On Tue, Feb 8, 2011 at 13:30, Marko Kreen
<markokr@gmail.com> wrote:
Yeah, seems your analysis is right and it's libpq (design?) bug,
so no way to work around it.
Nice to know that my analysis is right... but eventually it is a bad news because the only way to fix the issue is to split input data into many chunks or to directly modify libpq... By the way, I have posted my analysis to pgsql-hackers and asked what the list thinks about modify pqParseInput to handle this case. But I have never posted to this list before so I don't know what is the good way to ask for feedback.
Nicolas Grilly