profile
viewpoint
Pierre Chambart chambart OCamlPro & Nomadic Paris, France I do OCaml stuff

chambart/ocaml-bytes 9

Bytes module for ocaml <= 4.02

chambart/ocaml 5

GitHub mirror of the OCaml Subversion repository.

chambart/ocaml-1 3

Read-only mirror of INRIA SVN

chambart/ocaml-cstruct 3

Map OCaml arrays onto C-like structs

chambart/mlbdd 1

A not-quite-so-simple Binary Decision Diagrams implementation for OCaml

chambart/bitoduc.fr 0

A website about french words for computer concepts.

chambart/camlp4 0

Camlp4 tool

chambart/coq 0

Coq is a formal proof management system. It provides a formal language to write mathematical definitions, executable algorithms and theorems together with an environment for semi-interactive development of machine-checked proofs.

PullRequestReviewEvent

push eventOCamlPro/ocplib-endian

Pierre Chambart

commit sha 90439ca345230422233f687ab0532e1c9432f440

Add licence to opam file Also remove spurious src address

view details

push time in a month

pull request commentocaml-flambda/flambda-backend

Cause a C warning when CAMLreturn is missing in C stubs.

I didn't try, but is it missing something similar in CAMLnoreturn ?

stedolan

comment created time in a month

pull request commentocaml/ocaml

Add value kinds on blocks

Is there a mother-analysis that could be implemented once, and then partially-erased to derive simpler approximations?

If it is possible for other analyses, I'm all for it. We don't need something much more complex than that for a first version, but if is there is something to expose the shape of all kinds of value, we can use that.

As to which fuel/depth value to use as @lthls said, it is a trade-off between cost and benefit. We don't expect many situations where more than a depth 2 would be useful, so it might not be worth doubling the cost for those rare cases. While I did notice some normal examples where 2 was required. So this is very heuristic. If the representation was shared in memory between multiple annotations, it would be ok to build something really deeper. Otherwise, for most situations depth 1 would be sufficient with also an annotation on the getfield, but that would not be always sufficient. We might later want to increase that limit for some contexes. In particular to generate optimized versions of polymorphic comparisons.

The general type traversal is not going to be doable soon, so this is not an option right now. What would you suggest changing on that code for it to be ok ?

chambart

comment created time in a month

push eventchambart/opam-repository

Pierre Boutillier

commit sha 1603496d77031ef805d693ca5e0050cf0253994a

+tezos-node.10.2

view details

Pierre Boutillier

commit sha eec2975495fe43b88f1e1ad414d4b4188b8a09da

Tezos generic client packages v10.2 +tezos-client-base.10.2 +tezos-signer-services.10.2 +tezos-rpc-http-client.10.2 +tezos-rpc-http-client-unix.10.2 +tezos-signer-backends.10.2 +tezos-client-commands.10.2

view details

Pierre Boutillier

commit sha 0a124d9fd0e5a5de1dbf433edc8877cb39403e78

Tezos mockup client v10.2 +tezos-mockup-registration.10.2 +tezos-mockup-proxy.10.2 +tezos-mockup.10.2 +tezos-mockup-commands.10.2

view details

Pierre Boutillier

commit sha 8666727d5321d6f9611acc885b105cab6da284c2

+tezos-proxy.10.2

view details

Pierre Boutillier

commit sha 15d8f718512496ef8c8725656b0f39e14f09fd07

+tezos-client-base-unix.10.2

view details

Pierre Boutillier

commit sha 142a231707bf5d2ba8d841dba0aa397205ee515e

+tezos-signer.10.2

view details

Pierre Boutillier

commit sha f3846c725c3e2a4448e4977ff6440179641e1185

Tezos Mainnet protocols' parameters packages v10.2 +tezos-protocol-006-PsCARTHA-parameters.10.2 +tezos-protocol-008-PtEdo2Zk-parameters.10.2 +tezos-protocol-009-PsFLoren-parameters.10.2 +tezos-protocol-010-PtGRANAD-parameters.10.2

