caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Michael C Vanier <mvanier42@gmail.com>
To: "Nicolás Ojeda Bär" <nicolas.ojeda.bar@lexifi.com>
Cc: caml-list <caml-list@inria.fr>
Subject: Re: [Caml-list] Has anybody gotten delimcc to work?
Date: Thu, 22 Feb 2018 21:56:17 -0800	[thread overview]
Message-ID: <321eca40-e0df-a38f-6c17-a258dc2360c3@cms.caltech.edu> (raw)
In-Reply-To: <CADK7aFOFh0N1R_-Q4a34T69-D3EiSccN8oVPRHBbfRg6hr2DUA@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 8097 bytes --]

Thanks again, Nicolas!

Mike


On 2/22/18 9:15 PM, Nicolás Ojeda Bär wrote:
> Hello Mike,
>
> Indeed the opam-repository is just a bunch of metadata and the 
> opam-publish tool is used to update this metadata. Typically this is 
> done by the opam package maintainer which may or may not be the 
> package author (the latter in this case).
>
> In order to submit a patch to the delimcc package you need to figure 
> out the upstream repository. Typically you can do this by looking in 
> the opam metadata. In this case, the opam package is maintained at 
> https://github.com/zinid/delimcc, but there is no public repository of 
> the upstream source, rather it is distributed as a tar file by the 
> author (Oleg Kiselyov; see 
> http://okmij.org/ftp/continuations/implementations.html#caml-shift).
>
> I went ahead and submitted a patch to the package maintainer 
> (https://github.com/zinid/delimcc/pull/1) and also sent it by email to 
> Oleg so that the upstream sources can be amended.
>
> Best wishes,
> Nicolás
>
>
> On Fri, Feb 23, 2018 at 4:18 AM, Michael C Vanier <mvanier42@gmail.com 
> <mailto:mvanier42@gmail.com>> wrote:
>
>     So... I was trying to figure out how to make a pull request on the
>     delimcc repo (if there is one) and it was an exercise in
>     frustration.  The "opam-publish" tool doesn't seem to work anymore
>     with recent ocaml builds due to a nest of dependency issues:
>
>     =====
>
>     > opam install opam-publish
>     The following dependencies couldn't be met:
>       - opam-publish -> opam-lib (= 1.2.0 | = 1.2.2) -> dose < 3.4.0
>     -> extlib < 1.7.0
>       - opam-publish -> opam-lib (= 1.2.0 | = 1.2.2) -> dose < 3.4.0
>     -> extlib-compat < 1.7.0
>       - opam-publish -> opam-lib (= 1.2.0 | = 1.2.2) -> dose < 3.4.0
>     -> ocamlgraph <= 1.8.5
>       - opam-publish -> publish -> opam-publish > 0.3.4 -> opam-lib >
>     1.2.2 -> jsonm -> uutf (<= 0.9.4 | >= 1.0.0)
>       - opam-publish -> publish -> opam-publish > 0.3.4 -> opam-lib >
>     1.2.2 -> cmdliner <= 0.9.8
>       - opam-publish -> publish -> opam-lib (= 1.2.2 | > 1.2.2) ->
>     cmdliner <= 0.9.8
>       - opam-publish -> publish -> opam-lib (= 1.2.2 | > 1.2.2) ->
>     dose < 3.4.0 -> ocamlgraph <= 1.8.5
>       - opam-publish -> publish -> opam-lib (= 1.2.2 | > 1.2.2) ->
>     dose < 3.4.0 -> extlib-compat < 1.7.0
>       - opam-publish -> publish -> opam-lib (= 1.2.2 | > 1.2.2) ->
>     dose < 3.4.0 -> extlib < 1.7.0
>       - opam-publish -> publish -> opam-lib (= 1.2.2 | > 1.2.2) ->
>     jsonm -> uutf (<= 0.9.4 | >= 1.0.0)
>     Your request can't be satisfied:
>       - cmdliner.0.9.4 is in conflict with uutf.1.0.1
>       - extlib-compat<1.7.0 is not available because your system
>     doesn't comply with ocaml-version < "4.05.0".
>       - extlib<1.7.0 is not available because your system doesn't
>     comply with ocaml-version < "4.05.0".
>       - ocamlgraph.1.8.5 is not available because your system doesn't
>     comply with ocaml-version < "4.06.0".
>       - ocamlgraph<=1.8.5 is not available because your system doesn't
>     comply with ocaml-version < "4.06.0".
>       - uutf<=0.9.4 is not available because your system doesn't
>     comply with ocaml-version >= "4.00.0" & ocaml-version < "4.06.0".
>
>     No solution found, exiting
>
>     =====
>
>     I tried forking the opam-repository repo but there is only
>     metadata about the various packages and not the actual code to
>     make a patch on.  Is there some documentation on how to submit a
>     bug fix to a package? Sorry if this is a dumb question.
>
>     Mike
>
>
>     On 2/22/18 3:58 PM, Michael C Vanier wrote:
>>
>>     It works!  Thank you so much Nicolas!
>>
>>     Mike
>>
>>
>>     On 2/22/18 3:42 PM, Nicolás Ojeda Bär wrote:
>>>     Dear Mike,
>>>
>>>     I was able to successfully load delimcc into the toplevel after
>>>     applying the tiny patch below.
>>>     You can give it a try by doing:
>>>
>>>       opam source delimcc
>>>       cd delimcc.(version)
>>>       (apply patch)
>>>       opam pin add .
>>>
