digitalmars.D.learn - Debugging D shared libraries
- Russel Winder via Digitalmars-d-learn (39/39) Sep 19 2015 Calling D from Python. I have two functions in D, compiled to a shared
- John Colvin (6/10) Sep 19 2015 I heard it crashed during the talk. Bummer. I should really be
- Russel Winder via Digitalmars-d-learn (20/25) Sep 19 2015 [=E2=80=A6]
- ponce (4/7) Sep 19 2015 Friend call me like this IRL since forever.
- Russel Winder via Digitalmars-d-learn (16/20) Sep 19 2015 English and Spanish meanings of the word are very different. In UK (not
- ponce (3/6) Sep 20 2015 In French it means "to rub down with abrasive paper".
- ponce (21/52) Sep 19 2015 Try using an explicit TaskPool and destroying it with scope(exit).
- Russel Winder via Digitalmars-d-learn (14/42) Sep 19 2015 Hummm=E2=80=A6 I now do not get a segfault, and the code runs as expecte...
- ponce (4/8) Sep 19 2015 Where is it stuck?
- Russel Winder via Digitalmars-d-learn (26/38) Sep 19 2015 I commented out the :
- John Colvin (5/12) Sep 19 2015 version(LDC){ /* ... */ }
- Russel Winder via Digitalmars-d-learn (15/31) Sep 19 2015 Indeed, it works well. Well for LDC. DMD and GDC are still broken. My
- Martin Krejcirik (6/9) Sep 20 2015 All I need to do to make your example work, is to move the calls of
- Johannes Pfau (9/32) Sep 20 2015 Am Sat, 19 Sep 2015 17:41:41 +0100
- Johannes Pfau (4/39) Sep 20 2015 Just realized this thread is titled "Debugging D shared libraries" ;-)
- Russel Winder via Digitalmars-d-learn (14/19) Sep 22 2015 Conversely I thought it did due to the GCC toolchain thingy.
- Russel Winder via Digitalmars-d-learn (25/30) Sep 22 2015 [=E2=80=A6]
- John Colvin (4/25) Sep 22 2015 seeing as it's an error in core.stdc.config, I'd say it's
- Johannes Pfau (8/40) Sep 22 2015 It's indeed strange.
- Russel Winder via Digitalmars-d-learn (17/27) Sep 19 2015 On the otherhand using a:
- Laeeth Isharc (4/63) Sep 19 2015 What is the difference between shared static this and the global
- ponce (5/8) Sep 19 2015 Would like to know too. On OSX I've found that shared static
- John Colvin (4/9) Sep 19 2015 IIRC shared static this is only called by the runtime, so
- Laeeth Isharc (2/12) Sep 19 2015 Aha... Is there an equivalent in dmd to the global constructor ?
- Russel Winder via Digitalmars-d-learn (18/22) Sep 19 2015 [=E2=80=A6]
- Laeeth Isharc (9/40) Sep 19 2015 Btw have you looked at Colvin's prettypyd ? It's a nicer way to
- Laeeth Isharc (7/53) Sep 19 2015 In particular it should just work that way as PyD should deal
- rom (7/38) Sep 20 2015 Isn't it simply that when doing some work asynchronously, as with
- Russel Winder via Digitalmars-d-learn (21/40) Sep 22 2015 [=E2=80=A6]
Calling D from Python. I have two functions in D, compiled to a shared
object on Linux using LDC (but I get same problem using DMD).
The sequential code:
extern(C)
double sequential(const int n, const double delta) {
Runtime.initialize();
const pi =3D 4.0 * delta * reduce!(
(double t, int i){ immutable x =3D (i - 0.5) * delta; return t =
+ 1.0 / (1.0 + x * x); })(
0.0, iota(1, n + 1));
Runtime.terminate();
return pi;
}
works entirely fine. However the "parallel" code:
extern(C)
double parallel(const int n, const double delta) {
Runtime.initialize();
const pi =3D 4.0 * delta * taskPool.reduce!"a + b"(
map!((int i){ immutable x =3D (i - 0.5) * delta; return 1.0 / (1.=
0 + x * x); })(iota(1, n + 1)));
Runtime.terminate();
return pi;
}
causes an immediate segfault (with LDC and DMD. I am assuming that the
problem is the lack of initialization of the std.parallelism module and
hence the use of taskPool is causing a problem. I am betting I am
missing something very simple about module initialization, and that
this is not actually a bug.