view details

Pierre Boutillier

commit sha 96b343dfb01899cfa63ffe6ed94c2417cb856447

+tezos-protocol-alpha-parameters.10.2

view details

Pierre Boutillier

commit sha cd36ccd5e290271abdf356ac6e93528ba1218d18

Tezos client mainnet protocols' packages v10.2 +tezos-client-000-Ps9mPmXa.10.2 +tezos-client-001-PtCJ7pwo.10.2 +tezos-client-002-PsYLVpVv.10.2 +tezos-client-003-PsddFKi3.10.2 +tezos-client-004-Pt24m4xi.10.2 +tezos-client-005-PsBabyM1.10.2 +tezos-client-006-PsCARTHA.10.2 +tezos-client-007-PsDELPH1.10.2 +tezos-client-008-PtEdo2Zk.10.2 +tezos-client-009-PsFLoren.10.2 +tezos-client-010-PtGRANAD.10.2

view details

Pierre Boutillier

commit sha e142d1a34229e580d4c2c88ca27e244c8482fcb6

+tezos-client-genesis-carthagenet.10.2

view details

Pierre Boutillier

commit sha 9af3a7a27e7d007d9954f37a064793d0be749b58

Tezos client dev protocols' packages v10.2 +tezos-client-alpha.10.2 +tezos-client-demo-counter.10.2 +tezos-client-genesis.10.2

view details

Pierre Boutillier

commit sha ebe5b3a98a9a9ef8c11acfe5d5eb14c0706170e2

Tezos client commands Mainnet protocols' packages v10.2 +tezos-client-001-PtCJ7pwo-commands.10.2 +tezos-client-002-PsYLVpVv-commands.10.2 +tezos-client-003-PsddFKi3-commands.10.2 +tezos-client-004-Pt24m4xi-commands.10.2 +tezos-client-005-PsBabyM1-commands.10.2 +tezos-client-006-PsCARTHA-commands.10.2 +tezos-client-007-PsDELPH1-commands.10.2 +tezos-client-008-PtEdo2Zk-commands.10.2 +tezos-client-009-PsFLoren-commands.10.2 +tezos-client-010-PtGRANAD-commands.10.2

view details

Pierre Boutillier

commit sha b92964f98a0d775c86e18b788a5cbed905799221

Tezos client sapling primitives packages v10.2 +tezos-client-sapling-008-PtEdo2Zk.10.2 +tezos-client-sapling-009-PsFLoren.10.2 +tezos-client-sapling-010-PtGRANAD.10.2

view details

Pierre Boutillier

commit sha 4c2381fc7381ed6ae6ee1d921244cb3a191d2420

Tezos client command registerers Mainnet protocols' packages v10.2 +tezos-client-007-PsDELPH1-commands-registration.10.2 +tezos-client-008-PtEdo2Zk-commands-registration.10.2 +tezos-client-009-PsFLoren-commands-registration.10.2 +tezos-client-010-PtGRANAD-commands-registration.10.2

view details

Pierre Boutillier

commit sha f4b2c464a7add289d5d0313a423ba1431029be8c

Tezos client dev protocols' packages v10.2 +tezos-client-alpha-commands.10.2 +tezos-client-sapling-alpha.10.2 +tezos-client-alpha-commands-registration.10.2

view details

Pierre Boutillier

commit sha 4014c69417e0cb8f299867fcb164d600f81892f9

Tezos consensus contribution primitives Mainnet protocols' packages v10.2 +tezos-baking-008-PtEdo2Zk.10.2 +tezos-baking-009-PsFLoren.10.2 +tezos-baking-010-PtGRANAD.10.2

view details

Pierre Boutillier

commit sha b17c68d17e429f94e8d4dfca004663b1b87b9c17

Tezos consensus contribution commands Mainnet protocols' packages v10.2 +tezos-baking-008-PtEdo2Zk-commands.10.2 +tezos-baking-009-PsFLoren-commands.10.2 +tezos-baking-010-PtGRANAD-commands.10.2

