You're right, I was speaking sloppily. By "infallible" I mean construction of the AST should not need to be null-checked.
I would like to allocate a tremendous amount of virtual memory up front such that we can just crash. I take back what I said about 32bit fragmentation, this approach would fail when the address space is very fragmented. If we aspire to work even under heavily fragmented 32bit address spaces, then that kills this proposal. On Thu, Apr 21, 2016 at 5:46 PM, Jason Orendorff <[email protected]> wrote: > I want to understand your beefs before really responding... > > On Thu, Apr 21, 2016 at 4:05 PM, Shu-yu Guo <[email protected]> wrote: > >> The first is ergonomic. I want phased workloads like parsing + BCE and >> JIT compiling to be completely infallible. >> > > What do you mean by "infallible" here? To me it seems like virtually every > Parser method is inherently fallible because of SyntaxErrors. > > The allocation pattern of compilers is a series of many small allocations. >> I'd like to just allocate a huge buffer inside a LifoAlloc-like allocator >> in the beginning and call it a day. >> > > OK, but what should we do when the huge buffer runs out during parsing? > > -j > > -- shu _______________________________________________ dev-tech-js-engine-internals mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-js-engine-internals
