Re: [PATCH v3 1/5] GDB: Ignore `max-value-size' setting with value history accesses

[Tom Tromey <tom@tromey.com>]

>>>>> "Maciej" == Maciej W Rozycki <macro@embecosm.com> writes:

Maciej> We have an inconsistency in value history accesses where array element 
Maciej> accesses cause an error for entries exceeding the currently selected 
Maciej> `max-value-size' setting even where such accesses successfully complete 
Maciej> for elements located in the inferior, e.g.:

Thank you for the patch.

gdb has other inconsistencies when using the value history as well.  If
you print a struct holding a char*, the printers will generally fetch it
and print the string.  Then if you 'print $3' at some later point, the
string will be re-fetched and the new value displayed.  Occasionally
confusing; it seems to me it would be better for the value history and
the printing mechanism to conspire to keep track of all memory read
during the printing.

Maciej> Moreover a 
Maciej> statement in documentation says, concerning this setting, that: "Setting 
Maciej> this variable does not effect values that have already been allocated 
Maciej> within GDB, only future allocations."  While in the implementer-speak 
Maciej> the sentence may be unambiguous I think the outside user may well infer 
Maciej> that the setting only applies to values that need to be retrieved from 
Maciej> the debuggee.

Perhaps this could be reworded to mention "values previously computed"
or "printed" to tie it better to user actions rather than internal
state.

Regardless, I think this patch is ok.

Tom