Request for review r1833421 (was: Reminder and caution)

[Yann Ylavic <yl...@gmail.com>]

On Wed, Jul 25, 2018 at 4:43 AM, Bill wrote:
>
> Change Process
>
> Most changes (bug fixes and minor, commonsense feature adds) do not
> require review. Developers are encouraged to request review for:
[]
> - Changes to interfaces Changes that commit APR to one option out
>   of an exclusive set

IIUC, my crypto libs' linking changes, needed IMO for any sane
implementation of the new PRNG, would match this case.

We are probably already above the "request for review" state since
r1833421 was veto-ed, though I disagree with the arguments put
forward.

So could the next step be:

> - Any change the developer wants a second (or Nth) opinion on

or does the veto imply that I should revert the commit anyway, with no
further discussion and/or others' opinions?

Thanks for letting me know because I don't want to (and won't) dispute
indefinitely for a lost cause...
Since I'm not really satisfied by the apr_crypto code in its
driver/DSO shape, I can as well ignore it and move out the PRNG code
to APR core (by copying a reference implementation of the needed
crypto primitive, from public domain, and quite simple btw).

Regards,
Yann.

Re: Request for review r1833421 (was: Reminder and caution)

[Yann Ylavic <yl...@gmail.com>]

On Wed, Jul 25, 2018 at 12:05 PM, Yann Ylavic <yl...@gmail.com> wrote:
>
> Since I'm not really satisfied by the apr_crypto code in its
> driver/DSO shape, I can as well ignore it and move out the PRNG code
> to APR core (by copying a reference implementation of the needed
> crypto primitive, from public domain, and quite simple btw).

I probably missed the point here that r1833421 is not only about the
PRNG needs, it's mainly about the APR providing (to the user) a way to
initialize/terminate the crypto libraries used internally, but also
possibly externally in multiple flavors.