[Ironpython-users] IronPython SQLite interface

[Doug Blank]

On Fri, May 9, 2014 at 8:59 AM, Jeff Hardy <jdhardy at gmail.com> wrote:

> On Fri, May 9, 2014 at 12:01 PM, Doug Blank <doug.blank at gmail.com> wrote:
> > On Fri, May 9, 2014 at 4:42 AM, Jeff Hardy <jdhardy at gmail.com> wrote:
> >>
> >> On Thu, May 8, 2014 at 5:32 PM, Doug Blank <doug.blank at gmail.com>
> wrote:
> >> > Just getting back to attempting to use sqlite3 on computers other than
> >> > Windows, and it looks like this C# library has other problems on other
> >> > OSes.
> >> >
> >> > The first issue is that it appears that all activity must happen in
> the
> >> > same
> >> > thread (connection, cursor, execution). That can mitigated against if
> >> > that
> >> > is the way that it has to be. (A wrapper could be supplied to run
> >> > everything
> >> > in a specific thread).
> >> >
> >> > The next problem is a show stopper:
> >> >
> >> > import sqlite3
> >> > conn = sqlite3.connect("test.db")
> >> > cur = conn.cursor()
> >> > cur.execute("select * from tables;")
> >> >
> >> > This works on Windows, but gives a SystemError: LockFileEx on Linux. I
> >> > suspect that the information that Ivan pointed to (below) indicates
> that
> >> > these patches screwed up sqlite port so that it no longer will work on
> >> > Linux
> >> > and Mac.
> >> >
> >> > Anyone have ideas, on either a fix, or using a different library,
> >> > different
> >> > version?
> >>
> >> Right now, I don't know if there's much that can be done. The C#
> >> SQLite port it uses is pretty Windows-specific, and I doubt it's worth
> >> the effort to fix it since it's no longer maintained anyway. I've been
> >> pondering replacing the guts of the SQLite module with
> >> https://github.com/ericsink/SQLitePCL.raw, which wraps the *actual*
> >> SQLite library instead of the C# port.
> >>
> >> The module implementation shouldn't have to change that much, since
> >> most of the function names should be the same. I don't yet know if
> >> SQLitePCL works on Unix/Mac but it does support Android and iOS so
> >> it's probably going to be a lot easier to fix it than it would be for
> >> C#-SQLite.
> >>
> >> I hadn't planned on this until IronPython 3, but I'd be very happy to
> >> include it in 2.7.5 (or a possible 2.7.6) if someone else does the
> >> work.
> >
> >
> > Thanks for the information!
> >
> > As I am also interested in other .NET/Mono languages (and other DB
> > connections), I wonder if we are going to spend any time on this, if it
> > wouldn't be better to use (and perhaps refine, if needed) the ADO.NETSqlite
> > interface:
> >
> > http://www.mono-project.com/SQLite
> >
> > Perhaps it is better to move away from the sqlite-specific API and
> towards a
> > standard that would allow people to move between db backends? I suspect
> Vern
> > may have some information on this point, considering his recent
> announcement
> > here on adodbapi version 2.6.
>
> I actually tried to use that, originally, but Python's sqlite module
> exposes some low-level features of SQLite that can't be emulated with
> an ADO-type interface. The only way to implementing it faithfully is
> to have direct(ish) access to the SQLite library.
>
> Most real apps will use a higher-level wrapper, but IronPython still
> needs to provide the sqlite module, since most of the high-level
> wrappers will expect it to be there.
>

Agreed. But maybe there is a way to still use the Mono.Data.Sqlite code
base? Perhaps add in more functionality (wouldn't be exposed by ADO),
accessible via a intermediate API. It seems a waste to have two low-level
interfaces to sqlite3...

-Doug


>
> - Jeff
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/ironpython-users/attachments/20140509/78e93666/attachment.html>