Closed KennethAdamMiller closed 9 years ago
It looks like we needed to replace just a couple of functions to their tail-recursive versions. So I just did that. Please see my latest change. I like it better because it doesn't require adding another dependency and changing the whole codebase to use its conventions. Let me know if this works for you.
Made another change replacing remaining and potentially useful List functions with tail-recursive implementations: https://github.com/alavrik/piqi/commit/98cd2b59c3b52b7e3e351c7bbe7d406fbf5cae2c
Yeah, it's not building with just strictly the latest commits from your head. I have to revert shortly to my own to continue testing on my ends. I see in Travis CI that your builds are working, that's great, but here's my error:
File "piqi_common.ml", line 327, characters 2-6: Parse error: [semi] expected after [str_item](in [implem])
(* surrounding code of line 327 *) let trace_indent = ref 0
let print_trace_indent () = for i = 1 to !trace_indent do prerr_string " " done ...
try running make distclean && ./configure && make deps && make && make ocaml
. it looks like the optcomp
dependency ise missing or didn't get rebuilt after upgrading the compiler/camlp4.
I think something like that happened with camlp4 because opam flipped the hell out unreasonably. I think yesterday someone pushed an incompatible version requirement for LWT and it caused all kinds of issues when pinning and unpinning CMU's BAP. In any case, I was extra perplexed when I saw the Travis CI build working fine for that code. It works now, so whatever.
On Tue, Jun 30, 2015 at 3:33 AM, Anton Lavrik notifications@github.com wrote:
try running make distclean && ./configure && make deps && make && make ocaml. it looks like the optcomp dependency ise missing or didn't get rebuilt after upgrading the compiler/camlp4.
— Reply to this email directly or view it on GitHub https://github.com/alavrik/piqi/pull/51#issuecomment-117037700.
So, I didn't have time to make sure and remove Core in favor of Core_kernel-that's the only drawback. But I was relentless in chasing down this bug. Holy crap did this take a long time. I'm pretty sure that there is literally no instance of pervasives' List.map. I think I'll have to make a later pass for memory leaking, but for now, it's not blowing up due to stack exhaustion.