Short story: i fixed DBD::SQLite's annoying warning, diff at
http://rt.cpan.org/Ticket/Display.html?id=40383
Long story: today a friend proposed a patch to silence those warnings in
another module and i was finally annoyed enough to investigate.
First i've read the various reports about this, their analysis and
workarounds, then i glanced at the sqlite documentation and found
something promising.
I tried adding a fix in a quick and dirty style until i noticed sqlite
required an update first, as the used function was only available since
sqlite version 3.6.0.
After updating sqlite the fix still didn't work and i saw ``assignment
makes pointer from integer without a cast'' when compiling.
It took a few minutes until i realized that DBD::SQLite picks up the
header files of the bundled sqlite instead of the system-wide installed
one and thus the compiler had no idea about the new function i was using.
The first step to a solution was to zap the local sqlite headers and
then i had it working.
But that was still not very nice as it only helps me but i wanted to get
a result i could push upstream.
So the next logical thing was updating the bundled sqlite before i had a
sufficient patch to send to the DBD::SQLite authors.
A happy end to the story of this bug should be possible with the creation
of a proper ticket on RT.
Now we ``only'' have to wait for a new release of DBD::SQLite.
At least i did more than just moaning :-)
Re:DBD::SQLite::Amalgamation
janus on 2008-10-26T21:55:42
Thanks for the pointer, the same patch applies and i've added a ticket[0] with it.
I believe there's already permission from MSERGEANT for it to happen, at least to Matt Trout... I imagine we could get you in on the act.
