ArcsinX wrote:
I feel like we can't come into agreement here. So, I will shortly sum up my
position and maybe someone else can take a look at this:
- this patch is a try to address random crashes of clangd when modules are
imported within a header
- these crashes happens inside the clang (but not clangd) and relate to
precompiled headers (clangd uses preamble optimization to speedup reparsing at
file modification if `include`s section is not modified)
- the author confirms that this is clang (but not clangd) problem, but as he
said, this is not easy to fix on the clang side and we don't known when this
can be fixed because we have limited resources to support c++20 modules
functionality in the clang. So, inacceptable to him.
- this patch hides the problem instead of a real fix by skipping preamble after
it was build:
- this disables preamble optimization in clangd, leading to performance
degradation. We reparse the whole file at every file modification. This also
affects cases, when we don't have modules import within a header (these cases
had no problems before)
- I think disabling preamble optimization is still an option (because
modules import is some kind of replacement for precompiled headers), but we
should not build the preamble if we have no plans to use it. But the author
also rejects this option, because it's not so easy to implement and needs more
complex modification.
https://github.com/llvm/llvm-project/pull/187432
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
