Homework & Coursework Questions

Use and complete the template provided. The entire template must be completed. If you don't, your post may be deleted!

1. The problem statement, all variables and given/known data:
"Create a shell script that changes the selected word to another word in all files from selected archive. The script has 3 parameters: archive, old-word, new-word."



2. Relevant commands, code, scripts, algorithms:
This is a begginer's level exercise, we were instructed to use simple solutions such as loops



3. The attempts at a solution (include all code and scripts):This was my attempt:

Code:

for filename in *
do
     sed "s/old-word/new-word/g" "$filename" >temp
     mv temp "$filename"
done

My TA told me this but I believe I could not understand how to do that: "you should make an error-proof script that does not overwrite any file. If I
would have a file called 'temp', you will overwrite it. The same will happen if
two people will run your script simultaneously.

It should work with commandline parameters also. In case of any error it
should report the error to standard error output.

Perform testing on filenames (inside the archive, archive name, directory
names) containing spaces."


4. Complete Name of School (University), City (State), Country, Name of Professor, and Course Number (Link to Course):
CVUT "České vysoké učení technické v Praze", Czech Republic, Praha, Miroslav Bursa, Operating systems and networks, we don't have a page for the class that i know of. [/b]


Note: Without school/professor/course information, you will be banned if you post here! You must complete the entire template (not just parts of it).

Moderator's Comments:

Please use code tags <- click the link!

You shouldn't do mv like that, creating a new file instead of overwriting the old runs the risk of changing the permissions and ownership of the file. You should overwrite the existing file with cat afterwards instead.

I think what he wants is this:

Code:

TMP=`mktemp`

for filename in *
do
        sed 's/old-word/new-word/g' "$filename" > "$TMP"
        cat "$TMP" > "$filename"        
done

rm -f "$TMP"

'mktemp' generates a random, unique filename inside /tmp/ for you to use. You will never get the same filename twice, making it safe to run several of your script at the same time.

If you don't have mktemp, do TMP=/tmp/$$ which is almost as good a unique name, tied to your process ID ($$), but less secure because someone outside your process will be able to guess which file you are using.

Under Linux, you can do this all in one step with

Code:

sed -i 's/old-word/new-word/g' *

...but like your first solution, this runs the risk of changing the ownerships of files. Sometimes people have run sed -i as root and later found that all their files are unwritable -- because they're all owned by root now!