By default, a Failure exception is printed as such:
# print_endline (Printexc.to_string (Failure "uh\noh"));;
Failure("uh\noh")
For improved readability, we want to print the argument of Failure as is because we understand it should be human-readable. In the OCaml standard library, we would initialize an application with the following:
# Printexc.register_printer (function
| Failure s -> Some ("Failure: " ^ s)
| _ -> None
);;
The new behavior of Printexc.to_string would be:
# print_endline (Printexc.to_string (Failure "uh\noh"));;
Failure: uh
oh
Great. Now if we use the core_kernel library, first we can see that printing an exception is slightly different but not better to a human reader:
#require "core_kernel";;
# print_endline (Printexc.to_string (Failure "uh\noh"));;
(Failure "uh\
\noh")
Perhaps we can override this? Let's try.
# Printexc.register_printer (function
| Failure s -> Some ("Failure: " ^ s)
| _ -> None
);;
# print_endline (Printexc.to_string (Failure "uh\noh"));;
Failure: uh
oh
This works, but it's not using the printer that's part of Core_kernel. If we use it, we still get the same unreadable result:
# print_endline (Core_kernel.Exn.to_string (Failure "uh\noh"));;
(Failure "uh\
\noh")
And Core_kernel.Exn doesn't offer a register_printer function. So effectively, it looks like Core_kernel.Exn makes sure that we don't define custom exception printers. Is there another way or should we just not use Core_kernel.Exn then if we want to show human-readable error messages?
Edit: For context, our original problem is to print nested error messages nicely. For example, we would like to read something like the following:
Uncaught exception: Failure:
Uncaught exception in subprocess 1234: Failure:
something happened
trace line 1
trace line 2
trace line 1
trace line 2
trace line 3
where we use indentation for quoting and escaping, rather than double-quotes and backslash escape sequences.
Base.Exn (for which Core_kernel.Exn is an alias) prints errors as human readable s-expressions using Sexplib0 printer and converters. It is possible to add custom sexp converter for exceptions with Sexplib0.Exn_converter.add.
However, if you don't intend to print exceptions as s-expressions, I see indeed no reason to use the printer of Base.Exn.
EDIT: Since the problem seems to be the printing of S-expressions, one solution might be to use Base.Exn.sexp_of and then plug-in a custom S-expression printers that does not escape strings, nor print parenthesis:
let pp_sep ppf () = Format.fprintf ppf "# "
let pp_sexp ppf = function
| Atom s -> Format.pp_print_string ppf s
| List l ->
Format.fprintf ppf "#[<v 2> %a#]"
(Format.pp_print_list ~pp_sep pp_sexp) l
let to_string exn = Format.asprintf "%a" pp_sexp (Exn.sexp_of_t exn)
Related
I am trying to write a function that creates a directory but does not raise an error when the directory already exists. This is the function:
fun ensureDir s =
(OS.FileSys.mkDir s)
handle OS.SysErr (_, SOME Posix.Error.exist) => ()
I based the pattern OS.SysErr (_, SOME Posix.Error.exist) on the fact that OS.FileSys.mkDir fails with the following error message when the directory already exists:
Poly/ML:
Exception- SysErr ("File exists", SOME EEXIST) raised
SML/NJ:
uncaught exception SysErr [SysErr: File exists [exist]]
raised at: <mkdir.c>
However, I get this error when I try to define the function in the Poly/ML shell:
poly: : error: qualified name Posix.Error.exist illegal here
Static Errors
This is the error in SML/NJ:
Error: variable found where constructor is required: Posix.Error.exist
What mistake did I make in the function definition?
(Poly/ML 5.7.1; SML/NJ 110.79; Ubuntu 20.04)
Posix.Error.exist is a val and not a constructor (like SOME) so
you cannot use it as a pattern in a pattern matching construct. Here
is an equivalent code for what you intended to do.
fun ensureDir s =
(OS.FileSys.mkDir s)
handle e as (OS.SysErr (_, SOME err)) =>
if err = Posix.Error.exist
then ()
else raise e
;
I am currently creating a CLI application in OCaml and using Core.Command, the CLI parser included in core (v0.10), to parse the command line.
I want to have a global flag that can be used for any subcommand (like the --paginate or --git-dir flags in git for example).
For instance, I want a -debug flag so that the two following commands are valid
my-cli -debug hello world
my-cli -debug goodbye world
However, I could not find a way to do this with the Core.Command API.