Anyone any proposals?
--=20
Russel.
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D
Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n=
et
41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk
London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 19 2015
On Saturday, 19 September 2015 at 10:45:22 UTC, Russel Winder wrote:Calling D from Python. I have two functions in D, compiled to a shared object on Linux using LDC (but I get same problem using DMD). [...]I heard it crashed during the talk. Bummer. I should really be there, seeing as I live about 15 mins away. If you get a chance to talk to Alex Bishop, don't be too harsh on D to him, I'm trying to convince him to make the switch :)
Sep 19 2015
On Sat, 2015-09-19 at 11:07 +0000, John Colvin via Digitalmars-d-learn wrote:=20[=E2=80=A6]I heard it crashed during the talk. Bummer. I should really be=20 there, seeing as I live about 15 mins away. If you get a chance=20 to talk to Alex Bishop, don't be too harsh on D to him, I'm=20 trying to convince him to make the switch :)With the input from ponce (*) I now have the code working. Now all is need is a time machine so as to redo the demo successfully. Doctor, where are you? Oh you are committed to be on BBC1 in two hours time. Get Alex to have a chat with me and I can show the demonstration working. (*) ponce is arguably not the most positive or constructive name to go by. =20 --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 19 2015
On Saturday, 19 September 2015 at 16:32:18 UTC, Russel Winder wrote:(*) ponce is arguably not the most positive or constructive name to go by.Friend call me like this IRL since forever. It seems to be a swear word in english?
Sep 19 2015
On Sat, 2015-09-19 at 16:41 +0000, ponce via Digitalmars-d-learn wrote:=20[=E2=80=A6]Friend call me like this IRL since forever. =20 It seems to be a swear word in english?English and Spanish meanings of the word are very different. In UK (not sure about Canada, USA, Australia, New Zealand, South Africa,=E2=80=A6) it = is generally a somewhat derogatory term. --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 19 2015
ponce <contact gam3sfrommars.fr> On Saturday, 19 September 2015 at 17:02:37 UTC, Russel Winder wrote:English and Spanish meanings of the word are very different. In UK (not sure about Canada, USA, Australia, New Zealand, South Africa,…) it is generally a somewhat derogatory term.In French it means "to rub down with abrasive paper".
Sep 20 2015
On Saturday, 19 September 2015 at 10:45:22 UTC, Russel Winder
wrote:
Calling D from Python. I have two functions in D, compiled to a
shared object on Linux using LDC (but I get same problem using
DMD).
The sequential code:
extern(C)
double sequential(const int n, const double delta) {
Runtime.initialize();
const pi = 4.0 * delta * reduce!(
(double t, int i){ immutable x = (i - 0.5) * delta;
return t + 1.0 / (1.0 + x * x); })(
0.0, iota(1, n + 1));
Runtime.terminate();
return pi;
}
works entirely fine. However the "parallel" code:
extern(C)
double parallel(const int n, const double delta) {
Runtime.initialize();
const pi = 4.0 * delta * taskPool.reduce!"a + b"(
map!((int i){ immutable x = (i - 0.5) * delta; return
1.0 / (1.0 + x * x); })(iota(1, n + 1)));
Runtime.terminate();
return pi;
}
causes an immediate segfault (with LDC and DMD. I am assuming
that the problem is the lack of initialization of the
std.parallelism module and hence the use of taskPool is causing
a problem. I am betting I am missing something very simple
about module initialization, and that this is not actually a
bug.
Anyone any proposals?
Try using an explicit TaskPool and destroying it with scope(exit).
Also if using LDC, you can use global ctor/dtor to deal with the
runtime.
