I'm trying to run this command called codemaker which takes a filename as input but then writes the output to stdout instead of back to the file, so I have to redirect stdout to that file. That works fine, but I want to do this for a whole bunch of files at once, so I came up with this (based on https://stackoverflow.com/a/845928/65387):
ctouch() {
xargs -t -i -0 sh -c 'codemaker "$1" > "$1"' -- {} <<<"${(ps:\0:)#}"
}
But I can't quite get the syntax right. It looks like it's treating everything as a single arg still:
❯ ctouch foo.h bar.cc
sh -c 'codemaker "$1" > "$1"' -- 'foo.h bar.cc'$'\n'
Whereas I just want to run 2 commands:
codemaker foo.h > foo.h
codemaker bar.cc > bar.cc
How do I make an alias/function for that?
(And no, I'm not sure about that <<<"${(ps:\0:)#}" bit either. Really hard to Google. I want the usual "$#" to expand with null separators to feed to xargs)
I don't see a compelling reason to use xargs in your case. You just create additional processes unnecessarily (one for xargs, plus for each argument, one shell process).
A simpler solution (and IMO easier to understand) would be to do it with this zsh-function:
ctouch() {
for f
do
codemaker $f >$f
done
}
I think this is a lot easier to just do with printf.
ctouch() {
printf -- '%s\0' "$#" | xargs -t -i -0 sh -c 'codemaker "$1" > "$1"' -- {}
}
Related
Completely noob question but, using ls piped to grep, I need to find files or directories that have all capitals in their name, and directories need to have "/" appended to indicate that it is a directory. Trying to append the "/" is the only part I am stuck on. Again, I apologize for the amateur question. I currently have ls | grep [A-Z] and the example out should be: BIRD, DOG, DOGDIR/
It's an interesting question because it's a somewhat difficult thing to accomplish with a bash one-liner.
Here's what I came up with. It doesn't seem very elegant, but I'm not sure how to improve.
find /animals -type d -or -type f \
| grep '/[A-Z]*$' \
| xargs -I + bash -c 'echo -n $(basename +)$( test -d + && echo -n /),\\ ' \
| sed -e 's/, *$//'; echo
I'll break that down for you
find /animals -type d -or -type f writes out, once per line, the directories and files it found in /animals (see below for my test environment dockerfile - I created /animals to match your desired output). Find can't do a regex match as far as I know on the name, so...
grep '/[A-Z]*$' filter's find's output so that only paths are shown where the last part of the file or directory name, after the final /, is all uppercase
xargs -I + bash -c '...' when you're in a shell and you want to use a "for" loop, chances are what you should be using is xargs. Learn it, know it, love it. xargs takes its input, separated by default by $IFS, and runs the command you give it for each piece of input . So this is going to run a bash shell for each path. that passed the grep filter. In my case, -I + will make xargs replace the literal '+' character with its current input filename. -I also makes it pass one at a time through xargs. For more information, see the xargs manual page.
'echo -n $(basename +)$( test -d + && echo -n /),\\ ' this is the inner bash script that will be run by xargs for each path that got through grep.
basename + cuts the directory component off the path; from your example output you don't want eg /animals/DOGDIR/, you want DOGDIR/. basename is the program that trims the directories for us.
test -d + && echo -n / checks to see whether + (remember xargs will replace it with filename) is a directory ,and if so, runs echo -n /. the -n argument to echo suppresses the newline, important to get the output in the CSV format you specified.
now we can put it all together to see that we're echo -n the output of basename + , with / appended, if it's a directory, and then , appended to that. All the echos run with -n to suppress newlines to keep output CSV looking.
| sed -e 's/, *$//'; echo is purely for formatting. Adding , to each individual output was an easy way to get the CSV, but it leaves us with a final , at the end of the list. The sed invocation removes , followed by any number of spaces at the end of the output so far - eg the entire output from all the xargs invocations. And since we never did output a newline at the end of that output, the final echo is adding that.
Usually in unix shells, you probably wouldn't want a CSV style output. You'd probably instead want a newline-separated output in most cases, one matching file per line, and that would be somewhat simpler to do because you wouldn't need all that faffing with -n and , to make it CSV style. But, valid requirement if the need is there.
