This is the mail archive of the
mailing list for the newlib project.
Re: [PATCH] New configuration option for lite exit
- From: Jeff Johnston <jjohnstn at redhat dot com>
- To: newlib at sourceware dot org
- Date: Wed, 29 May 2013 15:29:36 -0400
- Subject: Re: [PATCH] New configuration option for lite exit
- References: <000001ce3b52$4646b880$d2d42980$ at arm dot com> <517BB415 dot 2080601 at embedded-brains dot de> <CAL0py27CtHj=h3tmr8W+CdnpM-zcF0GPbQxWeNJayYyqcB2REg at mail dot gmail dot com> <20130528102816 dot GE5264 at calimero dot vinschen dot de>
On 05/28/2013 06:28 AM, Corinna Vinschen wrote:
On May 28 18:12, Ye Joey wrote:
On Sat, Apr 27, 2013 at 7:18 PM, Sebastian Huber
why can't you use _exit() or _Exit()?
This patch makes calls to expensive clean up functions as weak
reference, so one can still keep original functionality by explicitly
pull-in symbols. Comparing to simply skipping them to _exit or _Exit,
this approach is more flexible and testing friendly. In another word,
user can still get a full functional libc even if it is built with
Please let me know if you still have concerns.
I don't know. I guess I defer the decision to Jeff.
Sorry I have been busy with an up-coming Eclipse release and two
Looking at it, this needs documentation as it is not straight-forward
with regards to how the whole design ties together (i.e. what are the
behaviour rules for the application / compiler). You might also state
what this accomplishes on your initial platform.
AFAICT, if a call is made to __cxa_atexit, nothing will occur unless a
call is also made to __cxa_finalize or atexit() or on_exit(). The exit
list won't be run on exit if just on_exit() is called because
__call_exitprocs() won't be brought in.
-- Jeff J.