Alain Frisch

The problem is actually that ppx_sedlex.ml applies super#expr instead of this#expr to immediate sub-expressions (actions).

I don't think there is a need to change the OCaml lexer, all special characters not interpreted verbatim in strings literal are ASCII, and any occurrence of those bytes in...

The concrete syntax is the one of OCaml patterns, following their precedence. So `Star r1, r2` is `(Star r1), r2`.

Don't hesitate to suggest a specific patch or pull request. I assume that this is _not_ just about moving the ppx directive to the ppx sub-package, which already exists but...

If you only want to source transformation without running the compiler, you can use the `rewriter` tool, part of this package. Otherwise, I'm not against splitting ppx_metaquot into a reusable...

Thanks for your reply. I have switched to using zlib internally (just replace the content of pdfflate.ml with `include Zlib` and not linking the correponding .c file) and confirm everything...

For future reference: the same problems exists with the vendored AES implementation (the original rijndael implementation + the OCaml stubs from https://github.com/xavierleroy/cryptokit). Whatever solution is put in place for zlib...

Thanks for looking into this! I'm unsure about the direction. I think OCaml developers would prefer if camlpdf relied on upstream camlzip (which is itself self-contained) and libzip, and the...

About the patch for malformed streams : either the patch is not needed with zlib, which would suggests that relying on official versions (camlzip+zlib) is indeed better; or it is...

Note: observed with ocamlopt, not ocamlc!