view details

Pierre Boutillier

commit sha 53db6c4755f086459f753d513682018ae21ecf00

+tezos-baking-alpha.10.2

view details

Pierre Boutillier

commit sha e44dfd7fd1129f2555cdeb329f030ab3b1b868ac

+tezos-baking-alpha-commands.10.2

view details

Pierre Boutillier

commit sha 90673aa47fccd0b8991b6ffec15b8f3fa96332ce

Tezos consensus contribution binaries Mainnet v10.2 +tezos-accuser-008-PtEdo2Zk.10.2 +tezos-accuser-009-PsFLoren.10.2 +tezos-accuser-010-PtGRANAD.10.2 +tezos-baker-008-PtEdo2Zk.10.2 +tezos-baker-009-PsFLoren.10.2 +tezos-baker-010-PtGRANAD.10.2 +tezos-endorser-008-PtEdo2Zk.10.2 +tezos-endorser-009-PsFLoren.10.2 +tezos-endorser-010-PtGRANAD.10.2

view details

push time in a month

push eventchambart/opam-repository

Pierre Chambart

commit sha bc8975596826bd8f3911fe779b521e2ea544956f

Package ocplib-endian.1.2

view details

push time in a month

push eventchambart/opam-repository

Ryan M. Moore

commit sha e870dfb7d571e5cb8aad3492aec109640f3ef594

Package pyml_bindgen.0.1.1

view details

Thomas Leonard

commit sha f9c923555ce1334844b4136c500f1bf6382bda45

[new release] capnp-rpc, capnp-rpc-unix, capnp-rpc-net, capnp-rpc-mirage and capnp-rpc-lwt (1.2.1) CHANGES: - Fix Fmt deprecation warnings (@tmcgilchrist mirage/capnp-rpc#239). - Update to latest X509 and DNS APIs (@tmcgilchrist mirage/capnp-rpc#240).

view details

Thierry Martinez

commit sha 697b0986eeb181e0266fb8cb5e7ef871e7e2d634

[new release] redirect.0.2.1 This PR adds the new release for redirect.0.2.1: - Fix `with_channel_from_string` for long strings that do not fit in pipe, by using a thread for outputing the string. - Some minor optimizations, using the fact that re-raising an exception preserves the backtrace and the constraint on OCaml version (>=4.03) ensures that match-with-exception is available.

view details

Kate

commit sha 8d1225bcf57b2d7be8d485f948e71d443fc587c3

Merge pull request #19965 from thierry-martinez/redirect.0.2.1 [new release] redirect.0.2.1

view details

Kate

commit sha d814acf43af6b577737a332637d2b36a3d51be60

Merge pull request #19954 from talex5/release-capnp-rpc-v1.2.1 [new release] capnp-rpc, capnp-rpc-unix, capnp-rpc-net, capnp-rpc-mirage and capnp-rpc-lwt (1.2.1)

view details

Nate Foster

commit sha 86395b95f636294f4018d47c1bdd9125b93e3035

Package frenetic.5.0.5

view details

Kate

commit sha c57448965ab9998ef5bbfe1149fb0d2691de8586

Merge pull request #19561 from jnfoster/opam-publish-frenetic.5.0.5 Package frenetic.5.0.5

view details

Kate

commit sha 7b361ce572d98a700d2e0018cb6e9c72db9a3675

Merge pull request #19949 from mooreryan/opam-publish-pyml_bindgen.0.1.1 Package pyml_bindgen.0.1.1

view details

Pierre Chambart

commit sha acbb25ac6c745cadfa2f88d9acd4e956199a21db

Package ocplib-endian.1.2

view details

push time in a month

created tagOCamlPro/ocplib-endian

tag1.2

Optimised functions to read and write int16/32/64 from strings, bytes and bigarrays, based on primitives added in version 4.01.

