preg functions: use single or double quotes?

[sinister]

The examples in the online manual all seem to use double quotes, e.g. at
http://us3.php.net/preg_replace

Why? (The behavior is different with single quotes, and presumably simpler
to understand.)

[Chung Leong]

When it comes to regular expressions, single quotes don't have a distinct
advantage, in terms of simpler syntax, over double quotes, since you have to
escape backslashes in both type of strings. Meanwhile you can do variable
interpolation in doubly quoted strings. So I think "" has the advantage.

Uzytkownik "sinister" <si******@nospam.invalid> napisal w wiadomosci
news:nk*******************@nwrddc01.gnilink.net...

The examples in the online manual all seem to use double quotes, e.g. at
http://us3.php.net/preg_replace

Why? (The behavior is different with single quotes, and presumably simpler to understand.)

[John Dunlop]

sinister wrote:

The examples in the online manual all seem to use double quotes, e.g. at
http://us3.php.net/preg_replace

Why?
Dunno. You might not be too wrong if you were to ascribe that to
inertia. Remember that examples are just examples -- illustrations of
particular methods, if you like -- and aren't intended as best-
practice guidelines. The Manual doesn't urge you to copy their
constructs character for character. It's sometimes necessary to
question authority. ;-)
(The behavior is different with single quotes, and presumably simpler
to understand.)

Indeed the behaviour is different, but I don't believe it's that much
simpler to understand, as long as you know how single- and double-
quoted strings are parsed differently.

http://www.php.net/manual/en/language.types.string.php

As regards to the question in your Subject line (the Subject line is
no substitute for the body of a message; at least yours is
informative though), which was "preg functions: use single or double
quotes?", that's up to you. If, for instance, you need variables to
be parsed in your pattern, then you can't use single-quotes.

There's a nice example given in the documentation for preg_replace on
the 1st Nov. 2003. It concerns how to go about matching the two-
character sequence "\n" (a backslash followed by a lowercase letter
"n"). Consider the string

$foo = '\n'

or,

$foo = "\\n"

The sequence "\n" is special in a regular expression, in that the two
characters stand for a linefeed character. Now, in a single-quoted
pattern, you must precede the "n" by three backslashes:

preg_match('`\\\n`',$foo)

The reason is that a backslash followed by another backslash results
in a single literal backslash in single-quoted strings. The Manual
isn't at all clear on this point IMO, although it does say as much,
in a roundabout kind of way. If we had just two backslashes, the
regular expression would only match a newline character, just as
would happen with a single backslash. (A single and double backslash
behave identically here because a single backslash followed by a
letter is not special in single-quoted strings, but a double
backslash is transformed into a single backslash.)

In a double-quoted pattern however, we need four backslashes before
the "n". This is because in double-quoted strings, a backslash
following a backslash results in a single literal backslash, just as
in single-quoted strings, but additionally, special "escape
sequences" are recognised within double-quoted strings. "\n" means a
linefeed character. To stop "\n" from having that meaning, we need to
escape the backslash with another backslash. Thus, we must use four
backslashes in total:

preg_match("`\\\\n`",$foo)

--
Jock

[sinister]

"Chung Leong" <ch***********@hotmail.com> wrote in message
news:Z6********************@comcast.com...

When it comes to regular expressions, single quotes don't have a distinct
advantage, in terms of simpler syntax, over double quotes, since you have to escape backslashes in both type of strings. Meanwhile you can do variable
interpolation in doubly quoted strings. So I think "" has the advantage.
I agree with what you say. By "simpler" I meant that fewer backslashes are
needed.

Uzytkownik "sinister" <si******@nospam.invalid> napisal w wiadomosci
news:nk*******************@nwrddc01.gnilink.net...

The examples in the online manual all seem to use double quotes, e.g. at
http://us3.php.net/preg_replace

Why? (The behavior is different with single quotes, and presumably

simpler

to understand.)

