Given the lack of other responses, I'll take a guess:
- FLAGs are per-process, and mutable at runtime. When threaded cctests
mutate them, weird behavior tends to result, as most code paths aren't
prepared to handle randomly changing FLAG values.
- Whenever activation of a given feature depends on more than just the flag
value (e.g., implied by other flags, unsupported in special cases, ...),
it's cleaner and more flexible to compute the resulting state in a local
variable. It can be argued that this should be done consistently; it may
well be that someone applied such consistency to parts of the codebase ;-)

On Fri, Jul 15, 2016 at 11:21 PM, Adam Klein <[email protected]> wrote:

> It seems the current practice of copying flags into allow_ fields began
> around https://codereview.chromium.org/13450007. That's also when
> cctest/test-parsing was updated to have a set of flags for parser testing.
>
> My question is this: is there any concrete reason that we need separate
> fields for these, rather than simply accessing the FLAGs directly in the
> parser?
>
> Asking because a new contributor asked me on irc, and I couldn't come up
> with a good reason (I'd thought it might have something to do with
> threading, since the parser can run off the main thread, but that doesn't
> seem to be the case, since we do access some flags in the parser).
>
> - Adam
>
> --
> --
> v8-dev mailing list
> [email protected]
> http://groups.google.com/group/v8-dev
> ---
> You received this message because you are subscribed to the Google Groups
> "v8-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>

-- 
-- 
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
--- 
You received this message because you are subscribed to the Google Groups 
"v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to