UNIX for Dummies Questions & Answers

Yes, the 'd' option would delete the whole line that contains the pattern; in this case the line containing "PhoneBooth".

For any regular expression engine, certain characters have a different meaning than it's literal one. For e.g., a question mark '?' would mean match the previous character zero or one time. So, if you want to tell the regex engine that you want to match a literal question mark and not want to implement the special meaning, then you have to escape it. In simple terms, escape a character if you want to use it in the literal sense, which otherwise would have a special meaning.

When I downloaded your file- Compound Eye.qtz, I ran ->

Code:

file Compound\ Eye.qtz
Compound Eye.qtz: Apple binary property list

PlistBuddy operates on binary property lists (strange that it's missing on your computer). You have to convert the binary property list to xml format in order to use any of the other solutions offered.

Code:

/usr/bin/plutil -convert xml1 /path/to/file.qtz

#to convert back to binary format

/usr/bin/plutil -convert binary1 /path/to/file.qtz

balajesuri has already answered most of your questions, so i will try to shed some additional light on it:

As i have explained in detail in the mentioned thread: it just looks that way.

LOL!

Yes, more or less.

Whenever you construct a regexp you have to ask three questions:

1) does it match what i want to match?
As you want to delete lines containing "PhotoBooth", this is obviously the case. OK.

2) does it match any instance i want to match?
For instance, there might be lines which do not read "PhotoBooth", but "Photobooth" and which should also be deleted. The regexp above won't catch these, because regexps are case-sensitive. You would have to use another - slightly different - regexp to catch both versions.

3) does it match things i don't want to match?
For instance, you might have lines in your files which contain "PhotoBooth", but which shouldn't be deleted. In this case you would have to enlarge your regexp, maybe to match only instances of "PhotoBooth" inside "<string>"-tags.

Note, i do not know your files, so i don't know if these points apply to your project. What i said is just a how-to for constructing regexps in general.

Regarding the "d" command for deleting lines: the general form of a command in "sed" is one of the three versions:

Code:

address,address command
address command
command

"address" can be a regexp, a line number or the symbol "$", which means the last line.
"command" can be either a single command, like "s" (=substitute, change), "d" (=delete), etc. or a compound command, consisting of several singular commands: "{ cmd1; cmd2; cmd3; ...}"

The first form applies "command" to consecutive lines, starting with the first address and ending with the last:

Code:

10,15d
/foo/,/bar/ s/x/y/

The first will delete lines 10-15, the second will change "x" to "y" in a block of lines from a line containing "foo" to a line containing "bar".

The second form will apply the command to a certain line or all the lines for which the regexp matches. Your command is this pattern:

Code:

10d
/foo/ s/x/y/

The first example will delete line 10, the second will change "x" to "y" only in lines containing "foo".

The third variation, finally, will apply "command" to every line. This is what most times is used.

Code:

d
s/x/y/

The first would delete every line (not very helpful, LOL), the second would replace "x" by "y" in every line.

You do not have to protect it from the shell, but from the regexp-engine. Like the shell, regexp engines have characters too which have a special meaning to them, as bajalesuri has already explained. These characters are called "metacharacters" and all have this property: they do not stand for themselves (their "literal" meaning), but have some special function. Every time you want to use them literally, you have to escape them to tell the regexp engine it shouldn't treat them in their special capacity but as normal characters. You might want to read this little introduction about regular expressions.

I hope this helps.

bakunin