Re: tsearch2 column update produces "word too long"error

Patch submitted to 7.5devel and REL7_4_STABLE

Markus Wollny wrote:
> Hi!
>
> Now I really couldn't code C to save my life, but I managed to elicit
> some more debugging info. It's still dumb-user-interaction as suspected,
> but this is an issue I have to take into account as a basis; here's the
> "patch" for ts_cfg.c:
>
> if (lenlemm >= MAXSTRLEN)
> ereport(ERROR,
> (errcode(ERRCODE_SYNTAX_ERROR),
> ! errmsg("word is too long(%d):
> %s",lenlemm,lemm)));
>
> Now when I try
>
> UPDATE ct_com_board_message
> SET ftindex=to_tsvector('default',coalesce(user_login,'') ||'
> '|| coalesce(title,'') ||' '|| coalesce(text,''));
>
> I eventually get:
>
> ERROR: word is too long(2724):
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajjajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajjajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajjajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajjajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajjajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajjajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajjajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajjajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajjajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajjajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaja
> jajajajajajajajajajajajajajajajajajajajajajajajajjajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
> ajajajajajajajajajajajajajajajajajajajajajajajajajajajajajaj
>
> This is a brightly shining example of utterly wanton user-stupidity, I
> think: A 2k+ string of |:ja:|. Input like that cannot be helped, though
> - if he'd been a bit more imaginative, he could have used a few dozen
> "Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch" in a row or
> anything else; unfortunately there's no app that could automatically
> whack a user if he's doing something stupid.
>
> But on the other hand I cannot think of any reason why crap like that
> should be indexed in the first place. Therefore I would like to see some
> sort of option allowing me to still use tsearch2 but actually
> automatically excluding anything exceeding MAXSTRLEN - so the UPDATE
> might throw a NOTICE (if anything at all) but still get on with the
> rest.
>
> An alteration like that does however exceed my limited abilities with C
> by far and I don't want to mess with something I do not fully understand
> and then use that mess in a production environment. Is there a way to
> get around this problem with oversized words?
>
> Kind regards
>
> Markus
>
>
>
>>-----Ursprüngliche Nachricht-----
>>Von: Oleg Bartunov [mailto:oleg(at)sai(dot)msu(dot)su]
>>Gesendet: Freitag, 21. November 2003 15:13
>>An: Markus Wollny
>>Cc: pgsql-general(at)postgresql(dot)org
>>Betreff: Re: AW: [GENERAL] tsearch2 column update produces "word too
>>long"error
>>
>>
>>On Fri, 21 Nov 2003, Markus Wollny wrote:
>>
>>
>>>Hello!
>>>
>>>
>>>>Von: Oleg Bartunov [mailto:oleg(at)sai(dot)msu(dot)su]
>>>>Gesendet: Freitag, 21. November 2003 13:06
>>>>An: Markus Wollny
>>>>Cc: pgsql-general(at)postgresql(dot)org
>>>>
>>>>Word length is limited by 2K. What's exactly the word
>>>>tsearch2 complained on ?
>>>>'Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch'
>>>>is fine :)
>>>
>>>This was a silly example, I know - it is a long word, but
>>
>>not too long
>>
>>>to worry a machine. The offending word will surely be much
>>
>>longer, but
>>
>>>as a matter of fact, I cannot think of any user actually
>>
>>typing a 2k+
>>
>>>string without any spaces in between. I'm not sure on which word
>>>tsearch2 complained, it doesn't tell and even logging did
>>
>>not provide me
>>
>>>with any more detail:
>>>
>>>2003-11-21 14:06:44 [26497] ERROR: 42601: word is too long
>>>LOCATION: parsetext_v2, ts_cfg.c:294
>>>STATEMENT: UPDATE ct_com_board_message
>>> SET
>>>ftindex=to_tsvector('default',coalesce(user_login,'') ||' '||
>>>coalesce(title,'') ||' '|| coalesce(text,''));
>>>
>>>Is there some way to find the exact position?
>>
>>I'm afraid you need to hack ts_cfg.c:294 yourself to print the word
>>which's bugging you :)
>>
>>
>>>>btw, don't forget to configure properly dictionaries, so you
>>>>don't have a lot of unique words.
>>>
>>>I won't forget that; I justed wanted to run a quick-off first test
>>>before diving deeper into Ispell and other issues which are
>>
>>as yet a bit
>>
>>>of a mystery to me.
>>>
>>>Kind Regards
>>>
>>> Markus
>>>
>>
>> Regards,
>> Oleg
>>_____________________________________________________________
>>Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
>>Sternberg Astronomical Institute, Moscow University (Russia)
>>Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
>>phone: +007(095)939-16-83, +007(095)939-23-83
>>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org

--
Teodor Sigaev E-mail: teodor(at)sigaev(dot)ru