created time in a month

PullRequestReviewEvent

Pull request review commentocaml/opam-repository

Package ocplib-endian.1.2

+opam-version: "2.0"+synopsis: "Optimised functions to read and write int16/32/64 from strings and bigarrays"+description: """+The library implements three modules:+* [EndianString](https://github.com/OCamlPro/ocplib-endian/blob/master/src/endianString.mli) works directly on strings, and provides submodules BigEndian and LittleEndian, with their unsafe counter-parts;+* [EndianBytes](https://github.com/OCamlPro/ocplib-endian/blob/master/src/endianBytes.mli) works directly on bytes, and provides submodules BigEndian and LittleEndian, with their unsafe counter-parts;+* [EndianBigstring](https://github.com/OCamlPro/ocplib-endian/blob/master/src/endianBigstring.mli) works on bigstrings (Bigarrays of chars), and provides submodules BigEndian and LittleEndian, with their unsafe counter-parts.+"""+maintainer: "pierre.chambart@ocamlpro.com"+authors: "Pierre Chambart"+homepage: "https://github.com/OCamlPro/ocplib-endian"+bug-reports: "https://github.com/OCamlPro/ocplib-endian/issues"+doc: "https://ocamlpro.github.io/ocplib-endian/ocplib-endian/"+license: "LGPL-2.1-or-later WITH OCaml-LGPL-linking-exception"+depends: [+  "base-bytes"+  "ocaml" {>= "4.03.0"}+  "cppo" {>= "1.1.0" & build}+  "dune" {>= "1.0"}+]+build: [+  ["dune" "build" "-p" name "-j" jobs+   "@install"+   "@runtest" {with-test}+   "@doc" {with-doc}]+]+dev-repo: "git+https://github.com/OCamlPro/ocplib-endian.git"+url {+  src: "https://github.com/OCamlPro/ocplib-endian/archive/1.1.tar.gz"

Good catch. I didn't even look at it, it was produced by opam-publish so I had supposed that it would be ok. I probably forgot to push the tag. opam-publish took the version from the local git for some things and from the remote for the archive ! @AltGr Is that a bug ?

chambart

comment created time in a month

create barnchchambart/opam-repository

branch : opam-publish-ocplib-endian.1.2

created branch time in a month

PR opened ocaml/opam-repository

Package ocplib-endian.1.2

ocplib-endian.1.2

Optimised functions to read and write int16/32/64 from strings and bigarrays The library implements three modules:

  • EndianString works directly on strings, and provides submodules BigEndian and LittleEndian, with their unsafe counter-parts;
  • EndianBytes works directly on bytes, and provides submodules BigEndian and LittleEndian, with their unsafe counter-parts;
  • EndianBigstring works on bigstrings (Bigarrays of chars), and provides submodules BigEndian and LittleEndian, with their unsafe counter-parts.

  • Homepage: https://github.com/OCamlPro/ocplib-endian
  • Source repo: git+https://github.com/OCamlPro/ocplib-endian.git
  • Bug tracker: https://github.com/OCamlPro/ocplib-endian/issues

:camel: Pull-request generated by opam-publish v2.1.0

+33 -0

0 comment

1 changed file

pr created time in a month

push eventOCamlPro/ocplib-endian

Pierre Chambart

commit sha a060c60e459f2cbd889a8d20375b113d8970cc16

Annotate all relevant functions with [@inline]

view details

Pierre Chambart

commit sha db020c52d63ae1d40f982d0873fbf15940f9071e

Deprecate EndianString.set * functions

view details

Pierre Chambart

commit sha ab77bc18eaaffe0f883222284f9b56bcc6b89dc1

Remove build mark on dune dependency in .opam file This is in general incorrect as dune generate some files

view details

Pierre Chambart

commit sha 10292cd3ffa4d23d737e3f855ad04f22d3d95460

Update CHANGES.md

view details

push time in a month

push eventOCamlPro/ocplib-endian

Hugo Heuzard

commit sha 602f20fd8ec6bc8fecebe526373cf142c7819abb

fix the implementation for js_of_ocaml

view details

push time in a month

PR merged OCamlPro/ocplib-endian

Fix the implementation for js_of_ocaml

Ints are 32bit in jsoo (not 31bit/63bit).

