="HOME"
TITLE="Advanced Bash-Scripting Guide"
HREF="index.html"><LINK
REL="PREVIOUS"
TITLE="Awk"
HREF="awk.html"><LINK
REL="NEXT"
TITLE="Exit Codes With Special Meanings"
HREF="exitcodes.html"><META
HTTP-EQUIV="Content-Style-Type"
CONTENT="text/css"><LINK
REL="stylesheet"
HREF="common/kde-common.css"
TYPE="text/css"><META
HTTP-EQUIV="Content-Type"
CONTENT="text/html; charset=iso-8859-1"><META
HTTP-EQUIV="Content-Language"
CONTENT="en"><LINK
REL="stylesheet"
HREF="common/kde-localised.css"
TYPE="text/css"
TITLE="KDE-English"><LINK
REL="stylesheet"
HREF="common/kde-default.css"
TYPE="text/css"
TITLE="KDE-Default"></HEAD
><BODY
CLASS="APPENDIX"
BGCOLOR="#FFFFFF"
TEXT="#000000"
LINK="#AA0000"
VLINK="#AA0055"
ALINK="#AA0000"
STYLE="font-family: sans-serif;"
><DIV
CLASS="NAVHEADER"
><TABLE
SUMMARY="Header navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TH
COLSPAN="3"
ALIGN="center"
>Advanced Bash-Scripting Guide: An in-depth exploration of the art of shell scripting</TH
></TR
><TR
><TD
WIDTH="10%"
ALIGN="left"
VALIGN="bottom"
><A
HREF="awk.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="80%"
ALIGN="center"
VALIGN="bottom"
></TD
><TD
WIDTH="10%"
ALIGN="right"
VALIGN="bottom"
><A
HREF="exitcodes.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="100%"></DIV
><DIV
CLASS="APPENDIX"
><H1
><A
NAME="PATHMANAGEMENT"
></A
>Appendix D. Parsing and Managing Pathnames</H1
><P
>Emmanual Rouat contributed the following example of parsing
        and transforming <I
CLASS="FIRSTTERM"
>filenames</I
> and, in
        particular, <A
HREF="special-chars.html#PATHNAMEREF"
>pathnames</A
>. It draws
        heavily on the functionality of <I
CLASS="FIRSTTERM"
>sed</I
>.</P
><P
><TABLE
BORDER="0"
BGCOLOR="#E0E0E0"
WIDTH="100%"
><TR
><TD
><PRE
CLASS="PROGRAMLISTING"
>   1&nbsp;#!/usr/bin/env bash
   2&nbsp;#-----------------------------------------------------------
   3&nbsp;# Management of PATH, LD_LIBRARY_PATH, MANPATH variables...
   4&nbsp;# By Emmanuel Rouat &#60;no-email&#62;
   5&nbsp;# (Inspired by the bash documentation 'pathfuncs' and on
   6&nbsp;# discussions found on stackoverflow:
   7&nbsp;# http://stackoverflow.com/questions/370047/
   8&nbsp;# http://stackoverflow.com/questions/273909/#346860 )
   9&nbsp;# Last modified: Sat Sep 22 12:01:55 CEST 2012
  10&nbsp;#
  11&nbsp;# The following functions handle spaces correctly.
  12&nbsp;# These functions belong in .bash_profile rather than in
  13&nbsp;# .bashrc, I guess.
  14&nbsp;#
  15&nbsp;# The modular aspect of these functions should make it easy
  16&nbsp;# to expand them to handle path substitutions instead
  17&nbsp;# of path removal etc....
  18&nbsp;#
  19&nbsp;# See http://www.catonmat.net/blog/awk-one-liners-explained-part-two/
  20&nbsp;# (item 43) for an explanation of the 'duplicate-entries' removal
  21&nbsp;# (it's a nice trick!)
  22&nbsp;#-----------------------------------------------------------
  23&nbsp;
  24&nbsp;# Show $@ (usually PATH) as list.
  25&nbsp;function p_show() { local p="$@" &#38;&#38; for p; do [[ ${!p} ]] &#38;&#38;
  26&nbsp;echo -e ${!p//:/\\n}; done }
  27&nbsp;
  28&nbsp;# Filter out empty lines, multiple/trailing slashes, and duplicate entries.
  29&nbsp;function p_filter()
  30&nbsp;{ awk '/^[ \t]*$/ {next} {sub(/\/+$/, "");gsub(/\/+/, "/")}!x[$0]++' ;}
  31&nbsp;
  32&nbsp;# Rebuild list of items into ':' separated word (PATH-like).
  33&nbsp;function p_build() { paste -sd: ;}
  34&nbsp;
  35&nbsp;# Clean $1 (typically PATH) and rebuild it
  36&nbsp;function p_clean()
  37&nbsp;{ local p=${1} &#38;&#38; eval ${p}='$(p_show ${p} | p_filter | p_build)' ;}
  38&nbsp;
  39&nbsp;# Remove $1 from $2 (found on stackoverflow, with modifications).
