Hi,
I find the exposure of an implementation detail (usage of Apache HTTP
client) of the Elasticsearch client a bit problematic. If they change
this to another HTTP client, our SPI would break.
Did you instead consider to just let users provide their custom
instance of org.elasticsearch.client.Re
On Fri, 2 Jun 2017, 08:56 Gunnar Morling, wrote:
> Hi,
>
> I find the exposure of an implementation detail (usage of Apache HTTP
> client) of the Elasticsearch client a bit problematic. If they change
> this to another HTTP client, our SPI would break.
>
Yes the very point of exposing that detai
2017-06-02 10:30 GMT+02:00 Sanne Grinovero :
>
>
> On Fri, 2 Jun 2017, 08:56 Gunnar Morling, wrote:
>>
>> Hi,
>>
>> I find the exposure of an implementation detail (usage of Apache HTTP
>> client) of the Elasticsearch client a bit problematic. If they change
>> this to another HTTP client, our SPI
>
> Did you instead consider to just let users provide their custom
> instance of org.elasticsearch.client.RestClient? It's still leaking an
> implementation detail of Hibernate Search, but at least it's one
> indirection less.
The only way to add AWS authentication to the RestClient is to use Ap
2017-06-02 11:58 GMT+02:00 Yoann Rodiere :
>> Did you instead consider to just let users provide their custom
>> instance of org.elasticsearch.client.RestClient? It's still leaking an
>> implementation detail of Hibernate Search, but at least it's one
>> indirection less.
>
>
> The only way to add
On 2 June 2017 at 10:58, Yoann Rodiere wrote:
>> Did you instead consider to just let users provide their custom
>> instance of org.elasticsearch.client.RestClient? It's still leaking an
>> implementation detail of Hibernate Search, but at least it's one
>> indirection less.
>
>
> The only way to
Hi Sanne, **
*Announcing: Early-Access builds of JDK 9 for Alpine Linux/musl at
jdk.java.net/9/ [1]
*
* As of today there are pre-built Early-Access (EA) JDK binaries for
Alpine Linux/musl at jdk.java.net/9/**
o look for “Alpine Linux”. [1]
* The Alpine Linux build is compatible
> There's an important difference: one exposes Apache HTTP client in
> HSEARCH's SPI, whereas the other just requires usage of Apache HTTP
> client within one specific implementation. For users it doesn't change
> much, but the latter is cleaner from HSEARCH's perspective.
RestClient is not an int
2017-06-02 12:49 GMT+02:00 Yoann Rodiere :
>> There's an important difference: one exposes Apache HTTP client in
>> HSEARCH's SPI, whereas the other just requires usage of Apache HTTP
>> client within one specific implementation. For users it doesn't change
>> much, but the latter is cleaner from H
2017-06-02 12:29 GMT+02:00 Sanne Grinovero :
> On 2 June 2017 at 10:58, Yoann Rodiere wrote:
>>> Did you instead consider to just let users provide their custom
>>> instance of org.elasticsearch.client.RestClient? It's still leaking an
>>> implementation detail of Hibernate Search, but at least it
> There's no exposure of HTTP Client in this SPI. Yes, if people need to
> customize the HTTP client to be used by the returned RestClient
> instance, they'll naturally depend on that. But this SPI isn't tied to
> such detail of how RestClient works - if ES folks decided to use
> OkHttp instead, ou
Sure, as I said, an implementor of that SPI will very likely deal with
the HTTP client. But our SPI isn't tainted by that.
I also don't see much value in re-defining all the options from
RestClientBuilder. ElasticsearchHttpClientConfigurer resembles
HttpClientConfigCallback. Would you also re-defi
I made a mistake, Header is an Apache HTTP Client type and FailureListener
depends on one. So we're down to one useful customization, which we should
probably support as an Hibernate Search option by the way:
- org.elasticsearch.client.RestClientBuilder.setPathPrefix(String)
Yoann Rodière
Hib
> What's the benefit of this catch-up game?
Not tainting our SPI with RestClient :)
Yoann Rodière
Hibernate NoORM Team
yo...@hibernate.org
On 2 June 2017 at 14:59, Gunnar Morling wrote:
> Sure, as I said, an implementor of that SPI will very likely deal with
> the HTTP client. But our SPI isn'
I reckon you are making fun of me. Anyways, I've had my say :)
2017-06-02 15:03 GMT+02:00 Yoann Rodiere :
>> What's the benefit of this catch-up game?
>
> Not tainting our SPI with RestClient :)
>
> Yoann Rodière
> Hibernate NoORM Team
> yo...@hibernate.org
>
> On 2 June 2017 at 14:59, Gunnar Morl
> Anyways, I've had my say :)
And I'm grateful you took the time to do it. I guess we both would have
preferred to agree on the same solution, but... at least we now have
something to process.
Yoann Rodière
Hibernate NoORM Team
yo...@hibernate.org
On 2 June 2017 at 15:11, Gunnar Morling wrote:
16 matches
Mail list logo