On Mon, Aug 26, 2019 at 11:54 AM Junio C Hamano <gitster@xxxxxxxxx> wrote:
>
> Carlo Marcelo Arenas Belón <carenas@xxxxxxxxx> writes:
>
> > e87de7cab4 ("grep: un-break building with PCRE < 8.32", 2017-05-25)
> > added a restriction for JIT support that is no longer needed after
> > pcre_jit_exec() calls were removed.
>
> I was initially puzzled by this statement, until I realized that the
> removal of pcre_jit_exec() happens in the topic still in flight that
> this patch builds on top of, namely 685668fa ("grep: stop using a
> custom JIT stack with PCRE v1", 2019-07-26).
sorry about that, I thought I had mentioned it in the cover letter
(since the hash is likely to change and so is not fit for a commit
message) but it is not there either.
how could this be tracked more effectively?
> So the logic is that because we do no longer call pcre_jit_exec()
> that weren't available between 8.20 and 8.32, these slightly older
> versions can now do JIT just like the ones post 8.32?
exactly; but also because it is no longer using the JIT fast path
which skipped UTF-8 validation, will need a way to disable that or
risk a regression as I mentioned in [1]
was planning in proposing a fix for PCRE1 based on [2] but wasn't sure
if it could be part of this series, an independent one that is also
based on ab/pcre-jit-fixes, and like this one, is mostly a consequence
of 685668fa ("grep: stop using a custom JIT stack with PCRE v1",
2019-07-26) or something else, specially considering that Ævar
dismissed it as a non issue in his commit message.
Carlo
[1] https://public-inbox.org/git/CAPUEspj4BJLjXorUXMiZnFtNcmhym_2QL5GUqeaGaCoxk=zjtw@xxxxxxxxxxxxxx/T/#m6acc8f68c398951457da469530bafa7e18811366
[2] https://public-inbox.org/git/20190721183115.14985-1-carenas@xxxxxxxxx/