----------------------->8---------------------
extern (C) {
pragma(LDC_global_crt_ctor, 0)
void initRuntime()
{
import core.runtime;
Runtime.initialize();
}
pragma(LDC_global_crt_dtor, 0)
void deinitRuntime()
{
import core.runtime;
Runtime.terminate();
}
}
----------------------->8---------------------
Sep 19 2015
On Sat, 2015-09-19 at 12:21 +0000, ponce via Digitalmars-d-learn wrote:
[=E2=80=A6]
=20
Try using an explicit TaskPool and destroying it with scope(exit).
=20
=20
Also if using LDC, you can use global ctor/dtor to deal with the=20
runtime.
=20
=20
----------------------->8---------------------
=20
extern (C) {
pragma(LDC_global_crt_ctor, 0)
void initRuntime()
{
import core.runtime;
Runtime.initialize();
}
pragma(LDC_global_crt_dtor, 0)
void deinitRuntime()
{
import core.runtime;
Runtime.terminate();
}
}
=20
----------------------->8---------------------
=20
Hummm=E2=80=A6 I now do not get a segfault, and the code runs as expected :
-))))) but the program never terminates. :-(
Also, what would I need to cover the DMD and the GDC situations?
--=20
Russel.
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D
Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n=
et
41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk
London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 19 2015
On Saturday, 19 September 2015 at 15:42:15 UTC, Russel Winder wrote:Hummm… I now do not get a segfault, and the code runs as expected : -))))) but the program never terminates. :-(Where is it stuck?Also, what would I need to cover the DMD and the GDC situations?I don't know. :(
Sep 19 2015
On Sat, 2015-09-19 at 15:58 +0000, ponce via Digitalmars-d-learn wrote:On Saturday, 19 September 2015 at 15:42:15 UTC, Russel Winder=20 wrote:I commented out the : //pragma(LDC_global_crt_dtor, 0) //void deinitRuntime() { //import core.runtime: Runtime; //Runtime.terminate(); //} and it now works fine :-)))))=20 Hummm=E2=80=A6 I now do not get a segfault, and the code runs as=20 expected : -))))) but the program never terminates. :-(=20 Where is it stuck?Let's hope DMD and GDC folk chip in :-) Sadly the: pragma(LDC_global_crt_ctor, 0) void initRuntime() { import core.runtime: Runtime; Runtime.initialize(); } will not compile under DMD :-( --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder=20 Also, what would I need to cover the DMD and the GDC situations?=20 I don't know. :(
Sep 19 2015
On Saturday, 19 September 2015 at 16:15:45 UTC, Russel Winder
wrote:
Sadly the:
pragma(LDC_global_crt_ctor, 0)
void initRuntime() {
import core.runtime: Runtime;
Runtime.initialize();
}
will not compile under DMD :-(
version(LDC){ /* ... */ }
not that it helps make things work correctly, but at least
they'll compile :)
Sep 19 2015
On Sat, 2015-09-19 at 16:33 +0000, John Colvin via Digitalmars-d-learn wrote:On Saturday, 19 September 2015 at 16:15:45 UTC, Russel Winder=20 wrote:Indeed, it works well. Well for LDC. DMD and GDC are still broken. My GDC problems are deeper that this code: Debian packages seem to have weird problems and Fedora do not package GDC. --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winderSadly the: =20 pragma(LDC_global_crt_ctor, 0) void initRuntime() { import core.runtime: Runtime; Runtime.initialize(); } =20 will not compile under DMD :-(=20 version(LDC){ /* ... */ } =20 not that it helps make things work correctly, but at least=20 they'll compile :)
Sep 19 2015
Dne 19. 9. 2015 v 18:41 Russel Winder via Digitalmars-d-learn napsal(a):Indeed, it works well. Well for LDC. DMD and GDC are still broken. My GDC problems are deeper that this code: Debian packages seem to have weird problems and Fedora do not package GDC.All I need to do to make your example work, is to move the calls of Runtime.initialize and Runtime.terminate to a wrapper functions called at the start and end of the main program. -- mk
Sep 20 2015
