[Tarantool-patches] [PATCH v2 0/4] Don't truncate float in update

[Vladislav Shpilevoy]

Hi!

Why did you decide not to push it into 2.2 and 1.10? This is
a bug, and it exists in these versions too.

On 20/02/2020 07:15, Kirill Yukhin wrote:
> Hello,
> 
> On 04 фев 23:53, Vladislav Shpilevoy wrote:
>> The patchset fixes float value truncation in
>> tuple/space:update/upsert. First commit fixes the actual issue,
>> but introduces another bug that a float value can be stored into a
>> field having double type in its format.
>>
>> The rest 3 commits fix this new bug.
>>
>> However, there is an alternative how these 3 could be done. My
>> solution is that I calculate the result types in
>> xrow_update_*_store() functions. Because of that I don't know the
>> types during xrow_update_*_sizeof(), and it needs to return the
>> maximal possible size instead of exact size. The alternative is to
>> calculate types during sizeof. It would allow to return the exact
>> size. But it also would complicate the code. I can implement this
>> option if it looks better.
>>
>> Branch: http://github.com/tarantool/tarantool/tree/gerold103/gh-4701-update-float-truncate
>> Issue: https://github.com/tarantool/tarantool/issues/4701
> 
> LGTM. I've checked your patchset into 2.3 and master.
> 
> --
> Regards, Kirill Yukhin
>