Danny Yoo <d...@cs.wpi.edu> writes: > On Sunday, May 6, 2012, prad wrote: > > i'm trying to understand a consistent error i'm getting while using a > racket program to get data from a website. > > i'm accessing the site to download certain posts using regexps and store > them in sqlite3. > > things go perfectly till i accumulate 336 records and then this error > crashes my script: > > Hi prad, > > Are you doing the open-input-url in a loop, out of curiosity? If so, do you > close the opened ports between each loop iteration? > hi danny! i using get-pure-port and yes it's in a loop. i also delete-pure-port because i was getting a different error when i didn't which was to run out of file-descriptors (i think) since i kept getting a can't open database file error.
i found this: https://bloggingmath.wordpress.com/2008/12/15/sqlite-unable-to-open-database-file/ but instead of increasing the limits, i deleted the port. > 336 is a very specific observation. Is it repeatable? > very - 4 times in a row so far. > > tcp-connect: connection to www.30bananasaday.com, port 80 failed; host > not found (at step 1: Host name lookup failure; errno=2) > > so if it is a lookup failure that suggests the resolving server doesn't > like me after a certain number of lookups within a specific period of > time? is this reasoning correct? > > This particular error seems to be blaming host name lookup. As far as I > understand, the error's occuring even before you program can contact the > remote server, if I understand the error message. Odd. If it were some kind > of deliberate rate limiting, I'd expect a different class of error message. > i was thinking that since my resolving server was shawcable, may be they put limits on. i could try providing a ip, but i'm curious to see what is causing the error. -- in friendship, prad ____________________ Racket Users list: http://lists.racket-lang.org/users