>>>     (you have to replace the stuff between parentheses by whatever
>>>     is correct in your setup.)
>>>
>>>     diff --git a/stacks.c b/stacks.c
>>>     index fdab2a7..5765710 100644
>>>     --- a/stacks.c
>>>     +++ b/stacks.c
>>>     @@ -203,13 +203,13 @@ value copy_stack_fragment(const value vek1)
>>>      #endif
>>>        if (size < Max_young_wosize) {
>>>     -    block = alloc(size, 0);
>>>     +    block = caml_alloc(size, 0);
>>>          memcpy(&Field(block, 0), tp2, size * sizeof(value));
>>>        } else {
>>>     -    block = alloc_shr(size, 0);
>>>     +    block = caml_alloc_shr(size, 0);
>>>          mlsize_t i;
>>>          for (i = 0; i < size; i++)
>>>     -      initialize(&Field(block, i), tp2[i]);
>>>     +      caml_initialize(&Field(block, i), tp2[i]);
>>>        }
>>>        /* We check the invariants after the allocation of block,
>>>     which may
>>>     @@ -382,4 +382,3 @@ value dbg_note(const value message)
>>>        fprintf(stderr,"%s\n",String_val(message));
>>>        return Val_unit;
>>>      }
>>>
>>>     Hope it helps,
>>>     Nicolás
>>>
>>>
>>>
>>>     On Fri, Feb 23, 2018 at 12:18 AM, Michael C Vanier
>>>     <mvanier@cms.caltech.edu <mailto:mvanier@cms.caltech.edu>> wrote:
>>>
>>>         I've been trying to use the delimcc delimited continuation
>>>         library, but so far I've been unsuccessful.  I'm using OCaml
>>>         4.06.1 and I've tried it on both Mac OS X (High Sierra) and
>>>         Ubuntu MATE 17.10. I'm fine sticking to bytecode.  The opam
>>>         package installs correctly but when you try to run it you
>>>         get errors from the dynamically-linked libraries. On Mac OS
>>>         X I get this:
>>>
>>>         # #require "delimcc";;
>>>         Cannot load required shared library dlldelimcc.
>>>         Reason:
>>>         /Users/mvanier/.opam/4.06.1/lib/stublibs/dlldelimcc.so:
>>>         dlopen(/Users/mvanier/.opam/4.06.1/lib/stublibs/dlldelimcc.so,
>>>         10): Symbol not found: _alloc
>>>           Referenced from:
>>>         /Users/mvanier/.opam/4.06.1/lib/stublibs/dlldelimcc.so
>>>           Expected in: flat namespace
>>>          in /Users/mvanier/.opam/4.06.1/lib/stublibs/dlldelimcc.so.
>>>
>>>         And on Ubuntu I get this:
>>>
>>>         # #require "delimcc";;
>>>         /home/mvanier/.opam/4.06.1/lib/delimcc: added to search path
>>>         /home/mvanier/.opam/4.06.1/lib/delimcc/delimcc.cma: loaded
>>>         Cannot load required shared library dlldelimcc.
>>>         Reason:
>>>         /Users/mvanier/.opam/4.06.1/lib/stublibs/dlldelimcc.so:
>>>         dlopen(/Users/mvanier/.opam/4.06.1/lib/stublibs/dlldelimcc.so,
>>>         10): Symbol not found: initialize.
>>>
>>>         However, in either case I can do:
>>>
>>>         # open Delimcc;;
>>>
>>>         and it reports no errors, but then if I try e.g.:
>>>
>>>         # shift;;
>>>         Reference to undefined global `Delimcc`
>>>
>>>         so it appears the library isn't there or isn't functional.
>>>         Compiling from the delimcc source also appears to work, but
>>>         gcc warns about implicit declarations of "alloc",
>>>         "alloc_shr" and "initialize" when compiling stacks.c, which
>>>         makes sense.  Compiling any of the test programs also fails.
>>>
>>>         So it appears that this library no longer works.  Does
>>>         anyone know any way to make it work?
>>>
>>>         Thanks in advance,
>>>
>>>         Mike
>>>
>>>
>>>         -- 
>>>         Caml-list mailing list.  Subscription management and archives:
>>>         https://sympa.inria.fr/sympa/arc/caml-list
>>>         <https://sympa.inria.fr/sympa/arc/caml-list>
>>>         Beginner's list:
>>>         http://groups.yahoo.com/group/ocaml_beginners
>>>         <http://groups.yahoo.com/group/ocaml_beginners>
>>>         Bug reports: http://caml.inria.fr/bin/caml-bugs
>>>         <http://caml.inria.fr/bin/caml-bugs>
>>>
>>>
>>
>
>


[-- Attachment #2: Type: text/html, Size: 15233 bytes --]

  reply	other threads:[~2018-02-23  5:56 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-02-22 23:18 Michael C Vanier
2018-02-22 23:42 ` Nicolás Ojeda Bär
2018-02-22 23:58   ` Michael C Vanier
2018-02-23  3:18     ` Michael C Vanier
2018-02-23  5:15       ` Nicolás Ojeda Bär
2018-02-23  5:56         ` Michael C Vanier [this message]
2018-02-26  5:42         ` Oleg

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=321eca40-e0df-a38f-6c17-a258dc2360c3@cms.caltech.edu \
    --to=mvanier42@gmail.com \
    --cc=caml-list@inria.fr \
    --cc=nicolas.ojeda.bar@lexifi.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).