Gotcha, sounds fair. This is the kind of mistake that arises when we don’t talk
and merely see bits go by.
Once a few more folks confirm that 1843-feature-bigcouch stands up and can at
least do validate_doc_update on a few other platforms, I’ll consider the
rebar.config.script work in a good enough shape to discuss it for merging. It’s
broadly what you had except it turns out rebar itself can do the os:type
selection using regex’s and it’s possible to pass different environments for
each port (so couch_icu_driver is no longer linked to spidermonkey for no
reason).
B.
On 4 Feb 2014, at 16:06, Benoit Chesneau <bchesn...@gmail.com> wrote:
> On Tue, Feb 4, 2014 at 4:50 PM, Robert Samuel Newson
> <rnew...@apache.org>wrote:
>
>> Hi,
>>
>> It's largely because we're all busy with wholly bigcouch internal things
>> on our 1843-feature-bigcouch branch, there's nothing *yet* to
>> cooperate/feedback about. The closest thing is the bit I'm hacking on
>> ("Build with rebar").
>>
>> I agree that we should not be inventing new things during the merge, but
>> then I saw you were splitting out parts of couch_httpd elsewhere, which
>> counts as that.
>>
>
> Actually :
>
> https://github.com/refuge/couch_core/tree/master/apps/couch_httpd
>
> The difference is that couch_changes is not its own app like it is in
> rcouch:
>
> https://github.com/refuge/couch_core/tree/master/apps/couch_changes
>
> because i ported the latest but non released in that branch yet of the
> code. I also not integrated the usage of barrel, to easy any merge. So
> nothing really new in that split. I just simplified it compared to rcouch
> (if you see in the history this is how it started anyway) in view of
> speeding the final merge and propose other changes as patches later.
>
>
>
>>
>> Before bigcouch can be merged, it needs to build and run and perform as
>> many of the features as possible, that simply can't happen until 'make'
>> works. I deliberately started from your rebar.config.script rather than the
>> code that bigcouch uses (which uses Scons) as it's the right approach. That
>> it's not identical is not a problem, in my opinion, but we can judge that
>> at merge time. It needs to work first.
>>
>> I appreciate your "during 2 years" comment but let's not go there, I can
>> pull a larger value, as can everyone else at Cloudant, but that's going to
>> be tiresome and divisive and it achieves nothing.
>>
>> Short version: If it was as simple as dropping in the rcouch versions of
>> these pieces, I'd have finished a week ago. Software is not so simple.
>>
>
>
> ok... I will forget to answer to that part. I will rather focus on
> finishing these latest patches.
>
> - benoit
