Sphinx 2.3.2: segfault in some queries about manticoresearch

[tomatolog]

Could you provide data that cause this crash? I need index files along with config (tokenizer and dictionary settings) or another way to get reproducible example.

[outcassed]

Unlike the creator of that issue, I don't know what query will reproduce it. Not very helpful I know!

I haven't switched to Manticore yet but when I start running it I will skip this patch and let you know.

[tomatolog]

In case it crashed without patch provided at original ticket please attach crash log from searchd and we try to figure out root of this crash

[tomatolog]

I've just looked at case #2644 and unable to reproduce issue described there with either options:

• multi-query
• subtree_*_cache options
• query to multiple indexes with one of them empty

Seems crash depends on data that is why I need the data that causes crash or case that reproduces this crash locally here.

I also can not apply the patch provided as it violates function NodeCacheContainer_t::WarmupCache contract, there is explicit assert at top of function assert ( m_pSetup ) - that means m_pSetup CAN NOT be NULL there. That is why that crash should be investigated and properly fixed not with just quick hack.

Feel free to post more info or catch up original reporter as I unable to get this info on my own.

[airolg]

The data for reproducing the crash were not provided so far, so I'm closing the issue for a while. You can always re-open it, when your data will be ready. Thank you