UNIX for Dummies Questions & Answers

I make this script that uses "strace" to trace programs. The script looks for files that the program opens. It looks like this

Code:

open("/lib/libtermcap.so.2", O_RDONLY)  = 3
open("/dev/tty", O_RDWR|O_NONBLOCK|O_LARGEFILE) = 3

I want to extract the path without the quotes, preferably using grep and cut, but it can be also sed or awk (I just don't know these much). I worked out a regular expression egrep -o "\(\"[^\"]*\"," | cut -f 2 -d \" but then I realized the path can contain characters like " as well.

I thought the solution could look something like this egrep -o "\(\"[^{[^\\]\",[:space:]}]*\"," (any characters except string "[^\\]\",[:space:]" and these characters repeating)... Because I found out that strace doesn't work correctly if "\" without quotes is at the end of the file. So I thought I could use it as a benefit. However, it doesn't work .. I'm not very good in regular expressions indeed.

EDIT: And I'm sorry ... [:space:] is not the solution, because the script should work in FreeBSD as well and it uses "truss" instead of "strace" and there's no space after ",

Code:

sed '/^open/s/.[^"]*\"//;s/\".*//'

Unfortunately doesn't work for input like this:

Code:

open("/homes/eva/xs/username/script\",\"", O_RDONLY|O_LARGEFILE) = 3

Guys, really don't know anybody? Just how to extract the path from the first " up to the last sequence ", I can't figure it out.

Here's another example of what this script needs to be capable extract (in this case /homes/xx):

Code:

openat(AT_FDCWD, "/homes/xx", O_WRONLY)

Code:

$ echo 'openat(AT_FDCWD, "/homes/xx", O_WRONLY)' | sed 's/[^"][^"]*"\([^"][^"]*\).*/\1/'
/homes/xx