[cwn] Attn: Development Editor, Latest OCaml Weekly News
Alan Schmitt
alan.schmitt at polytechnique.org
Mon Jul 9 23:49:20 PDT 2018
━━━━━━━━━━━━━━━━━━━
OCAML WEEKLY NEWS
━━━━━━━━━━━━━━━━━━━
Hello
Here is the latest OCaml Weekly News, for the week of July 03 to 10,
2018.
For people receiving this by email, I'm trying to send it as a
multipart/alternative email with the text and html version. Please let
me know if there is an issue.
Table of Contents
─────────────────
Sightings of OCaml around the Web
Looking for OCaml developers for a Tezos based Decentralised Betting Exchange!
Dune 1.0.0 is coming soon, what about jbuilder projects?
New release of craml, a tool to test command line applications
How to read/understand function signature with GADT?
Format module from the Standard Library
First release of mastodon-archive-viewer (0.1)
Other OCaml News
Old CWN
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
Sightings of OCaml around the Web
═════════════════════════════════
Archive:
[https://discuss.ocaml.org/t/sightings-of-ocaml-around-the-web/1867/23]
Yotam Barnoy announced
──────────────────────
Here's a must-read article on the dev team's [Plans for OCaml 4.08] by
@lpw25 . If you're a long-time user, you'll really enjoy this.
Also, I'll just mention that [OCamlverse] has had quite a few updates
recently. If you haven't checked it out yet, please do – everyone can
request and receive automatic contribution rights, so if there's
something you want to fix up/contribute to, you're welcome to do so.
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
[Plans for OCaml 4.08] https://blog.janestreet.com/plans-for-ocaml-408/
[OCamlverse] https://ocamlverse.github.io/
Looking for OCaml developers for a Tezos based Decentralised Betting Exchange!
══════════════════════════════════════════════════════════════════════════════
Archive:
[https://discuss.ocaml.org/t/looking-for-ocaml-developers-for-a-tezos-based-decentralised-betting-exchange/2236/1]
Suraj Tirupati announced
────────────────────────
My team and I are working on a decentralised betting exchange called
BettingWin - [https://www.bettingwin.online]
We want to create a platform through which users can place bets in a
decentralised environment; bettors bet, market makers create markets
and reporters report events. It's still in development and we are
looking to work with some new developers.
We are currently looking to expand our team and bring on some
developers to complete paid tasks or become a permanent member of the
team depending on our compatibility :) We are really interested in
working with smart, insightful developers who are familiar with OCaml
and have experience with programming smart contracts. Also, any
developers with an interest in Tezos are very welcome as that is the
blockchain we are using.
We can arrange Skype calls to discuss things in detail. Please contact
me at suraj at bettingwin.online
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
Dune 1.0.0 is coming soon, what about jbuilder projects?
════════════════════════════════════════════════════════
Archive:
[https://discuss.ocaml.org/t/dune-1-0-0-is-coming-soon-what-about-jbuilder-projects/2237/1]
Jérémie Dimino announced
────────────────────────
The 1.0.0 release of dune is planned for next week. It will be the
first release of jbuilder as dune, so what to expect for existing
jbuilder projects?
The good news is that dune is fully backward compatible with
jbuilder. The `dune' package in opam will install both `dune' and
`jbuilder' binaries and both will be usable on existing jbuilder
projects.
We do want to stop supporting jbuilder projects eventually, so the
plan is to continue to support the `jbuilder' binary and `jbuild'
files for another year. 6 months from now dune will start displaying
warnings when using the `jbuilder' binary or `jbuild' files. A more
detailied [migration plan][mp] is available in the manual. In
particular it contains a precise list of things to do to migrate a
jbuilder project to a dune project.
We hope that the new features in Dune 1.0.0 and subsequent releases
will be enough of an incentive for users to eagerly switch to Dune :)
Important changes in Dune
╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
This section is not an exhaustive list of changes, but it describes
the important changes to expect when switching from `jbuilder' to
`dune'.
◊ Versioning and configuration files
A big part of the work in Dune 1.0.0 was getting the versioning story
right. With jbuilder it was hard for us to make the language evolve
without breaking the build of existing projects. Dune can handle that
gracefully and in the future it will be much easier to introduce
breaking changes.