FROM debian
RUN mkdir -p /animals
WORKDIR /animals
RUN mkdir -p DOGDIR lowerdir && touch DOGDIR/DOG DOGDIR/lowerDOG2 lowerdir/BIRD
ENTRYPOINT [ "/bin/bash" ]
CMD [ "-c" , "find /animals -type d -or -type f | grep '/[A-Z]*$'| xargs -I + bash -c 'echo -n $(basename +)$( test -d + && echo -n /),\\ ' | sed -e 's/, *$//'; echo"]
$ docker run --rm test
BIRD, DOGDIR/, DOG
You can start looking at
ls -F | grep -v "[[:lower:]]"
I did not add something for a comma-seperated line, because this is the wrong method: Parsing ls should be avoided ! It will go wrong for filenames like
I am a terribble filename,
with newlines inside me,
and the ls command combined with grep
will only show the last line
BECAUSE THIS LINE HAS NO LOWERCASE CHARACTERS
To get the files without a pipe, you can use
shopt -s extglob
ls -dp +([[:upper:]])
shopt -u extglob
An explanation of the extglob and uppercase can be found at https://unix.stackexchange.com/a/389071/57293
When you want the output in one line, you can get troubles with filenames that have newlines or commas in its name. You might want something like
# parsing ls, yes wrong and failing for some files
ls -dp +([[:upper:]]) | tr "\n" "," | sed 's/,$/\n/'
What is the most elegant way in zsh to test, whether a file is either a readable regular file?
I understand that I can do something like
if [[ -r "$name" && -f "$name" ]]
...
But it requires repeating "$name" twice. I know that we can't combine conditions (-rf $name), but maybe some other feature in zsh could be used?
By the way, I considered also something like
if ls ${name}(R.) >/dev/null 2>&1
...
But in this case, the shell would complain "no matches found", when $name does not fulfil the criterium. Setting NULL_GLOB wouldn't help here either, because it would just replace the pattern with an empty string, and the expression would always be true.
In very new versions of zsh (works for 5.0.7, but not 5.0.5) you could do this
setopt EXTENDED_GLOB
if [[ -n $name(#qNR.) ]]
...
$name(#qNR.) matches files with name $name that are readable (R) and regular (.). N enables NULL_GLOB for this match. That is, if no files match the pattern it does not produce an error but is removed from the argument list. -n checks if the match is in fact non-empty. EXTENDED_GLOB is needed to enable the (#q...) type of extended globbing which in turn is needed because parenthesis usually have a different meaning inside conditional expressions ([[ ... ]]).
Still, while it is indeed possible to write something up that uses $name only once, I would advice against it. It is rather more convoluted than the original solution and thus harder to understand (i.e. needs thinking) for the next guy that reads it (your future self counts as "next guy" after at most half a year). And at least this solution will work only on zsh and there only on new versions, while the original would run unaltered on bash.
How about make small(?) shell functions as you mentioned?
tests-raw () {
setopt localoptions no_ksharrays
local then="$1"; shift
local f="${#[-1]}" t=
local -i ret=0
set -- "${#[1,-2]}"
for t in ${#[#]}; do
if test "$t" "$f"; then
ret=$?
"$then"
else
return $?
fi
done
return ret
}
and () tests-raw continue "${#[#]}";
or () tests-raw break "${#[#]}";
# examples
name=/dev/null
if and -r -c "$name"; then
echo 'Ok, it is a readable+character special file.'
fi
#>> Ok, it is...
and -r -f ~/.zshrc ; echo $? #>> 0
or -r -d ~/.zshrc ; echo $? #>> 0
and -r -d ~/.zshrc ; echo $? #>> 1
# It could be `and -rd ~/.zshrc` possible.
I feel this is somewhat overkill though.
Unfortunately, due to the limitations of our Unix Tru64 environment, I am unable to use the GREP -r switch to perform my search for strings within files across multiple directories and sub directories.
Ideally, I would like to pass two parameters. The first will be the directory I want my search is to start on. The second is a file containing a list of all the strings to be searched. This list will consist of various directory path names and will include special characters:
ie:
/aaa/bbb/ccc
/eee/dddd/ggggggg/
etc..
The purpose of this exercise is to identify all shell scripts that may have specific hard coded path names identified in my list.
There was one example I found during my investigations that perhaps comes close, but I am not sure how to customize this to accept a file of string arguments:
eg: find etb -exec grep test {} \;
where 'etb' is the directory and 'test', a hard coded string to be searched.
This should do it:
find dir -type f -exec grep -F -f strings.txt {} \;
dir is the directory from which searching will commence
strings.txt is the file of strings to match, one per line
-F means treat search strings as literal rather than regular expressions
-f strings.txt means use the strings in strings.txt for matching
You can add -l to the grep switches if you just want filenames that match.
Footnote:
Some people prefer a solution involving xargs, e.g.
find dir -type f -print0 | xargs -0 grep -F -f strings.txt
which is perhaps a little more robust/efficient in some cases.
By reading, I assume we can not use the gnu coreutil, and egrep is not available.
I assume (for some reason) the system is broken, and escapes do not work as expected.
Under normal situations, grep -rf patternfile.txt /some/dir/ is the way to go.
a file containing a list of all the strings to be searched
Assumptions : gnu coreutil not available. grep -r does not work. handling of special character is broken.