Am Sat, 19 Sep 2015 17:41:41 +0100 schrieb Russel Winder via Digitalmars-d-learn <digitalmars-d-learn puremagic.com>:On Sat, 2015-09-19 at 16:33 +0000, John Colvin via Digitalmars-d-learn wrote:Have you tried using a newer GDC version? The debian jessie version probably uses the 2.064.2 frontend. I wanted to add attribute(cctor/cdtor) support for some time now, I even wrote the code some time but didn't push it to the main repo for some reason. I'll put it on the TODO list but I can't work on this for the next 2-3 weeks.On Saturday, 19 September 2015 at 16:15:45 UTC, Russel Winder wrote:Indeed, it works well. Well for LDC. DMD and GDC are still broken. My GDC problems are deeper that this code: Debian packages seem to have weird problems and Fedora do not package GDC.Sadly the: pragma(LDC_global_crt_ctor, 0) void initRuntime() { import core.runtime: Runtime; Runtime.initialize(); } will not compile under DMD :-(version(LDC){ /* ... */ } not that it helps make things work correctly, but at least they'll compile :)
Sep 20 2015
Am Sun, 20 Sep 2015 17:47:00 +0200 schrieb Johannes Pfau <nospam example.com>:Am Sat, 19 Sep 2015 17:41:41 +0100 schrieb Russel Winder via Digitalmars-d-learn <digitalmars-d-learn puremagic.com>:Just realized this thread is titled "Debugging D shared libraries" ;-) GDC does not yet support shared libraries.On Sat, 2015-09-19 at 16:33 +0000, John Colvin via Digitalmars-d-learn wrote:Have you tried using a newer GDC version? The debian jessie version probably uses the 2.064.2 frontend. I wanted to add attribute(cctor/cdtor) support for some time now, I even wrote the code some time but didn't push it to the main repo for some reason. I'll put it on the TODO list but I can't work on this for the next 2-3 weeks.On Saturday, 19 September 2015 at 16:15:45 UTC, Russel Winder wrote:Indeed, it works well. Well for LDC. DMD and GDC are still broken. My GDC problems are deeper that this code: Debian packages seem to have weird problems and Fedora do not package GDC.Sadly the: pragma(LDC_global_crt_ctor, 0) void initRuntime() { import core.runtime: Runtime; Runtime.initialize(); } will not compile under DMD :-(version(LDC){ /* ... */ } not that it helps make things work correctly, but at least they'll compile :)
Sep 20 2015
On Sun, 2015-09-20 at 17:49 +0200, Johannes Pfau via Digitalmars-d -learn wrote:[=E2=80=A6] =20 Just realized this thread is titled "Debugging D shared libraries" ; -) GDC does not yet support shared libraries.Conversely I thought it did due to the GCC toolchain thingy. I'm using DMD and LDC just now so I have no actual data. --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 22 2015
On Sun, 2015-09-20 at 17:47 +0200, Johannes Pfau via Digitalmars-d -learn wrote:Am Sat, 19 Sep 2015 17:41:41 +0100 =20[=E2=80=A6]=20 Have you tried using a newer GDC version? The debian jessie version probably uses the 2.064.2 frontend.Debian Jessie is far too out of date to be useful. I'm on Debian Sid (still quite old), and Fedora Rawhide (not quite so old). Sadly GDC on Debian Sid tells me 5.2.1 20150911 which may well tell Iain which DMD is being used, but I haven't a clue. :-) [=E2=80=A6] The real problem using GDC is: gdc -I. -O3 -fPIC -c -o processAll_library_d.o processAll_library_d.d /usr/include/d/core/stdc/config.d:28:3: error: static if conditional cannot be at global scope static if( (void*).sizeof > int.sizeof ) ^ I haven't had chance to sit down and see if this is reasonable or not. --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 22 2015
On Tuesday, 22 September 2015 at 14:37:11 UTC, Russel Winder wrote:On Sun, 2015-09-20 at 17:47 +0200, Johannes Pfau via Digitalmars-d -learn wrote:seeing as it's an error in core.stdc.config, I'd say it's definitely not reasonable[...][…][...]Debian Jessie is far too out of date to be useful. I'm on Debian Sid (still quite old), and Fedora Rawhide (not quite so old). Sadly GDC on Debian Sid tells me 5.2.1 20150911 which may well tell Iain which DMD is being used, but I haven't a clue. :-) […] The real problem using GDC is: gdc -I. -O3 -fPIC -c -o processAll_library_d.o processAll_library_d.d /usr/include/d/core/stdc/config.d:28:3: error: static if conditional cannot be at global scope static if( (void*).sizeof > int.sizeof ) ^ I haven't had chance to sit down and see if this is reasonable or not.
