caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
* [Caml-list] Delimcc bug
@ 2017-06-02 14:32 Evgeny Khramtsov
  2017-06-02 15:05 ` oleg
  0 siblings, 1 reply; 7+ messages in thread
From: Evgeny Khramtsov @ 2017-06-02 14:32 UTC (permalink / raw)
  To: caml-list

Does anyone know where to report bugs in Delimcc?
Is it even supported?
Under some conditions I have segfaults in it when compiled natively in
4.04.1.

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [Caml-list] Delimcc bug
  2017-06-02 14:32 [Caml-list] Delimcc bug Evgeny Khramtsov
@ 2017-06-02 15:05 ` oleg
  2017-06-02 15:31   ` Ivan Gotovchits
                     ` (2 more replies)
  0 siblings, 3 replies; 7+ messages in thread
From: oleg @ 2017-06-02 15:05 UTC (permalink / raw)
  To: Evgeny Khramtsov; +Cc: caml-list

On Fri, Jun 02, 2017 at 05:32:21PM +0300, Evgeny Khramtsov wrote:
> Does anyone know where to report bugs in Delimcc?
> Is it even supported?
> Under some conditions I have segfaults in it when compiled natively in
> 4.04.1.
> 

You can send me directly, or send to the list if you want. But be sure
you fetch the latest version from my web page. The problem is usually
GCC, which optimizes too much. So, special flags are required to switch
off some optmizations. You can see in the Makefile. If you have a very
recent version of GCC, chances are it needs some more flags to disable
too aggressive optimizations.

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [Caml-list] Delimcc bug
  2017-06-02 15:05 ` oleg
@ 2017-06-02 15:31   ` Ivan Gotovchits
  2017-06-02 15:52   ` Evgeny Khramtsov
  2017-06-03  5:26   ` Evgeny Khramtsov
  2 siblings, 0 replies; 7+ messages in thread
From: Ivan Gotovchits @ 2017-06-02 15:31 UTC (permalink / raw)
  To: oleg; +Cc: Evgeny Khramtsov, caml-list

[-- Attachment #1: Type: text/plain, Size: 1289 bytes --]

Wouldn't it be better to add directly to the implementation a GCC pragma
[1] or an attribute [2] that will explicitly disable the optimizations?


[1]:
https://gcc.gnu.org/onlinedocs/gcc/Function-Specific-Option-Pragmas.html#Function-Specific-Option-Pragmas
[2]:
https://gcc.gnu.org/onlinedocs/gcc/Common-Function-Attributes.html#Common-Function-Attributes



On Fri, Jun 2, 2017 at 5:05 PM, <oleg@okmij.org> wrote:

> On Fri, Jun 02, 2017 at 05:32:21PM +0300, Evgeny Khramtsov wrote:
> > Does anyone know where to report bugs in Delimcc?
> > Is it even supported?
> > Under some conditions I have segfaults in it when compiled natively in
> > 4.04.1.
> >
>
> You can send me directly, or send to the list if you want. But be sure
> you fetch the latest version from my web page. The problem is usually
> GCC, which optimizes too much. So, special flags are required to switch
> off some optmizations. You can see in the Makefile. If you have a very
> recent version of GCC, chances are it needs some more flags to disable
> too aggressive optimizations.
>
> --
> Caml-list mailing list.  Subscription management and archives:
> https://sympa.inria.fr/sympa/arc/caml-list
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs
>

[-- Attachment #2: Type: text/html, Size: 2317 bytes --]

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [Caml-list] Delimcc bug
  2017-06-02 15:05 ` oleg
  2017-06-02 15:31   ` Ivan Gotovchits
@ 2017-06-02 15:52   ` Evgeny Khramtsov
  2017-06-03  5:26   ` Evgeny Khramtsov
  2 siblings, 0 replies; 7+ messages in thread
From: Evgeny Khramtsov @ 2017-06-02 15:52 UTC (permalink / raw)
  To: caml-list

Fri, 2 Jun 2017 11:05:57 -0400
oleg@okmij.org wrote:

> You can send me directly, or send to the list if you want. But be sure
> you fetch the latest version from my web page. The problem is usually
> GCC, which optimizes too much. So, special flags are required to
> switch off some optmizations. You can see in the Makefile. If you
> have a very recent version of GCC, chances are it needs some more
> flags to disable too aggressive optimizations.

