mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexander Rukletsov" <ruklet...@gmail.com>
Subject Re: Review Request 35714: Added a new HTTP response type: PreconditionFailed.
Date Wed, 24 Jun 2015 13:55:38 GMT


> On June 22, 2015, 8:04 p.m., Ben Mahler wrote:
> > Adding this seems good since it's agnostic to how it is used. But for reservations
specifically, any reason not to use just BadRequest?
> > 
> > From what I can tell, this is generally used for conditional requests (based on
headers or some other conditionals): http://stackoverflow.com/a/5369582
> > What will the pre-conditions be for /reserve? If the request does not contain explicit
pre-conditions, it seems a little non-idiomatic to return 412..?
> 
> Michael Park wrote:
>     I was hoping to keep `BadRequest` to situations where the user has provided bad/invalid
arguments. e.g. invalid JSON format, missing parameters, invalid resources, etc.
>     
>     I was aiming to use `PreconditionFailed` to represent the case where the user have
provided perfectly valid arguments, but we don't currently have sufficient resources to satisfy
the request. The precondition here is whether we have sufficient resources or not. I'm of
course open to using `BadRequest` with an explanation or what went wrong, or perhaps using
a different status. (e.g. `Conflict` was suggested)
>     
>     I'm not an expert on HTTP statuses and so I don't have a strong stance here, do you
have a strong opinion or stance on this? I'm really open to anything.

I think we should definitely distinguish between invalid JSON format, failed authorization
(not in MVP) and inability to satisfy a valid request. For the latter case, I'm not sure `Conflict`
is good choice, I would propose `403 Forbidden`.


- Alexander


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/35714/#review88839
-----------------------------------------------------------


On June 22, 2015, 5:08 a.m., Michael Park wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/35714/
> -----------------------------------------------------------
> 
> (Updated June 22, 2015, 5:08 a.m.)
> 
> 
> Review request for mesos, Adam B and Jie Yu.
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> Needed in subsequent patch for /reserve master endpoint.
> 
> 
> Diffs
> -----
> 
>   3rdparty/libprocess/include/process/http.hpp e47cc7afbc8110759edf25a2dc05d09eda25c417

> 
> Diff: https://reviews.apache.org/r/35714/diff/
> 
> 
> Testing
> -------
> 
> `make check`
> 
> 
> Thanks,
> 
> Michael Park
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message