On 2/20/06, Kon Lovett <[EMAIL PROTECTED]> wrote:
It turns out that the problem is that sha2-base.c now defines versions of SHA256_Final (and other functions) that is incompatible with that in libssl. You switched the order of arguments to most of the functions in sha2-base.c, which is what caused this. Since libpq depends on libssl, if it's loaded first, the sha2 egg will call the wrong version of SHA256_Final. This was probably an error waiting to happen anyway, because the version in libssl is probably subtly different even without a different prototype, or will be someday.
So what should we do? Mangle the names of everything in sha2-base.c, or make it depend on libssl?
Reed Sheridan
Could be something I did, message-digest is mine, and it was recently
updated. However, it doesn't explain the above. (I will look into
this more tomorrow.)
It turns out that the problem is that sha2-base.c now defines versions of SHA256_Final (and other functions) that is incompatible with that in libssl. You switched the order of arguments to most of the functions in sha2-base.c, which is what caused this. Since libpq depends on libssl, if it's loaded first, the sha2 egg will call the wrong version of SHA256_Final. This was probably an error waiting to happen anyway, because the version in libssl is probably subtly different even without a different prototype, or will be someday.
So what should we do? Mangle the names of everything in sha2-base.c, or make it depend on libssl?
Reed Sheridan
_______________________________________________ Chicken-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/chicken-users