[sinister]

"John Dunlop" <jo*********@johndunlop.info> wrote in message
news:MP************************@news.freeserve.net ...

sinister wrote:

The examples in the online manual all seem to use double quotes, e.g. at
http://us3.php.net/preg_replace

Why?
Dunno. You might not be too wrong if you were to ascribe that to
inertia. Remember that examples are just examples -- illustrations of
particular methods, if you like -- and aren't intended as best-
practice guidelines. The Manual doesn't urge you to copy their
constructs character for character. It's sometimes necessary to
question authority. ;-)

That's an interesting aside. I understand that reference material cannot
dwell exclusively on best practices, but IMHO they ought to be emphasized.

(The behavior is different with single quotes, and presumably simpler
to understand.)
Indeed the behaviour is different, but I don't believe it's that much
simpler to understand, as long as you know how single- and double-
quoted strings are parsed differently.

http://www.php.net/manual/en/language.types.string.php

As regards to the question in your Subject line (the Subject line is
no substitute for the body of a message; at least yours is
informative though), which was "preg functions: use single or double
quotes?", that's up to you. If, for instance, you need variables to
be parsed in your pattern, then you can't use single-quotes.

Right. In my example, I didn't need variable interpolation.
There's a nice example given in the documentation for preg_replace on
the 1st Nov. 2003. It concerns how to go about matching the two-
character sequence "\n" (a backslash followed by a lowercase letter
"n"). Consider the string

$foo = '\n'

or,

$foo = "\\n"

The sequence "\n" is special in a regular expression, in that the two
characters stand for a linefeed character. Now, in a single-quoted
pattern, you must precede the "n" by three backslashes:

preg_match('`\\\n`',$foo)

The reason is that a backslash followed by another backslash results
in a single literal backslash in single-quoted strings. The Manual
isn't at all clear on this point IMO, although it does say as much,
In the section on preg_replace, or in another section on quoted strings?
in a roundabout kind of way. If we had just two backslashes, the
regular expression would only match a newline character, just as
would happen with a single backslash. (A single and double backslash
behave identically here because a single backslash followed by a
letter is not special in single-quoted strings, but a double
backslash is transformed into a single backslash.)

In a double-quoted pattern however, we need four backslashes before
the "n". This is because in double-quoted strings, a backslash
following a backslash results in a single literal backslash, just as
in single-quoted strings, but additionally, special "escape
sequences" are recognised within double-quoted strings. "\n" means a
linefeed character. To stop "\n" from having that meaning, we need to
escape the backslash with another backslash. Thus, we must use four
backslashes in total:

preg_match("`\\\\n`",$foo)
That's what I meant by more complicated.

Thanks for the detailed explanation!

Best,

S

--
Jock

[John Dunlop]

sinister wrote:

"John Dunlop" <jo*********@johndunlop.info> wrote in message
news:MP************************@news.freeserve.net ...

Dunno. You might not be too wrong if you were to ascribe that to
inertia. Remember that examples are just examples -- illustrations of
particular methods, if you like -- and aren't intended as best-
practice guidelines. The Manual doesn't urge you to copy their
constructs character for character. It's sometimes necessary to
question authority. ;-)

That's an interesting aside. I understand that reference material cannot
dwell exclusively on best practices, but IMHO they ought to be emphasized.

I agree. I honestly don't know why they've used double-quoted strings
when they're not making use of the benefits that come with them.

The reason is that a backslash followed by another backslash results
in a single literal backslash in single-quoted strings. The Manual
isn't at all clear on this point IMO, although it does say as much,

In the section on preg_replace, or in another section on quoted strings?

In the section on single-quoted strings it says:

| To specify a literal single quote, you will need to escape it with
| a backslash (\), like in many other languages. If a backslash needs
| to occur before a single quote or at the end of the string, you
| need to double it. Note that if you try to escape any other
| character, the backslash will also be printed! So usually there is
| no need to escape the backslash itself.

http://www.php.net/manual/en/language.types.string.php

It says "Note that if you try to escape any other character, the
backslash will also be printed!". The prior sentence stated in what
position a backslash *must* be escaped (with another backslash). It
doesn't explicitly say that if you escape a backslash anywhere in a
single-quoted string, only a single backslash is left after parsing.
But that is what it's saying, implicitly. I'm fairly sure it could be
worded better. Maybe that's just me though.

--
Jock