Now, you have working awk ? no ?. It makes life so much easier. But lets be on the safe side.
Assume : working sed ,one of od OR hexdump OR xxd (from vim package) is available.
Lets call this patternfile.txt
1. Convert list into a regexp that grep likes
Example patternfile.txt contains
/foo/
/bar/doe/
/root/
(example does not print special char, but it's there.) we must turn it into something like
(/foo/|/bar/doe/|/root/)
Assuming echo -en command is not broken, and xxd , or od, or hexdump is available,
Using hexdump
cat patternfile.txt |hexdump -ve '1/1 "%02x \n"' |tr -d '\n'
Using od
cat patternfile.txt |od -A none -t x1|tr -d '\n'
and pipe it into (common for both hexdump and od)
|sed 's:[ ]*0a[ ]*$::g'|sed 's: 0a:\\|:g' |sed 's:^[ ]*::g'|sed 's:^: :g' |sed 's: :\\x:g'
then pipe result into
|sed 's:^:\\(:g' |sed 's:$:\\):g'
and you have a regexp pattern that is escaped.
2. Feed the escaped pattern into broken regexp
Assuming the bare minimum shell escape is available,
we use grep "$(echo -en "ESCAPED_PATTERN" )" to do our job.
3. To sum it up
Building a escaped regexp pattern (using hexdump as example )
grep "$(echo -en "$( cat patternfile.txt |hexdump -ve '1/1 "%02x \n"' |tr -d '\n' |sed 's:[ ]*0a[ ]*$::g'|sed 's: 0a:\\|:g' |sed 's:^[ ]*::g'|sed 's:^: :g' |sed 's: :\\x:g'|sed 's:^:\\(:g' |sed 's:$:\\):g')")"
will escape all characters and enclose it with (|) brackets so a regexp OR match will be performed.
4. Recrusive directory lookup
Under normal situations, even when grep -r is broken, find /dir/ -exec grep {} \; should work.
Some may prefer xargs instaed (unless you happen to have buggy xargs).
We prefer find /somedir/ -type f -print0 |xargs -0 grep -f 'patternfile.txt' approach, but since
this is not available (for whatever valid reason),
we need to exec grep for each file,and this is normaly the wrong way.
But lets do it.
Assume : find -type f works.
Assume : xargs is broken OR not available.
First, if you have a buggy pipe, it might not handle large number of files.
So we avoid xargs in such systems (i know, i know, just lets pretend it is broken ).
find /whatever/dir/to/start/looking/ -type f > list-of-all-file-to-search-for.txt
IF your shell handles large size lists nicely,
for file in cat list-of-all-file-to-search-for.txt ; do grep REGEXP_PATTERN "$file" ;
done ; is a nice way to get by. Unfortunetly, some systems do not like that,
and in that case, you may require
cat list-of-all-file-to-search-for.txt | split --help -a 4 -d -l 2000 file-smaller-chunk.part.
to turn it into smaller chunks. Now this is for a seriously broken system.
then a for file in file-smaller-chunk.part.* ; do for single_line in cat "$file" ; do grep REGEXP_PATTERN "$single_line" ; done ; done ;
should work.
A
cat filelist.txt |while read file ; do grep REGEXP_PATTERN $file ; done ;
may be used as workaround on some systems.
What if my shell doe not handle quotes ?
You may have to escape the file list beforehand.
It can be done much nicer in awk, perl, whatever, but since we restrict our selves to
sed, lets do it.
We assume 0x27, the ' code will actually work.
cat list-of-all-file-to-search-for.txt |sed 's#['\'']#'\''\\'\'\''#g'|sed 's:^:'\'':g'|sed 's:$:'\'':g'
The only time I had to use this was when feeding output into bash again.
What if my shell does not handle that ?
xargs fails , grep -r fails , shell's for loop fails.
Do we have other things ? YES.
Escape all input suitable for your shell, and make a script.
But you know what, I got board, and writing automated scripts for csh just seems
wrong. So I am going to stop here.
Take home note
Use the tool for the right job. Writing a interpreter on bc is perfectly
capable, but it is just plain wrong. Install coreutils, perl, a better grep
what ever. makes life a better thing.
When there is no files inside the folder the below script goes inside the for loop. Not sure what i can modify so that it doesn't go inside the for loop. Also when there is no files inside the directory exit status should be success. Wrapper script checks the exit status of the below script
FILESRAW ="/exp/test1/folder" .
for fspec in "$FILESRAW"/* ; do
echo "$fspec"
if [[ -f ${fspec} ]] ; then
..... processing logic
else
... processing logic
fi
done
if using bash,
you can set nullglob
shopt-s nullglob
if you have hidden files,
shopt -s dotglob
with ksh,
#!/bin/ksh
set -o noglob
for file in /path/*
do
....
done
for fspec in `dir $FILESRAW` ; do
To exit if $FILESRAW is empty:
[ $( ls "$FILESRAW" | wc -l ) -eq 0 ] && exit 0
If this test precedes the loop, it will prevent execution from reaching the for loop if $FILESRAW is empty.
