Re: gdb/665: target_changed_event & store_register/target_write_memory

[Andrew Cagney <ac131313 at ges dot redhat dot com>]

The following reply was made to PR gdb/665; it has been noted by GNATS.

From: Andrew Cagney <ac131313@ges.redhat.com>
To: keiths@redhat.com
Cc: gdb-gnats@sources.redhat.com
Subject: Re: gdb/665: target_changed_event & store_register/target_write_memory
Date: Wed, 21 Aug 2002 12:18:09 -0400

 > It would be awefully nice if we could just put target state event notifications into store_register and target_write_memory instead of having to scatter notifications all over the place. Unfortunately, setting breakpoints will interfere with this.
 Just FYI, there is a function read_memory_no_bp() that hides memory 
 breakpoints.  It would be good though, if the default memory write 
 funciton always hid breakpoints and hence could be used.
 
  >  (What about setting a local variable? A bit of an over-reaction, no?)
 
 No.
 
 -  GUI refresh after a variable write should never take longer than a 
 single-step GUI refresh.  Given that single step has to be <1 second for 
 GDB to be considered useable, this is more than suffient.
 
 -  GDB's just had a month old register (and register variable) write bug 
 fixed.  No one noticed!  It was found, by accident, while reading the code.
 
 So, provided they work, GUI writes <=single-step (<1sec) are sufficient.
 
 enjoy,
 Andrew