Marcus Ilgner wrote: > > I'm one of the authors of the SQLite-WinRT component for Windows 8. >
Cool, the one on CodePlex? > > Regarding this release, can you tell me what configuration was used to > compile it? > The default compilation options from the MSVC makefile were used, along with the extra options necessary for WinRT (e.g. SQLITE_OS_WINRT, WINAPI_FAMILY, etc). > > I'm asking because we wondered why the binary is about double the size > of the one we had previously. > >From where? With what compilation options? > > Also, there's a strange phenomenon occurring which is probably some > kind of race condition in the way that multi-threading is done: > sqlite3_step() returns a generic error code 1 (for a valid statement) > but when I proceed to call sqlite3_errmsg16(), the error message for > error 6 (table locked) is returned. > That is not necessarily a race condition. It could be the sqlite3_step is returning SQLITE_ERROR and the last error code stored for the database could be something else, like SQLITE_LOCKED. Also, the SQLite DLLs for WinRT are compiled with SQLITE_THREADSAFE=1. > > Generally speaking, is a locked table something that should be handled > or do you think there may be another problem at play here? I see that > SQLite does a lot of work to support seamless multi-threading and since > the one application is the only process working on the database, I > suspect that the problem may lie elsewhere. > How many processes and/or threads are attempting to access the database simultaneously? There can be only one writer at a time. Also, unless WAL is used, a writer may block a reader. -- Joe Mistachkin _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