The versioning of all dune files in a project is controlled via a
single `dune-project' file at the root of the project. This file
contains the version of the dune language and extensions used in the
project.
While this file is required in order to ensure compatibility with
future versions of Dune, it is not required to write one by hand every
time you start a new project or want to do a quick experiment. Dune
will create and edit this file for you as needed, so simply start by
creating a `dune' file (the new `jbuild' files) and dune will take
care of the rest.
Note that the syntax inside `dune', `dune-project', … files is
slightly different from the one inside `jbuild' files. In particular
the language requires fewer parentheses and the syntax of variables
changed from `${...}' to `%{...}' to avoid issues with shell
commands. This is detailed in the [migration plan][mp] as well.
◊ Build profiles and default settings
Up to now `jbuilder' supported a `--dev' option in order to enable
stricter flags. There have been several changes regarding this. First
of all you are now able to define as many build profiles as you want:
`dev', `release', `perf', … These can be selected either via the
`--profile' command line option or via the `dune-workspace' file. You
are also able to define what the default compilation flags mean for
each profile via an [env stanza][env].
Another important change is that the default build profile for the
`dune' binary is `dev' rather than `release'. We found that most of
the time developpers want the development mode by default. Please
remember to add `"-p" name' to all invocations of `dune' in opam
files! This is now even more important than before in order to ensure
the release profile is selected for opam builds.
◊ Default target
Dune allows to configure what happens when you run `dune build'
without any specific targets. This is done via the introduction of a
[default alias][defal].
◊ More parellelism by default
If you had `(jobs N)' in you `~/.config/dune/config' file, you can
remove it. Dune now automatically sets the number of jobs that can be
run in parallel to the number of processing units available. The
previous default was 4. For opam files, you should remember to add
`"-j" jobs' in order to let opam decide.
Future of Dune
╌╌╌╌╌╌╌╌╌╌╌╌╌╌
Now that we have a good versioning story and with the new testing
tools that are emerging, I'm hoping that at some point we will be able
to declare that the master of Dune is stable. At this point opam
releases will just become a monthly or even weekly formality. This
should make the experience nicer for both developers and users of
Dune.
[mp]: [https://dune.readthedocs.io/en/latest/migration.html]
[env]: [https://dune.readthedocs.io/en/latest/dune-files.html#env]
[defal]:
[https://dune.readthedocs.io/en/latest/usage.html#default-alias]
Jérémie Dimino then added
─────────────────────────
One thing I forgot to mention: if you have a `~/.config/dune/config'
file, you need to add the following line at the beginning of this
file: `(lang dune 1.0)'. This line is the one we use in every
configuration file that is versionned and versioning the
`~/.config/dune/config' file seemed natural. The `jbuilder' binary
doesn't require it for backward compatibility reasons, but the `dune'
binary does.
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
New release of craml, a tool to test command line applications
══════════════════════════════════════════════════════════════
Archive:
[https://discuss.ocaml.org/t/ann-new-release-of-craml-a-tool-to-test-command-line-applications/2241/1]
Thomas Gazagnaire announced
───────────────────────────
I am happy to announce the first release of [craml], a toool for
testing command-line applications.
To install it:
┌────
│ opam install craml
└────
The tool is already used by [real-world OCaml v2] to test shell script
examples and [by merlin]. Feedback is alway welcome, so feel free to
open new issues on the tracker to discuss about new features.
Here is a copy of the README, with a more complete description of the
project:
—
[craml] https://github.com/realworldocaml/craml
[real-world OCaml v2]
https://github.com/realworldocaml/book/blob/master/examples/code/json/build_github_atd.sh
[by merlin]
https://github.com/ocaml/merlin/blob/master/tests/type-expr/test.t
craml: a tool for testing command-line applications
╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
CRAM is a is functional testing framework for command line
applications. `craml' is freely inspired by the [Python tool], which
was itself based on Mercurial's [unified test format].
`craml' is released as a single binary (called `craml') and supports
the following syntax:
• Lines beginning with two spaces, a dollar sign, and a space are
*commands* and will be run in the shell.
• Multi-lines commands end by `\' and continue with two spaces and a
`>' sign on the next line:
┌────
│ $ <line1> \
│ > <line2> \
│ > <line3>
└────
• Lines beginning with two spaces are considered command *output*.
• Command outputs can contains *ellipsis*: `...'. These will match any
possible outputs (on zero, one or multiple lines).
