[Pkg-db-devel] Bug#437742: Bug#437742: libdb4.6: FTBFS on arm: in testsuite, run_subsystem: env env007: db open:invalid argument

[Steve Langasek]

On Wed, Sep 03, 2008 at 07:52:30PM +0000, Clint Adams wrote:
> On Tue, Sep 02, 2008 at 01:33:54AM -0700, Steve Langasek wrote:
> >   echo "source ../test/test.tcl; run_std env007" | tclsh8.5

> > How do these tcl commands map to C, for better debugging?

> Without looking into the tcl bindings, I would guess that
>  berkdb_env -create -home ./TESTDIR -txn -encryptaes $passwd
>  berkdb_open -create -btree -mode 0644 -env $env -encrypt a.db

> maps (modulo error checking) to something like

>     DB_ENV *env;
>     DB *d;
>     int i;
> 
>     i = db_env_create(&env, 0);
>     i = env->set_encrypt(env, "passw0rd", DB_ENCRYPT_AES);
>     i = env->open(env, "./TESTDIR", DB_INIT_TXN | DB_CREATE, 0666);
>     i = db_create(&d, env, 0);
>     i = d->open(d, NULL, "a.db", NULL, DB_BTREE, DB_CREATE, 0644);

> or something.

> As for disabling encryption on a release arch, I don't know.

Had to refine this a bit to reproduce the error from the tcl test:

#include <stdio.h>
#include <string.h>
#include <db.h>

int main() {
    DB_ENV *env;
    DB *d;
    int i;

    i = db_env_create(&env, 0);
    i = env->set_encrypt(env, "passw0rd", DB_ENCRYPT_AES);
    i = env->open(env, "./TESTDIR", DB_INIT_MPOOL | DB_INIT_TXN | DB_CREATE, 0666);
    i = db_create(&d, env, 0);
    i = d->set_flags(d, DB_ENCRYPT);
    i = d->open(d, NULL, "a.db", NULL, DB_BTREE, DB_CREATE, 0644);
    printf("error is: %s\n",strerror(i));

}

This gives the EINVAL from the tcl test, and appears to be otherwise
correct.

But figuring out where this EINVAL originates seems to be non-trivial.

-- 
Steve Langasek                   Give me a lever long enough and a Free OS
Debian Developer                   to set it on, and I can move the world.
Ubuntu Developer                                    http://www.debian.org/
slangasek at ubuntu.com                                     vorlon at debian.org