Re: "avoid running GC when SCM_I_CURRENT_THREAD is unset"

[Neil Jerram]

Andy Wingo <address@hidden> writes:

> Hi Neil!
>
> Adding guile-devel; I'm wrong a lot lately, and folks should know ;-)
>
> On Wed 30 Mar 2011 21:11, Neil Jerram <address@hidden> writes:
>
>> I saw this change and couldn't fully understand it.  Given that you
>> ended up adding GC_disable() and GC_enable() around the GC_malloc()
>> call, couldn't you just have done that without all the other changes
>> from "t->" to "t." ?
>>
>> I'm probably missing something, but could you point it out?
>
> Indeed, I probably could have just done a GC_disable before the whole
> thing, and an enable() afterwards.  That's probably clearer, no?

I doubt it's worth going back, now that the change is made.

I really just wanted to make sure I understood - because I'm seeing a
problem on ARM that looks GC-related, and so was reviewing recent
GC-area changes.  I'll post a bit more about that in a new thread.

      Neil