Unicode collation sequences for Kexi - the solution

[Jaroslaw Staniek]

On 18 December 2011 22:48, Roger Binns <rogerb at rogerbinns.com> wrote:

> The preferred solution for most SQLite using developers is to include a
> private copy of SQLite.  You just add the single amalgamation source file
> to your project and are no longer at the mercy of whatever goes on on the
> platform.

Roger, thanks for the portion of convincing notes.
My hope is to be finally able to use a private copy of sqlite, what I
actually have been doing in 2004-2008 or so. Then I switched to
'system' SQLite and only had no visible problems only because so far
Kexi uses only basic features of sqlite3 and does not try to optimize
its usage.

While talking to Linux distributors a bit I've noticed their reasoning
for refusing private copies, e.g. see comments posted under my blog
entry http://blogs.kde.org/node/4156. While I understand this, SQLite
is special case for the reasons you explained very well. It would save
weeks of developing workarounds if only we could make some coordinated
effort to turn these recommendations to something official in eyes of
distributors and make them allow for exceptions in their rules. I mean
for Linux, since deployment for Mac and Windows can be controlled.

Another reason for wanting the private copy is specific approach to
versioning in sqlite. The file format changed in  areas relevant for
most users, and without a bump in major version. Thus distributors
unfortunately do not bump major version of libsqlite3.so, so in turn,
there's not way to have two versions of 'system' sqlite3 installed
side-by-side unless, e.g. to enable data migration tools. Private
copies fix this as well.

-- 
regards / pozdrawiam, Jaroslaw Staniek
 http://www.linkedin.com/in/jstaniek
 Kexi & Calligra (kexi-project.org, identi.ca/kexi, calligra-suite.org)
 KDE Software Development Platform on MS Windows (windows.kde.org)