[PATCH] Add stackframes to public libsmbclient functions

[Gerald (Jerry) Carter]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Derrell Lipman wrote:

> In this case, I'm not the original author of the code, so 
> unless I've had reason to change it, I've just maintained
> the original organization of the code.  Even with a single
> entry and exit point, though, I'd still have a hard time
> justifying "magic" inputs to the lower-layer functions.
> Without some strong reason to the contrary, I prefer the
> black-box concept with inputs and outputs well defined by
> parameters and return value.

I understand.  And since I'm not the one working on libsmbclient,
it's certainly not my place to attempt to heavily influence
it's direction.

The talloc_stackframe() is a good idiom in my opion for

(a) removing the short lived tmp_ctx pointers that are needed
    randomly throughout the code.
(b) Allowing a caller to ensure that all short lived temp
    contexts are correctly cleaned up.

But then I've been bitten more than once of using the wrong
talloc_ctx and seeing huge memory blow up.






cheers, jerry
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHQa4DIR7qMdg1EfYRApBfAJ9BCXvgXxxNJ4vm69q9cDiUp3ELYgCfc025
OXN0NH0fys8qp9mJ49n1+8A=
=Nf/l
-----END PGP SIGNATURE-----