[tarantool-patches] Re: [PATCH v1 1/3] box: rfc for stacked diagnostic area in Tarantool

Vladislav Shpilevoy v.shpilevoy at tarantool.org
Tue Aug 6 23:50:24 MSK 2019


Hi! Thanks for the answers!

CCed tarantool-patches back.

On 06/08/2019 10:05, Kirill Shcherbatov wrote:
>> 3. VShard uses unpack: https://github.com/tarantool/vshard/blob/a6826b26f9ab38e338a34cb7dd7f46dabbc67af5/vshard/error.lua#L149
>> Is it compatible with what you are doing? Here I need to turn an error into a table to
>> set a different serialization method.
>>
>> 4. Have you considered an idea to unpack errors not as an array, but
>> as a list? When each error object has a field 'reason' pointing to
>> another object, and so on. It would allow not to change unpack() output
>> format - it would still return the newest error, but with a new field.
> 
> I designed :unpack() method to produce an output similar to backtrace.
> But maybe it is not really our priority and taking into account your (3)th comment
> we should better use reason-based list.

Please, ask other people.

>> 9. Please, describe the new binary protocol in the similar way as it
>> is done here: https://github.com/tarantool/tarantool/blob/master/doc/rfc/3328-wire_protocol.md#body
>> And add it to the docbot request with code values included.
> 
> Is I sad in other letter, designing of iproto errors transfer may not be implemented in scope of this
> patch series. Maybe it worth to remove this paragraph from RFC at all.
> 

Ok, then remove, please, or move to a section 'plans', 'alternatives', etc.




More information about the Tarantool-patches mailing list