Re: Error reporting should be improved

[Bob Friesenhahn]

On Wed, 30 Dec 2009, Peter Rosin wrote:
> The thread safety thing can be ignored as libltdl isn't thread safe
> and needs to be mutexed anyway. Care to test this patch?

There used to be some useless effort to make it thread safe. 
Apparently that is gone now.  Regardless, there is also a potential 
re-entrancy issue, which could be encountered if a legacy OS interface 
is supported (re-implemented) via dlopen() or if some other 
implementation logic (e.g. lazy symbol resolution) uses dlsym().

An alternative approach would be to store loader-specific errors in 
loader-specific storage and iterate through the loaders in order to 
obtain their error messages.

With this patch, how will the accumulated error strings be reported to 
the API user? Some of the error reports will be useless and so it may 
not be useful to combine them since a useless report would have the 
same significance as an important one.  The origin (id of the 
reporting loader) would be useful to record.  If the origin was part 
of the string (or the string is owned by the loader) then that would 
serve to keep them distinct.

Unfortunately, lt_dladderror() is publically exposed so it is 
difficult to change its arguments even though it is not documented.

Bob
--
Bob Friesenhahn
address@hidden, http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/