Here is a simplified version what I currently have.
open Core
let initialize_logger debug =
Logs.set_reporter (Logs_fmt.reporter ());
let log_level = if debug then Logs.Debug else Logs.Info in
Logs.set_level (Some log_level)
let some_func_with_logging () =
Logs.debug (fun m -> m "the flag debug was passed!")
let hello name =
some_func_with_logging ();
Printf.printf "Hello %s!\n" name
let goodbye name =
some_func_with_logging ();
Printf.printf "Goodbye %s!\n" name
let hello_command =
let open Command.Let_syntax in
Command.basic
~summary:"says hello"
[%map_open
let name = anon ("name" %: string)
and debug = flag "debug" no_arg ~doc:"debug" in
fun () ->
initialize_logger debug;
hello name
]
let goodbye_command =
let open Command.Let_syntax in
Command.basic
~summary:"says goodbye"
[%map_open
let name = anon ("name" %: string)
and debug = flag "debug" no_arg ~doc:"debug" in
fun () ->
initialize_logger debug;
goodbye name
]
let main_command =
Command.group ~summary:"a cool CLI tool"
[ ("hello", hello_command);
("goodbye", goodbye_command);
]
let () = Command.run main_command
There are two main issues here:
the debug flag as well as the call to initialize_logger is duplicated in every subcommand
the debug flag needs to be passed after the subcommand when invoking the command: my-cli hello world -debug instead of my-cli -debug hello world
Is there a clean way to handle this with Core.Command API?
get_ue_supported_srvcc([]) ->
?SRVCC_3GPP_NONE_SUPPORT;
get_ue_supported_srvcc([#sip_contactV{extensionsP = EP} | T]) ->
case b2bLib:support_tags_to_value(EP) of
?SRVCC_3GPP_NONE_SUPPORT ->
get_ue_supported_srvcc(T);
Flag ->
Flag
end.
I want create a unit test for this function,
Here is my unit test case:
get_ue_supported_srvcc_test() ->
Contact =
[#sip_contactV{extensionsP =
[{"+sip.instance",
{quoted_string,"<urn:gsma:imei:35502406-005233-0>"}},
{"+g.3gpp.icsi-ref",
{quoted_string,"urn%3Aurn-7%3A3gpp-service.ims.icsi.mmtel"}},
"+g.3gpp.mid-call",
"+g.3gpp.srvcc-alerting",
"+g.3gpp.ps2cs-srvcc-orig-pre-alerting",
"video"]}],
?assertEqual(7, b2bAtcfLib:get_ue_supported_srvcc(Contact)).
But when I run it, I get this error:
======================== EUnit ========================
module 'b2bAtcfLib'
b2bAtcfLib_tests: get_ue_supported_srvcc_test (module 'b2bAtcfLib_tests')...*failed*
in function b2bLib:support_tags_to_value/1
called as support_tags_to_value([{"+sip.instance",{quoted_string,"<urn:gsma:imei:35502406-005233-0>"}},
{"+g.3gpp.icsi-ref",
{quoted_string,"urn%3Aurn-7%3A3gpp-service.ims.icsi.mmtel"}},
"+g.3gpp.mid-call","+g.3gpp.srvcc-alerting",
"+g.3gpp.ps2cs-srvcc-orig-pre-alerting","video"])
in call from b2bAtcfLib:get_ue_supported_srvcc/1 (src/b2bAtcfLib.erl, line 1735)
in call from b2bAtcfLib_tests:'-get_ue_supported_srvcc_test/0-fun-0-'/1 (test/unit/b2bAtcfLib_tests.erl, line 49)
in call from b2bAtcfLib_tests:get_ue_supported_srvcc_test/0
**error:undef
output:<<"">>
[done in 0.008 s]
=======================================================
The error means b2bLib:support_tags_to_value/1 is undef.
The define for this function b2bLib:support_tags_to_value:
support_tags_to_value(FieldStr) ->
lists:sum([Val || {Tag, Val} <- ?TAGLIST, lists:member(Tag, FieldStr)]).
The error is:
**error:undef
That means that the test is calling a function that's not defined. Either the module couldn't be found, or the module in question doesn't define a function with that name and arity.
The whole error message is a bit confusing. Now that we know that we got a "function undefined" error, we should be looking at this line:
in function b2bLib:support_tags_to_value/1
Even though it says that the error occurred "in" this function, this is the function that's undefined.
