delete hashkey and hincrby twice, first get wrong answer and next get coredump

[wanghenshui]

Is this a regression?

No

Description and minimal reproduction of the bug

pika 4.0.0
repro

hincrby h 112000005 1000
hgetall h
del h
hincrby h 112000005 1000
hgetall h # already bug here
hincrby h 112000005 1000 # coredump

source code

    Int64ToStr(value_buf, 32, value);
    BaseDataValue internal_value(value_buf);
    batch.Put(handles_[kHashesDataCF], hashes_data_key.Encode(), internal_value.Encode());

Int64ToStr modified value_buf, value_buf implict convert to Slice, maybe bug here

Please provide the version you discovered this bug in (check about page for version information)

pika 4.0.0

[QlQlqiqi]

用 unstable 分支的最新代码（4bfb5e788a1a6fd92dff20d5f1fa24e26ee3d128）测了一下，没有发现问题

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

I tested it with the latest code from the unstable branch (4bfb5e7) and found no problem.

[QlQlqiqi]

我认为这个 issue 和 #2835 类似，不知是否是同一原因

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

I think this issue is similar to #2835, I don’t know if it’s the same reason

[wanghenshui]

我认为这个 issue 和 #2835 类似，不知是否是同一原因

确实是同一个问题，建议backport

[Issues-translate-bot]

Bot detected the issue body's language is not English, translate it automatically.

---

I think this issue is similar to #2835, I wonder if it is the same reason

It's indeed the same problem, I suggest you backport it.