Re: 2nd update on TOAST
От | Philip Warner |
---|---|
Тема | Re: 2nd update on TOAST |
Дата | |
Msg-id | 3.0.5.32.20000706131555.01e028e0@mail.rhyme.com.au обсуждение исходный текст |
Ответ на | Re: 2nd update on TOAST (Philip Warner <pjw@rhyme.com.au>) |
Ответы |
Re: 2nd update on TOAST
|
Список | pgsql-hackers |
At 12:50 6/07/00 +1000, Philip Warner wrote: >> >>Are you using current sources? Bruce committed Philip Warner's >>pg_dump rewrite a day or so ago (which I thought was way premature, >>but anyway...). Just want to know which pg_dump we're talking about. > >It's good that any bugs had a chnace to come out ASAP. > >However, I'd be interested to know what actually happens: I do recall that >pg_dump (both versions) have seomething like '#define COPY_BUFFER_SIZE >8192'; if I were to start looking, that's where I'd go first. > Further to this, looking at the code, it now uses 'archputs', which replaces 'fputs', to output the copy buffer (assuming the error is in the copy, not while dumping the definitions); if PQgetline fills the entire buffer (no trailing \0), then I could imagine both would croak. The simplest test would be to tell PQgetline that the buffer size if 1 byte smaller, and see if it fixes the problem. I'd be interested to hear from someone...at least to know a little more of the circumstaces of the error. ---------------------------------------------------------------- Philip Warner | __---_____ Albatross Consulting Pty. Ltd. |----/ - \ (A.C.N. 008 659 498) | /(@) ______---_ Tel: (+61) 0500 83 82 81 | _________ \ Fax: (+61) 0500 83 82 82 | ___________ | Http://www.rhyme.com.au | / \| | --________-- PGP key available upon request, | / and from pgp5.ai.mit.edu:11371 |/
В списке pgsql-hackers по дате отправления: