Converting a non-escaped regexp to an escaped version

Converting a non-escaped regexp to an escaped version

[HippoMan]

During development of elisp code that I'm writing, I test regular
expressions via the interactive use of M-x re-search-forward.  Once
I've come up with a regexp that works, I often will want
to yank it into my elisp code, for example, into a call to
`match-string'.

Of course, once I do this, I then have to go through the regexp and
double all of the backslashes within it.  In order to make this easier
for myself, I'd like to have an elisp function that will yank a
string while auto-doubling every backslash it contains.  I'm happy to
write such a function, but before I re-invent the wheel, I'm
wondering if such a thing has already been invented.

In case this isn't clear, here's what I want to do:

1. Test a regular expression interactively, for example, via
   the use of M-x re-search-forward.

2. Assume that I finally decide that the following regexp is
   the one that I want to use within my elisp code:

     \bfoo\b.*:.*\bbar\([0-9]\{9\}quack\.quack\)\b

3. Put this regexp to the kill ring.

4. Use my proposed function to yank it into my elisp code
   (for example, within an invocation of `match-string'), as follows:

     \\bfoo\\b.*:.*\\bbar\\([0-9]\\{9\\}quack\\.quack\\)\\b

Note that I'm not asking any of you to write this function for
me, as I can do this.  I just want to know whether something
like this already exists within the emacs code base or perhaps
somewhere else on the net.

Thanks in advance.

--
 hippoman@gmail.com

Re: Converting a non-escaped regexp to an escaped version

[Nikolaj Schumacher]

On 2007-06-07, at 14:32, HippoMan wrote:

> During development of elisp code that I'm writing, I test regular
> expressions via the interactive use of M-x re-search-forward.  Once
> I've come up with a regexp that works, I often will want
> to yank it into my elisp code, for example, into a call to
> `match-string'.

Not really a solution for your specific problem, but I'd suggest  
using `re-builder' for that.

regards,
Nikolaj Schumacher

Re: Converting a non-escaped regexp to an escaped version

[Juanma Barranquero]

On 6/7/07, HippoMan <hippoman@gmail.com> wrote:

> 1. Test a regular expression interactively, for example, via
>    the use of M-x re-search-forward.
>
> 2. Assume that I finally decide that the following regexp is
>    the one that I want to use within my elisp code:
>
>      \bfoo\b.*:.*\bbar\([0-9]\{9\}quack\.quack\)\b
>
> 3. Put this regexp to the kill ring.
>
> 4. Use my proposed function to yank it into my elisp code
>    (for example, within an invocation of `match-string'), as follows:
>
>      \\bfoo\\b.*:.*\\bbar\\([0-9]\\{9\\}quack\\.quack\\)\\b

You could use M-x re-builder to test your regexp, which is very nice,
supports several regexp syntaxes and can highlight subexpressions
(groups).

Then, inside the re-builder buffer, do C-c C-w (that's `reb-copy') to
copy the regexp.

             Juanma

Re: Converting a non-escaped regexp to an escaped version

[Joost Kremers]

Nikolaj Schumacher wrote:
> On 2007-06-07, at 14:32, HippoMan wrote:
>
>> During development of elisp code that I'm writing, I test regular
>> expressions via the interactive use of M-x re-search-forward.  Once
>> I've come up with a regexp that works, I often will want
>> to yank it into my elisp code, for example, into a call to
>> `match-string'.
>
> Not really a solution for your specific problem, but I'd suggest  
> using `re-builder' for that.

it can be a solution, though: re-builder supports different forms of input
syntax. by default, it requires you to escape backslashes, so that when you
type C-c C-w to copy the regexp to the kill-ring, you get it with every
backslash escaped.

if you don't want to type the double backslashes, you can change
re-builders input syntax (with C-c TAB)from "read" to "string". then,
before typing C-c C-w change it back again to "read": re-builder will
automatically adjust the syntax and escape all backslashes for you.

it's a pretty neat package, that re-builder.

-- 
Joost Kremers                                      joostkremers@yahoo.com
Selbst in die Unterwelt dringt durch Spalten Licht
EN:SiS(9)

Re: Converting a non-escaped regexp to an escaped version

[HippoMan]

On Jun 7, 10:35 am, Joost Kremers <joostkrem...@yahoo.com> wrote:
> Nikolaj Schumacher wrote:
> > On 2007-06-07, at 14:32, HippoMan wrote:
>
> >> During development of elisp code that I'm writing, I test regular
> >> expressions via the interactive use of M-x re-search-forward.  Once
> >> I've come up with a regexp that works, I often will want
> >> to yank it into my elisp code, for example, into a call to
> >> `match-string'.
>
> > Not really a solution for your specific problem, but I'd suggest
> > using `re-builder' for that.
>
> it can be a solution, though: re-builder supports different forms of input
> syntax. by default, it requires you to escape backslashes, so that when you
> type C-c C-w to copy the regexp to the kill-ring, you get it with every
> backslash escaped.
>
> if you don't want to type the double backslashes, you can change
> re-builders input syntax (with C-c TAB)from "read" to "string". then,
> before typing C-c C-w change it back again to "read": re-builder will
> automatically adjust the syntax and escape all backslashes for you.
>
> it's a pretty neat package, that re-builder.

Thank to all of you who suggested re-builder.  I can live with the
changing
of its input syntax from "read" to "string".  Hey, I might write my
own function(s)
for automating even this!

Re: Converting a non-escaped regexp to an escaped version

[Joost Kremers]

HippoMan wrote:
> Thank to all of you who suggested re-builder.  I can live with the
> changing of its input syntax from "read" to "string".  Hey, I might write
> my own function(s) for automating even this!

it may be possible to wrap reb-copy (the function called by C-c C-w in
re-builder) in a defadvice, so that it automatically changes the input
syntax for you before putting the regexp in the kill ring.


-- 
Joost Kremers                                      joostkremers@yahoo.com
Selbst in die Unterwelt dringt durch Spalten Licht
EN:SiS(9)

Re: Converting a non-escaped regexp to an escaped version

[martin rudalics]

You could also have a look at a thing I wrote

http://lists.gnu.org/archive/html/gnu-emacs-sources/2005-11/msg00004.html

It has two commands which permit to find the next and previous match
for the regexp you are about to write in the same Elisp buffer.

Re: Converting a non-escaped regexp to an escaped version

[Johan Bockgård]

HippoMan <hippoman@gmail.com> writes:

> 1. Test a regular expression interactively, for example, via
>    the use of M-x re-search-forward.
>
> 2. Assume that I finally decide that the following regexp is
>    the one that I want to use within my elisp code:
>
>      \bfoo\b.*:.*\bbar\([0-9]\{9\}quack\.quack\)\b

3. C-x ESC ESC

-- 
Johan Bockgård