  40&nbsp;function p_rm()
  41&nbsp;{ local d=$(echo $1 | p_filter) p=${2} &#38;&#38;
  42&nbsp;  eval ${p}='$(p_show ${p} | p_filter | grep -xv "${d}" | p_build)' ;}
  43&nbsp;
  44&nbsp;#  Same as previous, but filters on a pattern (dangerous...
  45&nbsp;#+ don't use 'bin' or '/' as pattern!).
  46&nbsp;function p_rmpat()
  47&nbsp;{ local d=$(echo $1 | p_filter) p=${2} &#38;&#38; eval ${p}='$(p_show ${p} |
  48&nbsp;  p_filter | grep -v "${d}" | p_build)' ;}
  49&nbsp;
  50&nbsp;# Delete $1 from $2 and append it cleanly.
  51&nbsp;function p_append()
  52&nbsp;{ local d=$(echo $1 | p_filter) p=${2} &#38;&#38; p_rm "${d}" ${p} &#38;&#38;
  53&nbsp;  eval ${p}='$(p_show ${p} d | p_build)' ;}
  54&nbsp;
  55&nbsp;# Delete $1 from $2 and prepend it cleanly.
  56&nbsp;function p_prepend()
  57&nbsp;{ local d=$(echo $1 | p_filter) p=${2} &#38;&#38; p_rm "${d}" ${p} &#38;&#38;
  58&nbsp;  eval ${p}='$(p_show d ${p} | p_build)' ;}
  59&nbsp;
  60&nbsp;# Some tests:
  61&nbsp;echo
  62&nbsp;MYPATH="/bin:/usr/bin/:/bin://bin/"
  63&nbsp;p_append "/project//my project/bin" MYPATH
  64&nbsp;echo "Append '/project//my project/bin' to '/bin:/usr/bin/:/bin://bin/'"
  65&nbsp;echo "(result should be: /bin:/usr/bin:/project/my project/bin)"
  66&nbsp;echo $MYPATH
  67&nbsp;
  68&nbsp;echo
  69&nbsp;MYOTHERPATH="/bin:/usr/bin/:/bin:/project//my project/bin"
  70&nbsp;p_prepend "/project//my project/bin" MYOTHERPATH
  71&nbsp;echo "Prepend '/project//my project/bin' \
  72&nbsp;to '/bin:/usr/bin/:/bin:/project//my project/bin/'"
  73&nbsp;echo "(result should be: /project/my project/bin:/bin:/usr/bin)"
  74&nbsp;echo $MYOTHERPATH
  75&nbsp;
  76&nbsp;echo
  77&nbsp;p_prepend "/project//my project/bin" FOOPATH  # FOOPATH doesn't exist.
  78&nbsp;echo "Prepend '/project//my project/bin' to an unset variable"
  79&nbsp;echo "(result should be: /project/my project/bin)"
  80&nbsp;echo $FOOPATH
  81&nbsp;
  82&nbsp;echo
  83&nbsp;BARPATH="/a:/b/://b c://a:/my local pub"
  84&nbsp;p_clean BARPATH
  85&nbsp;echo "Clean BARPATH='/a:/b/://b c://a:/my local pub'"
  86&nbsp;echo "(result should be: /a:/b:/b c:/my local pub)"
  87&nbsp;echo $BARPATH</PRE
></TD
></TR
></TABLE
></P
><P
>***</P
><P
>David Wheeler kindly permitted me to use his instructive
        examples.</P
><P
><TABLE
BORDER="0"
BGCOLOR="#E0E0E0"
WIDTH="100%"
><TR
><TD
><PRE
CLASS="PROGRAMLISTING"
>   1&nbsp;Doing it correctly: A quick summary
   2&nbsp;by David Wheeler
   3&nbsp;http://www.dwheeler.com/essays/filenames-in-shell.html
   4&nbsp;
   5&nbsp;So, how can you process filenames correctly in shell? Here's a quick
   6&nbsp;summary about how to do it correctly, for the impatient who "just want the
   7&nbsp;answer". In short: Double-quote to use "$variable" instead of $variable,
   8&nbsp;set IFS to just newline and tab, prefix all globs/filenames so they cannot
   9&nbsp;begin with "-" when expanded, and use one of a few templates that work
  10&nbsp;correctly. Here are some of those templates that work correctly:
  11&nbsp;
  12&nbsp;
  13&nbsp; IFS="$(printf '\n\t')"
  14&nbsp; # Remove SPACE, so filenames with spaces work well.
  15&nbsp;
  16&nbsp; #  Correct glob use:
  17&nbsp; #+ always use "for" loop, prefix glob, check for existence:
  18&nbsp; for file in ./* ; do          # Use "./*" ... NEVER bare "*" ...
  19&nbsp;   if [ -e "$file" ] ; then    # Make sure it isn't an empty match.
  20&nbsp;     COMMAND ... "$file" ...
