On 5/29/07, "Martin v. Löwis" <martin@v.loewis.de> wrote:
p &sqlite3InitCallback
(gdb) p $sqlite3InitCallback $1 = void grrrr.
Try "info shared" in gdb. Not sure whether that works on OSX, though.
Worked beautifully! The smoking gun: something is hauling in the system provided sqlite3 in addition to my static one. (Weren't you just saying that?) I'm going to investigate that further unless you tell me I'm an idiot. I suspect it's the official Qt binaries from Trolltech doing it indirectly. Plus that's the only way I can rationalize some sqlite functions having source available and some acting like they've been stripped. -- Darrin
Darrin Thompson schrieb:
On 5/29/07, "Martin v. Löwis" <martin@v.loewis.de> wrote:
p &sqlite3InitCallback
(gdb) p $sqlite3InitCallback $1 = void
Please try '&' instead of '$'. It's the address of that function I was after (to then find out whether it is in the address range of the extension module).
The smoking gun: something is hauling in the system provided sqlite3 in addition to my static one. (Weren't you just saying that?) I'm going to investigate that further unless you tell me I'm an idiot. I suspect it's the official Qt binaries from Trolltech doing it indirectly.
Plus that's the only way I can rationalize some sqlite functions having source available and some acting like they've been stripped.
Very likely. I think you will have to read up on "two-level namespaces" and stuff like that to resolve this. Regards, Martin
participants (2)
-
"Martin v. Löwis" -
Darrin Thompson