• Lines starting by `<--' are command *pre-conditions*; they will
determine the conditions where the command is run. Currently, only
non-deterministic modes are supported as pre-conditions (see below).
• Lines starting by `-->' are command *post-conditions*. Currently,
only exit codes are supported as post-conditions (see below).
• Anything else is a comment. It is not possible to put comments in
the middle of an output as it is not clear what should be done to
them when the output changes.
To run the tests described in a `<file>', use `craml <file>'. This
will run all the commands in sequence and will generated
`<file>.corrected' if one of the output do not match the expected
command outputs.
[Python tool] https://bitheap.org/cram/
[unified test format] https://www.selenic.com/blog/?p=663
◊ Non-deterministic Outputs
`craml' supports non-deterministic outputs:
┌────
│ <-- non-deterministic
│ $ <command>
│ <output>
└────
In that case, `craml <file>' will run the command but will not
generate `<file>.corrected' if the new output differs from the one
described in the file. Use `craml --non-deterministic <file>' to come
back to the default behaviour.
◊ Non-deterministic Commands
`craml' supports non-deterministic outputs:
┌────
│ <-- non-deterministic [skip]
│ $ <command>
│ <output>
└────
In that case, `craml <file>' will *not* run the command. Use `craml
--non-deterministic <file>' to come back to the default behaviour.
◊ Exit Codes
`craml' tests exit codes:
┌────
│ $ <command>
│ <output>
│ --> exit 10
└────
If `<command>' does not exit with code 10, then `craml <file>' will
generate `<file>.corrected' with the right exit code. Note that `@@
exit 0' will not be displayed.
Perry E. Metzger said and Daniel Bünzli replied
───────────────────────────────────────────────
> Just a small request: I note that the README, which many users will
never see on installing from opam,
Also note that an `odig readme craml' will do.
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
How to read/understand function signature with GADT?
════════════════════════════════════════════════════
Archive:
[https://discuss.ocaml.org/t/how-to-read-understand-function-signature-with-gadt/2250/1]
Volodymyr Melnyk asked
──────────────────────
For example, given the code:
┌────
│ let rec eval : type a . a term -> a = ...
└────
How to read/understand this part: `type a . a term -> a'? What does
the dot here mean? Why should we define type `a' in the function's
signature?
octachron
─────────
There is two things going on simultaneously here. The first is that
GADTs can only add equation to abstract types in a pattern
matching. For instance, let's consider the toy example
┌────
│ type a = Tag_a
│ type b = Tag_b
│ type _ t =
│ | A: a t
│ | B: b t
└────
then
┌────
│ let to_string = function
│ | A -> "a"
│ | B -> "b"
└────
yields a type error
┌────
│ > Error: This pattern matches values of type b t
│ but a pattern was expected which matches values of type a t
│ Type b is not compatible with type a
└────
because the type of the argument was inferred to `a t' in the first
branch of the pattern matching.
The solution is to make the type phantom parameter of `t' a locally
abstract type in `to_string'.
┌────
│ let to_string (type t) (x: t) = match x with
│ | A -> "a"
│ | B -> "b"
└────
The notation `(type t)' introduces a local abstract type named `t',
since the type is abstract, it can be refined to different types in
the branches of the pattern matching.
The second point is that recursive function involving GADTs often
needs to be recursive polymorphic. For instance, with a toy langage,
┌────
│ type _ term =
│ | Bool : bool -> bool term
│ | Int : int -> int term
│ | If : bool term * 'a term * 'a term -> 'a term
└────
an eval function would look like
┌────
│ let rec eval (type t) (x:t term): t = match x with
│ | Bool b -> b
│ | Int x -> x
│ | If (cond,then',else') -> if eval cond then eval then' else eval else'
└────
but in the `If' branch `eval' is called with an argument of a
different types, which result in a type error
┌────
│ > Error: This expression has type t term but an expression was expected of type
│ bool term
└────
on the first call with a different value for the `term' type
parameter. The long-winded solution here is to add an explicit
polymorphic annotation and then introduce locally abstract types
┌────
│ let rec eval: 't. 't term -> 't = fun (type t) (x: t term): t -> match x with
│ | Bool b -> b
│ | Int x -> x
│ | If (cond,then',else') -> if eval cond then eval then' else eval else'
└────
Since this is a bit of a mouthful, there is shorter variant that
directly combines explicit polymorphic annotation with locally
abstract types:
┌────
│ let rec eval: type t. t term -> t = function
│ | Bool b -> b
│ | Int x -> x
│ | If (cond,then',else') -> if eval cond then eval then' else eval else'
└────
Here `type t. … ~ declares ~t' as both a locally abstract types and
makes it explicitly polymorphic.
