Re: caller GCPROs?

Monday, 1 March 1999

        Gunnar Evermann <ge204(a)eng.cam.ac.uk&gt; writes:

...
 > That's no lie.  buffer_insert_string_1() can not only GC,
but
 > invoke arbitrary Lisp code!  Remember
 > {before,after}-change-functions.

 I know. Maybe I misunderstood all those 'can GC'-comments. I assumed
 them to mean:

  - during execution of this functions Lisp code might be called and
    potentially GC can occur 
Yes.

...
  - this code is GC-safe, meaning it GCPROs all local Lisp_Objects
    and therefore a GC won't collect vital stuff. 
No, I don't think "this function can GC" means that.  David Moore's
"GC checked <date here>" might mean that, but the way I see it,
vanilla "this function can GC" simply means "if you call this
function, be prepared that a GC might happen".

...
 My question was basically why we don't GCPRO 'str' and
then use
 RETURN_GCPRO, to ensure the second point. 
Dunno.  Hysterical reasons, probably.  We should probably check how
buffer_insert_string_1() can be called and make all the callers
properly GC-protect RELOC.

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998