So either the test is run in such a way that it doesn't find the b2bLib module, or that module doesn't define a function called support_tags_to_value taking one argument. If it's the former, add -pa path/to/ebin to the Erlang command line in order to add the right directory to the code path.
I currently have a project (Go to Python compiler) with the following files
ast.ml
parser.mly
lex.mll
weeder.ml
prettyPrint.ml
main.ml
Here are the dependencies:
parser: ast
lexer: parser, Core, Lexing
weeder: ast
prettyPrint: ast
main: ast, lex, parser, weeder, prettyPrint
I try to compile doing the following which should work according to the documentation I read:
$ menhir parser.mly
> Warning: you are using the standard library and/or the %inline keyword. We
recommend switching on --infer in order to avoid obscure type error messages.
$ ocamllex lex.mll
> 209 states, 11422 transitions, table size 46942 bytes
$ ocamlbuild -no-hygiene main.native
> File "parser.mli", line 77, characters 56-59:
Error: Unbound type constructor ast
Command exited with code 2.
Compilation unsuccessful after building 6 targets (2 cached) in 00:00:00.
ast.ml contains a list of type declarations in which I have a
type ast = ...
I spent a few hours now reading doc for ocamlfind, corebuild and ocamlopt and nothing. At some point it compiled by what seemed like a mere coincidence and never worked again. I'm open to using any tool.
Here is what is in parser.mly
%{
open Ast
exception ParserError of string
let rec deOptionTypeInList tupleList =
match tupleList with
| [] -> []
| (a, Some t)::tl -> (a, t)::(deOptionTypeInList tl)
| _ -> raise (ParserError "no type given in type declaration")
%}
[ ... long list of tokens ... ]
%type <ast> prog (* that seems to be the problem *)
%type <string> packDec
%type <dec> dec
%type <dec> subDec
[...]
%start prog
[ ... rules ... ]
And here is the line, the very last, that is refereed to in the error message.
val prog: (Lexing.lexbuf -> token) -> Lexing.lexbuf -> (ast)
The open Ast construct will not be exported to the .mli file where the type of symbols are mentioned. Try using
%type <Ast.ast>
Edit: also, your build commands are weird. You should not call ocamllex and menhir manually, and consequently not need -no-hygiene. Remove all generated files and just do
ocamlbuild -use-menhir main.byte
I wish to use OCaml to access the Yahoo Finance API. Essentially, it will be just a bunch of HTTP requests to get quotes from Yahoo Finance.
Which module I should use?
I wish to have async HTTP requests.
There are possibilities using lwt:
ocsigen has a quite complete and a bit complex implementation
cohttp is a bit simpler but lacks some usefull parts
using opam to install:
$ opam install ocsigenserver cohttp
For instance in a toplevel:
try Topdirs.dir_directory (Sys.getenv "OCAML_TOPLEVEL_PATH") with _ -> ();;
#use "topfind";;
#thread;;
#require "ocsigenserver";;
open Lwt
(* a simple function to access the content of the response *)
let content = function
| { Ocsigen_http_frame.frame_content = Some v } ->
Ocsigen_stream.string_of_stream 100000 (Ocsigen_stream.get v)
| _ -> return ""
(* launch both requests in parallel *)
let t = Lwt_list.map_p Ocsigen_http_client.get_url
[ "http://ocsigen.org/";
"http://stackoverflow.com/" ]
(* maps the result through the content function *)
let t2 = t >>= Lwt_list.map_p content
(* launch the event loop *)
let result = Lwt_main.run t2
and using cohttp:
try Topdirs.dir_directory (Sys.getenv "OCAML_TOPLEVEL_PATH") with _ -> ();;
#use "topfind";;
#require "cohttp.lwt";;
open Lwt
(* a simple function to access the content of the response *)
let content = function
| Some (_, body) -> Cohttp_lwt_unix.Body.string_of_body body
| _ -> return ""
(* launch both requests in parallel *)
let t = Lwt_list.map_p Cohttp_lwt_unix.Client.get
(List.map Uri.of_string
[ "http://example.org/";
"http://example2.org/" ])
(* maps the result through the content function *)
let t2 = t >>= Lwt_list.map_p content
(* launch the event loop *)
let v = Lwt_main.run t2
Notice that an implementation of cohttp for jane street async library is also available
Just for the record, there is also ocurl with curl multi API support.