  21&nbsp;   fi
  22&nbsp; done
  23&nbsp;
  24&nbsp;
  25&nbsp;
  26&nbsp; # Correct glob use, but requires nonstandard bash extension.
  27&nbsp; shopt -s nullglob  #  Bash extension,
  28&nbsp;                    #+ so that empty glob matches will work.
  29&nbsp; for file in ./* ; do        # Use "./*", NEVER bare "*"
  30&nbsp;   COMMAND ... "$file" ...
  31&nbsp; done
  32&nbsp;
  33&nbsp;
  34&nbsp;
  35&nbsp; #  These handle all filenames correctly;
  36&nbsp; #+ can be unwieldy if COMMAND is large:
  37&nbsp; find ... -exec COMMAND... {} \;
  38&nbsp; find ... -exec COMMAND... {} \+ # If multiple files are okay for COMMAND.
  39&nbsp;
  40&nbsp;
  41&nbsp;
  42&nbsp; #  This skips filenames with control characters
  43&nbsp; #+ (including tab and newline).
  44&nbsp; IFS="$(printf '\n\t')"
  45&nbsp; controlchars="$(printf '*[\001-\037\177]*')"
  46&nbsp; for file in $(find . ! -name "$controlchars"') ; do
  47&nbsp;   COMMAND "$file" ...
  48&nbsp; done
  49&nbsp;
  50&nbsp;
  51&nbsp;
  52&nbsp; #  Okay if filenames can't contain tabs or newlines --
  53&nbsp; #+ beware the assumption.
  54&nbsp; IFS="$(printf '\n\t')"
  55&nbsp; for file in $(find .) ; do
  56&nbsp;   COMMAND "$file" ...
  57&nbsp; done
  58&nbsp;
  59&nbsp;
  60&nbsp;
  61&nbsp; # Requires nonstandard but common extensions in find and xargs:
  62&nbsp; find . -print0 | xargs -0 COMMAND
  63&nbsp;
  64&nbsp; # Requires nonstandard extensions to find and to shell (bash works).
  65&nbsp; # variables might not stay set once the loop ends:
  66&nbsp; find . -print0 | while IFS="" read -r -d "" file ; do ...
  67&nbsp;   COMMAND "$file" # Use quoted "$file", not $file, everywhere.
  68&nbsp; done
  69&nbsp;
  70&nbsp;
  71&nbsp;
  72&nbsp; #  Requires nonstandard extensions to find and to shell (bash works).
  73&nbsp; #  Underlying system must include named pipes (FIFOs)
  74&nbsp; #+ or the /dev/fd mechanism.
  75&nbsp; #  In this version, variables *do* stay set after the loop ends,
  76&nbsp; #  and you can read from stdin.
  77&nbsp; #+ (Change the 4 to another number if fd 4 is needed.)
  78&nbsp;
  79&nbsp; while IFS="" read -r -d "" file &#60;&#38;4 ; do
  80&nbsp;   COMMAND "$file"   # Use quoted "$file" -- not $file, everywhere.
  81&nbsp; done 4&#60; &#60;(find . -print0)
  82&nbsp;
  83&nbsp;
  84&nbsp; #  Named pipe version.
  85&nbsp; #  Requires nonstandard extensions to find and to shell's read (bash ok).
  86&nbsp; #  Underlying system must include named pipes (FIFOs).
  87&nbsp; #  Again, in this version, variables *do* stay set after the loop ends,
  88&nbsp; #  and you can read from stdin.
  89&nbsp; # (Change the 4 to something else if fd 4 needed).
  90&nbsp;
  91&nbsp; mkfifo mypipe
  92&nbsp;
  93&nbsp; find . -print0 &#62; mypipe &#38;
  94&nbsp; while IFS="" read -r -d "" file &#60;&#38;4 ; do
  95&nbsp;   COMMAND "$file" # Use quoted "$file", not $file, everywhere.
  96&nbsp; done 4&#60; mypipe</PRE
></TD
></TR
></TABLE
></P
></DIV
><DIV
CLASS="NAVFOOTER"
><HR
ALIGN="LEFT"
WIDTH="100%"><TABLE
SUMMARY="Footer navigation table"
WIDTH="100%"
BORDER="0"
CELLPADDING="0"
CELLSPACING="0"
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
><A
HREF="awk.html"
ACCESSKEY="P"
>Prev</A
></TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
><A
HREF="index.html"
ACCESSKEY="H"
>Home</A
></TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
><A
HREF="exitcodes.html"
ACCESSKEY="N"
>Next</A
></TD
></TR
><TR
><TD
WIDTH="33%"
ALIGN="left"
VALIGN="top"
>Awk</TD
><TD
WIDTH="34%"
ALIGN="center"
VALIGN="top"
>&nbsp;</TD
><TD
WIDTH="33%"
ALIGN="right"
VALIGN="top"
>Exit Codes With Special Meanings</TD
></TR
></TABLE
></DIV
></BODY
></HTML
>