Leszek,
Thank you for turning me on to the possibility of using an interceptor of a
predefined object. This may be the gateway to several other approaches for
integrating parallelism to V8 that sidelined me. I'll admit the last time
I wrote code using V8 headers was v0.9 and now my full-time job has nothing
to do with this, but I'm approaching this as a long-term effort and your
response has breathed new life into an old idea.
Many of my previous interactions with the Node/V8 community have been
between awful and baffling so I want to reiterate my appreciation for your
response that goes beyond the technical advice.
-J
On Monday, July 26, 2021 at 5:27:21 AM UTC-4 les...@chromium.org wrote:
> On Sat, Jul 24, 2021 at 2:02 AM Jace Mogill <ja...@mogill.com> wrote:
>
>> Complementary data and task parallel JSON lexing implementations already
>> exists (https://github.com/simdjson/simdjson,
>> https://github.com/mogill/parallel-xml2json) but there's no way to store
>> the results in a way V8 can use, and thus an extra copy in/out step is
>> needed.
>>
>
> Note that V8's JSON parsing generates mutable, introspectable JavaScript
> objects, with all their prototypal inheritance, monkey patching
> possibility, iteration semantics, garbage collection compatibility,
> interaction with V8's inline caches for property access, etc.. The actual
> JSON _parsing_ of V8's JSON parser is quite competitive, it's the rest of
> this stuff that takes the additional time.
>
>
>> I see several "bookend" options which may be combined to varying degrees:
>> - A lowest common denominator data storage sequence is defined by V8
>> - Applications gain the ability to introspect about V8's object storage
>> sequences at runtime
>> - Applications can tell V8 how data is stored in memory and V8 can
>> adapt to an existing storage sequence
>>
>
> Since you are already using the C++ API, you can already do this by
> creating Object templates and registering interceptors on them which defer
> to your existing external storage. This will make the JSON parsing fast, at
> the cost of property access and memory management being slow, but there's
> no way for these to _not_ be slow without integrating with the rest of the
> system, at which point you're back to paying all the costs of V8's existing
> JSON parsing.
>
--
--
v8-users mailing list
v8-users@googlegroups.com
http://groups.google.com/group/v8-users
---
You received this message because you are subscribed to the Google Groups
"v8-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to v8-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/v8-users/109bb28a-0f99-4f3a-9db9-00143e677a7an%40googlegroups.com.
