We have done it for every getter. Now we do it only once as we ask for
the values. It simplifies the code and the test and could even improve
performance.
Change-Id: Ia7d4a33a77ec7c0a5fda548424fbf8b192f07511
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Sqlite has a function to get the last inserted rowid but very often you
want to get the updated rowid too.
Change-Id: Ie276a5039682813ad16597433996a2959f54d9ba
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
It is still only support references in columns but so far it is enough.
Change-Id: Iebb4866cf738d651270e54357b5e4a2837f05417
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
It adds a layer if you don't know if the type is integer, float or string.
It does not handle bytearrays here because so far there is no need. There
are two classes, Sqlite::Value and Sqlite::ValueView. Value owns the
string, ValueView holds only a view the string. So there is no allocation.
It is designed to hold Utf-8 string like Sqlite but it can be easily
converted in and from QString or QVariant but mind about that this is not
free. ValueView has no constructors on perpose because it would be
ambiguous if there would be constructors for the other primitives of
the Sqlite layer like "int64", "double" and "string view".
Change-Id: Ia39364eb2fc1998e5c59fdb4316add22c748507d
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
It can be happen that the entry is written by an other connection after
we tried to read and before we write. This would lead to a double entry
which be prevented by the unique index in the database. In that case we
simply try again and read the id from the database.
Change-Id: I6c9d94e95ae11556bb446813f64be0855be4ddbe
Reviewed-by: Ivan Donchevskii <ivan.donchevskii@qt.io>
The database is using file path integer ids to handle file paths because
otherwise we would save many redundant data. This patch is improving it
further with the introduction of a database based file path cache. The
entries are now divided in a directory path and file name. This is quite
handy for directory based file watching.
Change-Id: I03f2e388e43f3d521d6bf8e39dfb95eb2309dc73
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Introducing different exceptions for different error cases.
Change-Id: I4371d1e64d9dca2a9f68dcbaa4a891c55879c1f5
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Marco Bubke <marco.bubke@qt.io>
We use a name space now, so the prefix is not of much use.
Change-Id: I2b077576f94dab778add6ab2e54870f7ca18da78
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
We don't need the threading anymore, so we removed it. The indexer
will be run in its thread anyway, so an extra thread makes the code
only more complicated. And we added namespaces.
Change-Id: Ibcba306324763285cf653c28bb08122345e5f8da
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
This is a partial result of wip/clang-oop. More will follow.
This allows us to invoke the completion out of the Qt Creator process
and thus safes us as against libclang crashes.
At this point only the completion use case is supported.
Some notes on the individual components:
src/libs/codemodelbackendipc
* library encapsulating the inter process communication handling
* used by the backend application and in a follow-up change by the
creator integration
src/libs/3rdparty/sqlite
* version 3.8.10.2
* dependency of codemodelbackendipc, will be used to storage indexing
data, among others
src/tools/codemodelbackend
* the backend application
tests/unit:
* unit tests
Change-Id: I91a48e27467581a22fb760a18d8eb926008fea60
Reviewed-by: Alessandro Portale <alessandro.portale@theqtcompany.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@theqtcompany.com>
Reviewed-by: Marco Bubke <marco.bubke@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