Sep 22 2015
Am Tue, 22 Sep 2015 14:40:43 +0000 schrieb John Colvin <john.loughran.colvin gmail.com>:On Tuesday, 22 September 2015 at 14:37:11 UTC, Russel Winder=20 wrote:It's indeed strange. =20 Russel Winder if you can reproduce this with the latest GDC* please file a bug report. Even the error message doesn't make sense: static if works fine at global scope, AFAIK. * you could use the latest binaries from http://gdcproject.org/downloadsOn Sun, 2015-09-20 at 17:47 +0200, Johannes Pfau via=20 Digitalmars-d -learn wrote:=20 seeing as it's an error in core.stdc.config, I'd say it's=20 definitely not reasonable[...][=E2=80=A6][...]Debian Jessie is far too out of date to be useful. I'm on=20 Debian Sid (still quite old), and Fedora Rawhide (not quite so old). Sadly GDC on Debian Sid tells me 5.2.1 20150911 which may well=20 tell Iain which DMD is being used, but I haven't a clue. :-) [=E2=80=A6] The real problem using GDC is: gdc -I. -O3 -fPIC -c -o processAll_library_d.o=20 processAll_library_d.d /usr/include/d/core/stdc/config.d:28:3: error: static if=20 conditional cannot be at global scope static if( (void*).sizeof > int.sizeof ) ^ I haven't had chance to sit down and see if this is reasonable=20 or not.
Sep 22 2015
On Sat, 2015-09-19 at 17:15 +0100, Russel Winder wrote:=20[=E2=80=A6]Sadly the: =20 pragma(LDC_global_crt_ctor, 0) void initRuntime() { import core.runtime: Runtime; Runtime.initialize(); } =20 will not compile under DMD :-(On the otherhand using a: version(LDC) { =E2=80=A6 } solves the problem. --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 19 2015
On Saturday, 19 September 2015 at 12:21:02 UTC, ponce wrote:On Saturday, 19 September 2015 at 10:45:22 UTC, Russel Winder wrote:What is the difference between shared static this and the global constructor ? Russell, if you use shared static this for dmd does it work ? Laeeth.Calling D from Python. I have two functions in D, compiled to a shared object on Linux using LDC (but I get same problem using DMD). The sequential code: extern(C) double sequential(const int n, const double delta) { Runtime.initialize(); const pi = 4.0 * delta * reduce!( (double t, int i){ immutable x = (i - 0.5) * delta; return t + 1.0 / (1.0 + x * x); })( 0.0, iota(1, n + 1)); Runtime.terminate(); return pi; } works entirely fine. However the "parallel" code: extern(C) double parallel(const int n, const double delta) { Runtime.initialize(); const pi = 4.0 * delta * taskPool.reduce!"a + b"( map!((int i){ immutable x = (i - 0.5) * delta; return 1.0 / (1.0 + x * x); })(iota(1, n + 1))); Runtime.terminate(); return pi; } causes an immediate segfault (with LDC and DMD. I am assuming that the problem is the lack of initialization of the std.parallelism module and hence the use of taskPool is causing a problem. I am betting I am missing something very simple about module initialization, and that this is not actually a bug. Anyone any proposals?Try using an explicit TaskPool and destroying it with scope(exit). Also if using LDC, you can use global ctor/dtor to deal with the runtime. ----------------------->8--------------------- extern (C) { pragma(LDC_global_crt_ctor, 0) void initRuntime() { import core.runtime; Runtime.initialize(); } pragma(LDC_global_crt_dtor, 0) void deinitRuntime() { import core.runtime; Runtime.terminate(); } } ----------------------->8---------------------
Sep 19 2015
On Saturday, 19 September 2015 at 16:25:28 UTC, Laeeth Isharc wrote:What is the difference between shared static this and the global constructor ? Russell, if you use shared static this for dmd does it work ? Laeeth.Would like to know too. On OSX I've found that shared static this() was called by Runtime.initialize(), maybe it's different on Linux.
