# WAMR TSC Meeting 2022/12/1

**Attendees：**
Wang, Xin (Intel)
Huang, Wenyong (Intel)
Xu, Jun1 (Intel)
He, Liang (Intel)
Kolny, Marcin (Amazon)
Takashi Yamamoto (Midokura)
huang, Qi (Xiaomi)
Yan, Dongsheng (Sony)
Tang Wei (Ant)
Trenner, Thomas (Siemens)
Ermler, Rene (Siemens)

**Self-introduction**

## Topic collection

Xin (Intel)
- Intel team 2023 focus
- performance
- GC support
- Componenet model
- Developer tooling (emcc advisor, TS2Wasm..)
- Trusted Execution Environement(TEE) like SGX/TDX use cases
- Discuss TSC meeting cadance
- start with monthly, and adjust if needed.

Marcin(Amazon):
- roadmap transparency - how do we make sure the community knows what is already planned, how can they possibly get involved etc.
- one good example I can think of is the gitbook: it'd be great to know what are the plans for it, how community can contribute, is there opportunity for brainstorming etc.
- workstream for removing singleton/global objects and introduce a concept of contexts so it's possible to embed multiple VM instances into the process
- Debugging AOT is something we at Amazon will be probably looking into Q1 next year

Wenyong (Intel):
- When to release the next version, what is the version number, whether to merge Multi-tier JIT in this version or to merge it in next version? For example, release 1.1.2 and then merge Multi-tier JIT, or merge Multi-tier JIT and release 1.2.0 directly?
- Can each company share its roadmap? e.g. Source debugger enhancements, integration of perf tool and so on..

Dongsheng (Sony):
- introduce who we are(in 20221118_newsweek.pdf), and present something(in SSS-Agenda-for-WAMR-TSC.pdf) for our concerns and roadmap.
- Marcin: the rencent WASI talk about accessing HW.

Yamt (Midokura):
- how do you evaluate startup time for fast-jit? any tools available?
- Normally we calculate it from the time iwasm started to the time the first wasm function executed: print the time at the beginning of iwasm main function, and print the time at the beginning of the wasm app's main function

Thomas and Rene (Siemens):
- Plan to implement exception feature.

Thomas (Siemens) and Jun (Intel):
- Debugging support for multiple threads
- runtime launch multiple wasm app instance in multiple threads, and launch multiple lldb to debug each of the wasm instance ---- this is OK
- runtime launch one wasm app instance, this app use pthread API to spawn several child-threads, and launch one lldb to debug the multithread wasm app ---- this will not work for now
- The main problem is that current lldb doesn't hold thread id for every spawned instance, it will always use the same thread id to get stack frame information, so the debugger will display wrong data