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
Format module from the Standard Library
═══════════════════════════════════════
Archive:
[https://discuss.ocaml.org/t/format-module-from-the-standard-library/2254/1]
cloudyhug asked
───────────────
Hello everyone. Does any of you know where I can find a decent
tutorial for the OCaml Format module ? I cannot figure out how to use
it, and the few things I have found are the documentation and the
official tutorial, which are explicit but not clear at all. I have
spent the last 30 minutes trying to understand it, in vain.
Gabriel Radanne replied
───────────────────────
I would recommend reading this blog post:
[https://cedeela.fr/format-all-the-data-structures.html]
For understanding boxes, see
[https://ocaml.org/learn/tutorials/format.html]
Martin Jambon also replied
──────────────────────────
If that's an option for you, I'd suggest using [easy-format]. There's
a [complete example here] with various ways to format the same thing.
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
[easy-format] https://github.com/mjambon/easy-format
[complete example here]
https://mjambon.github.io/mjambon2016/easy_format_example.html
First release of mastodon-archive-viewer (0.1)
══════════════════════════════════════════════
Archive:
[https://discuss.ocaml.org/t/ann-first-release-of-mastodon-archive-viewer-0-1/2257/1]
Kate announced
──────────────
[Mastodon] is a decentralized social network similar to Twitter.
One of its many features is to allow people to get their archive as a
way to save their posts history in case one server is seized, crashed,
or temporary down. It is also useful in case you want to delete an old
account but still want to keep its archive.
Here I present you [mastodon-archive-viewer] which is a tool that
takes a mastodon archive, previously exported, downloaded and
extracted and gives you an offline web view of it.
If you want to give it a try, mastodon-archive-viewer is now available
on opam, and comes with a filter system to view only parts of your
archive.
┌────
│ opam install mastodon-archive-viewer
│ cd <archive-directory-previously-extracted>
│ mastodon-archive-viewer outbox.json > archive.html
└────
For the moment this tool is in an early stage of development so don't
hesitate to report bugs or feature requires on the issues tracker.
Happy tooting!
PS: if you are interested in using Mastodon, perhaps a legit choice of
instance might be [functional.cafe], for example :slight_smile:
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
[Mastodon] https://joinmastodon.org/
[mastodon-archive-viewer]
https://github.com/kit-ty-kate/mastodon-archive-viewer
[functional.cafe] https://functional.cafe
Other OCaml News
════════════════
From the ocamlcore planet blog
──────────────────────────────
Here are links from many OCaml blogs aggregated at [OCaml Planet].
• [Coq 8.8.1 is out]
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
[OCaml Planet] http://ocaml.org/community/planet/
[Coq 8.8.1 is out] https://coq.inria.fr/news/146.html
Old CWN
═══════
If you happen to miss a CWN, you can [send me a message] and I'll mail
it to you, or go take a look at [the archive] or the [RSS feed of the
archives].
If you also wish to receive it every week by mail, you may subscribe
[online].
――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――――
[Alan Schmitt]
[send me a message] mailto:alan.schmitt at polytechnique.org
[the archive] http://alan.petitepomme.net/cwn/
[RSS feed of the archives] http://alan.petitepomme.net/cwn/cwn.rss
[online] http://lists.idyll.org/listinfo/caml-news-weekly/
[Alan Schmitt] http://alan.petitepomme.net/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.idyll.org/pipermail/caml-news-weekly/attachments/20180710/2dbc2c85/attachment-0001.htm>
-------------- next part --------------
--
OpenPGP Key ID : 040D0A3B4ED2E5C7
Monthly Athmospheric CO₂, Mauna Loa Obs. 2018-06: 410.79, 2017-06: 408.84
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 528 bytes
Desc: not available
URL: <http://lists.idyll.org/pipermail/caml-news-weekly/attachments/20180710/2dbc2c85/attachment-0001.pgp>
More information about the caml-news-weekly
mailing list