GnuPG can't handle some V3 secret keys

Werner Koch wk at
Sat May 5 19:20:02 CEST 2001

On Fri, 4 May 2001, Stefan Bellon wrote:

> I'm not sure whether I got this right. You favour writing code without
> tests for NULL as the OS should react accordingly? Or did I
> misunderstand you?


If you read code written by me you will notice a lot assert()s and
similar constructs at many places even on places where you think,
this can't go wrong.  It has been proven over the years that this is
a very effective way to detect logical errors.

If I do not use an assert but have doubts that other parts of the
code might rely on a freed pointer for example, I set it to NULL and
a backtrace will show you what was wrong.

IIRC, the early implementation of mpi_read() aborted on unreadable
stuff.  I later changed this to return NULL but obviously didn't
fixed all callers.  Frankly the parsing code used to have no good
error handling at all.


Werner Koch        Omnis enim res, quae dando non deficit, dum habetur
g10 Code GmbH      et non datur, nondum habetur, quomodo habenda est.
Privacy Solutions                                        -- Augustinus

More information about the Gnupg-devel mailing list