When $FILESRAW is empty, "$FILESRAW"/* expands to "/exp/test1/folder/*", as ghostdog74 points out, you can change this behavior by setting nullglob with
shopt -s nullglob
If you want hidden files, set dotglob as well:
shopt -s dotglob
Alternately, you could use ls instead of globing. This has the advantage of working with very full directories (using a pipe, you won't reach the maximum argument limit):
ls "$FILESRAW" | while read file; do
echo "$file"
This becomes messier if you want hidden files, since you'll need to exclude . and .. to emulate globing behavior:
ls -a "$FILESRAW" | egrep -v '^(\.|\.\.)$' | while read file; do
echo "$file"
if you are using ksh,
try putting this in front of for loop so that it won't go inside it.
"set -noglob"
Even I have got the same problem, but I was able to resolve it by doing this.
I'm trying to write a (sh -bourne shell) script that processes lines as they are written to a file. I'm attempting to do this by feeding the output of tail -f into a while read loop. This tactic seems to be proper based on my research in Google as well as this question dealing with a similar issue, but using bash.
From what I've read, it seems that I should be able to break out of the loop when the file being followed ceases to exist. It doesn't. In fact, it seems the only way I can break out of this is to kill the process in another session. tail does seem to be working fine otherwise as testing with this:
touch file
tail -f file | while read line
do
echo $line
done
Data I append to file in another session appears just file from the loop processing written above.
This is on HP-UX version B.11.23.
Thanks for any help/insight you can provide!
If you want to break out, when your file does not exist any more, just do it:
test -f file || break
Placing this in your loop, should break out.
The remaining problem is, how to break the read line, as this is blocking.
This could you do by applying a timeout, like read -t 5 line. Then every 5 second the read returns, and in case the file does not longer exist, the loop will break. Attention: Create your loop that it can handle the case, that the read times out, but the file is still present.
EDIT: Seems that with timeout read returns false, so you could combine the test with the timeout, the result would be:
tail -f test.file | while read -t 3 line || test -f test.file; do
some stuff with $line
done
I don't know about HP-UX tail but GNU tail has the --follow=name option which will follow the file by name (by re-opening the file every few seconds instead of reading from the same file descriptor which will not detect if the file is unlinked) and will exit when the filename used to open the file is unlinked:
tail --follow=name test.txt
Unless you're using GNU tail, there is no way it'll terminate of its own accord when following a file. The -f option is really only meant for interactive monitoring--indeed, I have a book that says that -f "is unlikely to be of use in shell scripts".
But for a solution to the problem, I'm not wholly sure this isn't an over-engineered way to do it, but I figured you could send the tail to a FIFO, then have a function or script that checked the file for existence and killed off the tail if it'd been unlinked.
#!/bin/sh
sentinel ()
{
while true
do
if [ ! -e $1 ]
then
kill $2
rm /tmp/$1
break
fi
done
}
touch $1
mkfifo /tmp/$1
tail -f $1 >/tmp/$1 &
sentinel $1 $! &
cat /tmp/$1 | while read line
do
echo $line
done
Did some naïve testing, and it seems to work okay, and not leave any garbage lying around.
I've never been happy with this answer but I have not found an alternative either:
kill $(ps -o pid,cmd --no-headers --ppid $$ | grep tail | awk '{print $1}')
Get all processes that are children of the current process, look for the tail, print out the first column (tail's pid), and kill it. Sin-freaking-ugly indeed, such is life.
The following approach backgrounds the tail -f file command, echos its process id plus a custom string prefix (here tailpid: ) to the while loop where the line with the custom string prefix triggers another (backgrounded) while loop that every 5 seconds checks if file is still existing. If not, tail -f file gets killed and the subshell containing the backgrounded while loop exits.
# cf. "The Heirloom Bourne Shell",
# http://heirloom.sourceforge.net/sh.html,
# http://sourceforge.net/projects/heirloom/files/heirloom-sh/ and
# http://freecode.com/projects/bournesh
/usr/local/bin/bournesh -c '
touch file
(tail -f file & echo "tailpid: ${!}" ) | while IFS="" read -r line
do
case "$line" in
tailpid:*) while sleep 5; do
#echo hello;
if [ ! -f file ]; then
IFS=" "; set -- ${line}
kill -HUP "$2"
exit
fi
done &
continue ;;
esac
echo "$line"
done
echo exiting ...
'