Re: [rfc / remote protocol] Remote shared library events

[Daniel Jacobowitz <drow@false.org>]

On Tue, May 15, 2007 at 09:28:26AM -0400, Daniel Jacobowitz wrote:
> Here are just the infrun parts.  They've gained a test case for the
> trickiest part, a cleanup, a better changelog, and an explanation.
> 
> First of all, I don't remember why I was testing
> inferior_ignoring_startup_exec_events in the last version of this
> patch.  I discovered that now there are no references to it (there may
> have been one when I wrote it).  I didn't really need it, so I garbage
> collected it.
> 
> Secondly, I needed to make TARGET_WAITKIND_LOADED support work with
> the current solib.c implementation.  You can tell it hasn't been used
> in a while since it was conditioned on "#ifdef SOLIB_ADD"; there are
> almost no definitions of that left.
> 
> Third, while debugging shared library support for SymbianOS I kept
> encountering an unpleasant surprise.  I would disconnect from the
> remote target when it was stopped at a load event (usually because my
> GDB crashed :-).  When I reconnected, GDB would automatically resume
> the program!  That's where the ugly stop_soon test was in the last
> version of this.  For this version, I reviewed every setting and use
> of stop_soon and decided that none of them quite matched the case in
> stop_remote, so I added a new STOP_QUIETLY_REMOTE (and some comments
> about it).  This is just like STOP_QUIETLY_NO_SIGSTOP except that it
> does not need to hide a SIGSTOP if it finds one.  It means we're
> connecting to the target for the first time, so we shouldn't decide to
> resume it without giving the user a chance to look around.
> 
> I hope this version is clearer; what do you think of it?

Hi Mark, Kevin,

Did either of you have a chance to look at this?  You both had
concerns about the infrun parts of the combined patch, so I'd like to
make sure this version is better and clearer before I go on with the
other pieces.

-- 
Daniel Jacobowitz
CodeSourcery