summaryrefslogtreecommitdiff
path: root/tests/test_sqlite.cc
AgeCommit message (Collapse)AuthorFilesLines
2012-05-30Fix type error in test_sqlite.ccDaniel Hartwig1-3/+3
2009-08-22Hide statement execution behind an RAII class to make it harder to ↵Daniel Burrows1-107/+151
accidentally leave an active statement. SQLite gets into all sorts of bad states if you don't finish or reset all your statements -- e.g., it'll prevent you from rolling back transactions, which means you can never do anything with the file cache again.
2009-07-30Wrap sqlite3_blob_read and sqlite3_blob_write.Daniel Burrows1-0/+60
2009-07-28Write a wrapper for sqlite3_blob_open.Daniel Burrows1-0/+50
2009-07-27Add support for parameter binding to the sqlite wrapper.Daniel Burrows1-0/+167
2009-07-27Implement a system for caching prepared SQL statements.Daniel Burrows1-0/+45
Prepared statements created via this mechanism are placed into the cache for reuse once the user is done with them. Statements won't be reused as long as at least one proxy object still exists. This gives us a mechanism for safe reuse of statements without having to worry that two different pieces of code might use the same statement object and step on each other. Also, boost::multi_index is awesome.
2009-07-26Wrap enough functionality in the statement class to allow the user to ↵Daniel Burrows1-0/+127
execute statements and retrieve results. This explicitly doesn't try to deal with the inherent non-threadsafety of the sqlite statement type. That will only come up if you're trying to reuse statements, presumably for performance, and once we're doing that, the code that mediates reuse should ensure that only one client has its hands on a given statement at once. (even non-threaded reuse is dangerous!)
2009-07-26Start working on a C++ wrapper for sqlite.Daniel Burrows1-0/+42
This will be used to implement a backend cache for the various data that we download about packages (changelogs, screenshots, etc). This also includes the very first Boost-based unit test in the tree.