[R-sig-DB] [PATCH] segfault in RSQLite 0.5-4

[Seth Falcon]

"Ashish Kulkarni" <ashish.kulkarni using kalyptorisk.com> writes:

>> Adding such an experiment to the test cases would be most helpful. :-)
>> 
>> [snip]
>> 
>> How about renaming the test and using RUnit's DEACTIVATED().  This
>> way, the tests will run, but we'll be reminded to follow up on this.
>> 
>> Would you be willing to redo/add to your patch by adding the "add
>> column" test and the above rename/DEACTIVATE?
>
> done, attached updated patches. I've not been able to test them as I'm not
> at my regular workstation.

Thanks.  Getting closer.  I just tested against trunk (which includes
an update to SQLite 3.4.1) and I get:

RSQLite Test Suite - 21 test functions, 1 error, 0 failures
ERROR in testSchemaChangeDuringQuery2: Error in sqliteExecStatement(con, statement, bind.data) : 
  RS-DBI driver: (error in statement: no such table: t1)

This is surprising -- even if this is going to error out, the error
message doesn't make any sense to me.

Ashish: I really appreciate that you are sending patches and I would
like to update the code to the recent SQLite API changes, but I think
this one is not quite cooked yet.  It isn't clear to me whether these
are SQLite issues or an issue related to the patch, but it seems that
applying it results in a regression in robustness in the case of
multiple connections to the same SQLite DB.

+ seth

-- 
Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center
http://bioconductor.org