Use Sys.int_size (introduced in OCaml 4.03) instead of making assumption on int size.

+8 -8

2 comments

5 changed files

hhugo

pr closed time in a month

PullRequestReviewEvent

pull request commentOCamlPro/ocplib-endian

Fix the implementation for js_of_ocaml

Thanks the fix is sensible. I'll release soon

hhugo

comment created time in a month

push eventchambart/flambda-backend

Pierre Chambart

commit sha 4f6137f6ed2f0ae247fc95996ef705f263575671

Fix max_rec_depth off by one error

view details

Pierre Chambart

commit sha ef2170103e9e2825f1fde14d4b1582e3456f5e1c

Add an inlining decisions case for recursive

view details

Pierre Chambart

commit sha 92e4c83f2aa7e1cc609069af60b4f43f294294d2

Partially honnor the unroll attribute

view details

Mark Shinwell

commit sha fce90a02a8bd2c170df6a0642cba1fe194acb477

Expose max_rec_depth as a command-line option

view details

Pierre Chambart

commit sha ce8539d7a61adc01ae73d63db1c2ff7f62e92f80

Default max_rec_depth is 0

view details

push time in a month

push eventchambart/flambda-backend

Pierre Chambart

commit sha df2f7ee3a49803ead8d09bddab119c562f76b8f5

Actually use the command line argument

view details

push time in a month

create barnchchambart/flambda-backend

branch : cannot_inline_rec_functions

created branch time in a month

push eventchambart/ocaml-1

Pierre Chambart

commit sha cb03daabb84a6a4deb6bb5a9954db9796c0f03f5

Update .depend

view details

push time in a month

pull request commentocaml-flambda/flambda-backend

Proper Patricia tree map function

Also replace uses of map that can benefit from map_sharing

chambart

comment created time in a month

push eventchambart/flambda-backend

Pierre Chambart

commit sha a3e9387b5d888cfb2c16e0256a62f7ee1601ddc3

Use map_sharing when possible

view details

push time in a month

push eventchambart/ocaml-1

Pierre Chambart

commit sha 50979e2d1b9d2da4ffb1414b7d84079f3f3f902b

Remove the visited limit. No branch are cut by it that wouldn't be by the fuel limit

view details

push time in a month

pull request commentocaml-flambda/flambda-backend

Implement Patricia_tree.map_sharing properly

By the way I tried replacing map_sharing by assert false and it does not trigger while building the compiler. This is not the one used. Row_like uses Tag.Map which is from identifiable (but would probably benefit from using patricia trees instead) and Closure_id.Map which is not an integer.

lukemaurer

comment created time in a month

push eventchambart/flambda-backend

Pierre Chambart

commit sha df2d09f187deddd5331b9995a289afaa8f01a1db

Also do patricia mapi

view details

push time in a month

push eventchambart/ocaml-1

Pierre Chambart

commit sha 086e2756caa9475f98986e70444ad025c8024518

Fix recursion detection for value_kind

view details

Pierre Chambart

commit sha dfb181db7a258eff2fff18cd4ded5676cc4ea4ca

Remove the visited limit. No branch are cut by it that wouldn't be by the fuel limit

view details

push time in a month

pull request commentocaml-flambda/flambda-backend

Implement Patricia_tree.map_sharing properly

I didn't bench it, but I would suppose that we could win quite a lot from improving union

lukemaurer

comment created time in a month

PR opened ocaml-flambda/flambda-backend

Proper Patricia tree map function

I think that it is probably better to do it that way than #369

+21 -9

0 comment

1 changed file

pr created time in a month

more