Tom> http://sourceware.org/ml/gdb-patches/2008-08/msg00525.html

Ulrich> Ah, I'd forgotten about that thread.  I'll have a look at the
Ulrich> above patch.

Thanks!

Ulrich> In fact, I'm wondering if it wouldn't be nicer to also have something
Ulrich> like a get_default_print_options function instead of refering to the
Ulrich> user_print_options global in the top-level printing routines ...

I implemented this, and I fixed the other problems you mentioned.
The result is appended.  I still haven't tested this or written a
ChangeLog entry.

This version reduces the use of user_print_options to the absolute
minimum.  I don't think this is notably cleaner, but the difference
isn't extreme.  If we cared, we could make it mildly less ugly by
having get_user_print_options return its argument.

I still have a version of the patch without get_user_print_options in
case you change your mind on this.

Doing this did point out a couple things:

* As you mentioned, breakpoint-printing could probably use a small
  refactoring to give value_print_options arguments to various
  methods.  I can do this and roll it into this patch if you want;
  though my first reaction is to leave it as a follow-up patch (or
  just not do it).  Let me know what you want.

* print_scalar_formatted probably needs an options argument as well.
  What do you think?  This worries me a little since it is another 47
  call sites; so it could make the patch even bigger.  Maybe this
  ought to be a follow-up.

Ulrich> OK, I'm fine with this.  However, this means that we'll really have
Ulrich> to construct the option struct on the fly, we cannot have a global
Ulrich> struct hold pre-initialized pointers to current_language etc.

There's always "GCC style":

    #define current_language user_print_options.language

Semi-awful.

Or just s/current_language/user_print_options.current_language/
everywhere.  This, IMO, is not terrible, aside from its verbosity.

This is related to Joel's patch and to Daniel's question about a
policy for access to globals.  Before going too far down this road it
may be worth coming up with an agreement there.


To sum up, once I know how much more I should do, I will implement it
and then submit the patch for real.  I am not looking forward to
writing this ChangeLog entry...

Tom