Re: Provide a means to activate dynamically the debug traces

[Akim Demaille]

Le 10 janv. 09 à 22:25, Ralf Wildenhues a écrit :

> > So what do you recommend to store persistent information?  A function
> > that returns some static variable (sort of a singleton)?  Or an
> > additional argument everywhere (gee...).  Maybe I can do it using
> > advices to hand over the verbosity flags, but that does sound like a
> > nuisance for the API.
>
> A lt_dlget* and a lt_dlset* function, if the user needs to be able to
> provide a type.  Don't provide the user with a direct pointer if you
> don't have to.
>
> Note that I didn't say global variables are bad, although they are
> obviously bad, too, for two reasons: slow access, and, well,  
> globality.
> Assume that you write a library that uses ltdl internally, and I use
> your library in a program that also links against a completely  
> different
> library that also uses ltdl.  You'll most likely not want  
> interference.
> Well, in your case I'm not so sure.  But if not, then adding a module
> loading advisor to change verbosity would seem appropriate.

I thought about this, and I think that the best option would be to  
have a callback function with  a signature more or less like:

void log(void* user_data, level_type level, const char* format, ...);

(or maybe return an int in the spirit of ltdl to count errors?)

It could be given as a new kind of advise for the callback, and the  
callback user data.  It should cover all the needs, and should make it  
quite easy for C++ users to have wrappers that use the user_data to  
pass whatever log object they might use.

It is no longer ltdl that would decide whether to record or not the  
message, the level is given to the user.  I'm ambivalent wrt level:  
should it be a magnitude, or a bit-field for the various kinds of event?

One problem with the advise-based approach is that not everything  
works via advisors.  For a start lt_init does not.  It also makes it  
somewhat inconvenient for user who decided to use the basic interface  
to have logs easily.

So maybe it should rather be handled as a global inside ltdl, rather  
than an additional argument passed argument, possibly passed to  
lt_dlinit*, or even before lt_dlinit.

What do you think about it?