The wording on the various options could use some explanation. I don't know what style regexes are permitted, for example. And I don't what all those "default spaces" options do.

As an example, how do I search for all nodes written by me that include "on" as a word? I thought \bon\b would work, but I tried both (indecipherable) settings below the box, and neither seemed to do it. The placement of the radiobuttons and the word "-or-" is particularly troublesome:

(separate {button} strings -or- regexes {button} with {blank} -- default is spaces)
Does that mean the buttons choose between strings, or perhaps regexes with spaces (or something else?), or does the button choose between strings or regexes, and then I can pick something (what?) in place of spaces?

Even just a 30-second blurb of "what the hell was I thinking when I wrote that" from the author would be useful at this point.

-- Randal L. Schwartz, Perl hacker

Replies are listed 'Best First'.
(tye)Re: new super search
by tye (Sage) on Oct 08, 2002 at 19:49 UTC

    Enter "super search" after the "Match titles containing", enter "tye" after the "Match -or- Exclude authors", and then press the "Search" button.

    You should see quite a bit of information on (as you so tactfully put it) '"what the hell was I thinking when I wrote that" from the author'. In particular, read (tye)Re4: Newest Super Search.

    Update: I forgot to mention that regular expressions can only be used by saints (who can only use them on titles, not on node text). They are still a bit of a CPU drain but the idea is that saints will be responsible and not abuse the privilege. If it becomes an undue load, the privilege will be taken away.

    MySQL regexes don't support "\b". I translate several Perl regex features into MySQL regex constructs so you can use things like \d, \D, \s, etc. However, you hit on a weird one. You'll want to search on "\bon\B" since "\b" gets translated into MySQL's "start of word" pattern and "\B" gets translated into MySQL's "end of word" pattern.

    When it became clear that regex searching was too expensive, documenting how to do it became rather low on the priority list.

            - tye (the author in question)
[reply]
Re: new super search
by blakem (Monsignor) on Oct 08, 2002 at 21:14 UTC
    If you're content with defining the word on as the four characters space-oh-en-space, you can search for it using something like:
    ; on ;
    (semicolons included) Then set the default string separator to a semicolon. Thats how I've used it anyway.

    -Blake

[reply]
[d/l]
[reply]
        Seems to work for me... Do you see this text on the results page?
        where any text contains " on"
        written by any of merlyn
        I checked the first ten results and they all contained the word on. What results did you get that matched "question" or "one"?

        -Blake

[reply]
[d/l]
[select]
Re: new super search
by Aristotle (Chancellor) on Oct 09, 2002 at 21:22 UTC

    It's quite simple really. The text input defines a separator for parts of the search term. It's like the first parameter to split. If you say your separator is an ; and enter " on ;foo" in the search box, it will search for the string " on " and the string "foo", where the latter may be embedded in other words like "stfoobar".

    With the radio buttons, you just choose whether the individual parts of the search term (after splitting) are treated as plaintext words or regexen.

    Makeshifts last the longest.

[reply]

Back to Perl Monks Discussion