I'm using delimcc from opam, the url it's pointing to is
http://okmij.org/ftp/continuations/caml-shift.tar.gz
Sometimes it segfaults, sometimes it stops with "take_subcont: can't
happen". I tried with gcc 6.2.0 and 4.9.2. Do I need even older gcc? :)

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [Caml-list] Delimcc bug
  2017-06-02 15:05 ` oleg
  2017-06-02 15:31   ` Ivan Gotovchits
  2017-06-02 15:52   ` Evgeny Khramtsov
@ 2017-06-03  5:26   ` Evgeny Khramtsov
  2017-06-03  9:08     ` Jeremy Yallop
  2 siblings, 1 reply; 7+ messages in thread
From: Evgeny Khramtsov @ 2017-06-03  5:26 UTC (permalink / raw)
  To: caml-list

Fri, 2 Jun 2017 11:05:57 -0400
oleg@okmij.org wrote:

> So, special flags are required to switch
> off some optmizations.

I confirm, replacing -O2 with -O0 flag for compiling stacks-native.c
fixes the problem.
However, I think this (or something better) should be applied in the
patch used to compile delimcc by opam (makefile_fix.patch).

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [Caml-list] Delimcc bug
  2017-06-03  5:26   ` Evgeny Khramtsov
@ 2017-06-03  9:08     ` Jeremy Yallop
  2017-06-03 13:04       ` Evgeny Khramtsov
  0 siblings, 1 reply; 7+ messages in thread
From: Jeremy Yallop @ 2017-06-03  9:08 UTC (permalink / raw)
  To: Evgeny Khramtsov; +Cc: Caml List

On 3 June 2017 at 06:26, Evgeny Khramtsov <xramtsov@gmail.com> wrote:
> Fri, 2 Jun 2017 11:05:57 -0400
> oleg@okmij.org wrote:
>
>> So, special flags are required to switch
>> off some optmizations.
>
> I confirm, replacing -O2 with -O0 flag for compiling stacks-native.c
> fixes the problem.
> However, I think this (or something better) should be applied in the
> patch used to compile delimcc by opam (makefile_fix.patch).

The best way to achieve this is to send a pull request to the
opam-repository repository:

   https://github.com/ocaml/opam-repository/pulls

mentioning the package maintainer (@rgrinberg) in the description.

Kind regards,

Jeremy

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [Caml-list] Delimcc bug
  2017-06-03  9:08     ` Jeremy Yallop
@ 2017-06-03 13:04       ` Evgeny Khramtsov
  0 siblings, 0 replies; 7+ messages in thread
From: Evgeny Khramtsov @ 2017-06-03 13:04 UTC (permalink / raw)
  To: Caml List

Sat, 3 Jun 2017 10:08:12 +0100
Jeremy Yallop <yallop@gmail.com> wrote:

> The best way to achieve this is to send a pull request to the
> opam-repository repository:
> 
>    https://github.com/ocaml/opam-repository/pulls

Done: https://github.com/ocaml/opam-repository/pull/9401
Thanks for the tip :)

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2017-06-03 13:05 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-06-02 14:32 [Caml-list] Delimcc bug Evgeny Khramtsov
2017-06-02 15:05 ` oleg
2017-06-02 15:31   ` Ivan Gotovchits
2017-06-02 15:52   ` Evgeny Khramtsov
2017-06-03  5:26   ` Evgeny Khramtsov
2017-06-03  9:08     ` Jeremy Yallop
2017-06-03 13:04       ` Evgeny Khramtsov

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).