Regex negative lookahead failed

Regex negative lookahead failed

[Erwan Hingant]

Hello,

  I have some troubles with regex negative lookahead. When searching 
in 
all directories but one (say inbox), I do the following query:

 > notmuch search folder:"/^(?!inbox)/"

and I receive the following error:

     notmuch search: A Xapian exception occurred
     A Xapian exception occurred parsing query: Regexp error: Invalid 
preceding regular expression
     Query string was: folder:"/^(?!inbox)/"

Bests,

Erwan.


\r

Re: Regex negative lookahead failed

[David Bremner]

Erwan Hingant <erwan.hingant@mailo.com> writes:

> Hello,
>
>   I have some troubles with regex negative lookahead. When searching 
> in 
> all directories but one (say inbox), I do the following query:
>
>  > notmuch search folder:"/^(?!inbox)/"
>

As far as I know, lookaheads are not supported by POSIX regex, which is
what notmuch uses. regex(7) is a bit terse, but the relevant section
seems to be

       An atom is a regular expression enclosed in "()" (matching a match  for
       the  regular  expression),  an  empty  set  of  "()" (matching the null
       string)(!), a bracket expression (see below), '.' (matching any  single
       character),  '^' (matching the null string at the beginning of a line),
       '$' (matching the null string at the end of a line), a '\' followed  by
       one  of the characters "^.[$()|*+?{\" (matching that character taken as
       an ordinary character),  a  '\'  followed  by  any  other  character(!)
       (matching  that character taken as an ordinary character, as if the '\'
       had not been present(!)), or a single character with no other  signifi‐
       cance  (matching  that character).  A '{' followed by a character other
       than a digit is an ordinary character, not the beginning of a bound(!).
       It is illegal to end an RE with '\'.

I'm not sure your whole problem, but maybe regex is not the right answer
here. In general they should be a last resort in notmuch, for efficiency
reasons. Does "not folder:inbox"  do what you want?

Re: Regex negative lookahead failed

[Michael J Gruber]

David Bremner venit, vidit, dixit 2021-04-24 12:57:17:
> Erwan Hingant <erwan.hingant@mailo.com> writes:
> 
> > Hello,
> >
> >   I have some troubles with regex negative lookahead. When searching 
> > in 
> > all directories but one (say inbox), I do the following query:
> >
> >  > notmuch search folder:"/^(?!inbox)/"
> >
> 
> As far as I know, lookaheads are not supported by POSIX regex, which is
> what notmuch uses. regex(7) is a bit terse, but the relevant section
> seems to be
> 
>        An atom is a regular expression enclosed in "()" (matching a match  for
>        the  regular  expression),  an  empty  set  of  "()" (matching the null
>        string)(!), a bracket expression (see below), '.' (matching any  single
>        character),  '^' (matching the null string at the beginning of a line),
>        '$' (matching the null string at the end of a line), a '\' followed  by
>        one  of the characters "^.[$()|*+?{\" (matching that character taken as
>        an ordinary character),  a  '\'  followed  by  any  other  character(!)
>        (matching  that character taken as an ordinary character, as if the '\'
>        had not been present(!)), or a single character with no other  signifi‐
>        cance  (matching  that character).  A '{' followed by a character other
>        than a digit is an ordinary character, not the beginning of a bound(!).
>        It is illegal to end an RE with '\'.
> 
> I'm not sure your whole problem, but maybe regex is not the right answer
> here. In general they should be a last resort in notmuch, for efficiency
> reasons. Does "not folder:inbox"  do what you want?

Yes, that solves the problem as stated (all folders except inbox).

As for the error message: both "^" and "(?!inbox)" are assertions in
regex lingo, not atoms. A negative assertion like "(?!inbox)" asserts
whether the previous atom is considered a match or not (under the
assumption that the atom matches something).

The error message seems to say that you cannot use an assertion
without a preceding atom that it applies to.

But, assertions are really for the case where you want to base a match
result (true/false) on parts of the expression that you do not want to
be part of the match itself. As far as I understand, matching in xapian
is solely about the true/false result, so I'm wondering when you would
really need lookahead and lookbehind (instead of groups).

Michael

Re: Regex negative lookahead failed

[David Bremner]

Michael J Gruber <git@grubix.eu> writes:

> But, assertions are really for the case where you want to base a match
> result (true/false) on parts of the expression that you do not want to
> be part of the match itself. As far as I understand, matching in xapian
> is solely about the true/false result, so I'm wondering when you would
> really need lookahead and lookbehind (instead of groups).

Pedantic correction, the regex support in notmuch is not provided by
Xapian, but something bolted on by notmuch. That explains some (but not
all!) of its quirks.

d

Re: Regex negative lookahead failed

[Erwan Hingant]

Hello,

 Thanks a lot, this makes all clear and notmuch-search-terms(7) point the
 regex POSIX version you mention. My mistake. By the way, you completely
 solve my problem that of course no need regular expression has I
 thought...

Bests,
Erwan.

David Bremner <david@tethera.net> writes:

> Erwan Hingant <erwan.hingant@mailo.com> writes:
>
>> Hello,
>>
>>   I have some troubles with regex negative lookahead. When searching 
>> in 
>> all directories but one (say inbox), I do the following query:
>>
>>  > notmuch search folder:"/^(?!inbox)/"
>>
>
> As far as I know, lookaheads are not supported by POSIX regex, which is
> what notmuch uses. regex(7) is a bit terse, but the relevant section
> seems to be
>
>        An atom is a regular expression enclosed in "()" (matching a match  for
>        the  regular  expression),  an  empty  set  of  "()" (matching the null
>        string)(!), a bracket expression (see below), '.' (matching any  single
>        character),  '^' (matching the null string at the beginning of a line),
>        '$' (matching the null string at the end of a line), a '\' followed  by
>        one  of the characters "^.[$()|*+?{\" (matching that character taken as
>        an ordinary character),  a  '\'  followed  by  any  other  character(!)
>        (matching  that character taken as an ordinary character, as if the '\'
>        had not been present(!)), or a single character with no other  signifi‐
>        cance  (matching  that character).  A '{' followed by a character other
>        than a digit is an ordinary character, not the beginning of a bound(!).
>        It is illegal to end an RE with '\'.
>
> I'm not sure your whole problem, but maybe regex is not the right answer
> here. In general they should be a last resort in notmuch, for efficiency
> reasons. Does "not folder:inbox"  do what you want?
\r