Sep 19 2015
On Saturday, 19 September 2015 at 16:25:28 UTC, Laeeth Isharc wrote:On Saturday, 19 September 2015 at 12:21:02 UTC, ponce wrote:IIRC shared static this is only called by the runtime, so obviously you can't use it to initialise the runtime itself.[...]What is the difference between shared static this and the global constructor ? Russell, if you use shared static this for dmd does it work ? Laeeth.
Sep 19 2015
On Saturday, 19 September 2015 at 16:34:05 UTC, John Colvin wrote:On Saturday, 19 September 2015 at 16:25:28 UTC, Laeeth Isharc wrote:Aha... Is there an equivalent in dmd to the global constructor ?On Saturday, 19 September 2015 at 12:21:02 UTC, ponce wrote:IIRC shared static this is only called by the runtime, so obviously you can't use it to initialise the runtime itself.[...]What is the difference between shared static this and the global constructor ? Russell, if you use shared static this for dmd does it work ? Laeeth.
Sep 19 2015
On Sat, 2015-09-19 at 16:25 +0000, Laeeth Isharc via Digitalmars-d -learn wrote:=20[=E2=80=A6]What is the difference between shared static this and the global=20 constructor ? Russell, if you use shared static this for dmd=20 does it work ? Laeeth.I had no idea what to put in a: shared static this { =E2=80=A6 } so I don't have one. --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 19 2015
On Saturday, 19 September 2015 at 10:45:22 UTC, Russel Winder
wrote:
Calling D from Python. I have two functions in D, compiled to a
shared object on Linux using LDC (but I get same problem using
DMD).
The sequential code:
extern(C)
double sequential(const int n, const double delta) {
Runtime.initialize();
const pi = 4.0 * delta * reduce!(
(double t, int i){ immutable x = (i - 0.5) * delta;
return t + 1.0 / (1.0 + x * x); })(
0.0, iota(1, n + 1));
Runtime.terminate();
return pi;
}
works entirely fine. However the "parallel" code:
extern(C)
double parallel(const int n, const double delta) {
Runtime.initialize();
const pi = 4.0 * delta * taskPool.reduce!"a + b"(
map!((int i){ immutable x = (i - 0.5) * delta; return
1.0 / (1.0 + x * x); })(iota(1, n + 1)));
Runtime.terminate();
return pi;
}
causes an immediate segfault (with LDC and DMD. I am assuming
that the problem is the lack of initialization of the
std.parallelism module and hence the use of taskPool is causing
a problem. I am betting I am missing something very simple
about module initialization, and that this is not actually a
bug.
Anyone any proposals?
Btw have you looked at Colvin's prettypyd ? It's a nicer way to
wrap things. Just pdef!() before functions, aggregates and
fields to wrap them.
For demos, I should also think that showing Python code in one
Jupyter cell calling D code in another is a pretty nice way to
show interop. Just need to install the pyd Magic. Your D code
can import dub libraries too.
Sep 19 2015
On Saturday, 19 September 2015 at 17:41:39 UTC, Laeeth Isharc wrote:On Saturday, 19 September 2015 at 10:45:22 UTC, Russel Winder wrote:In particular it should just work that way as PyD should deal with runtime initialization and the like. https://github.com/DlangScience/PydMagic https://github.com/DlangScience/PydMagic/blob/master/examples/test.ipynb https://github.com/John-Colvin/ppydCalling D from Python. I have two functions in D, compiled to a shared object on Linux using LDC (but I get same problem using DMD). The sequential code: extern(C) double sequential(const int n, const double delta) { Runtime.initialize(); const pi = 4.0 * delta * reduce!( (double t, int i){ immutable x = (i - 0.5) * delta; return t + 1.0 / (1.0 + x * x); })( 0.0, iota(1, n + 1)); Runtime.terminate(); return pi; } works entirely fine. However the "parallel" code: extern(C) double parallel(const int n, const double delta) { Runtime.initialize(); const pi = 4.0 * delta * taskPool.reduce!"a + b"( map!((int i){ immutable x = (i - 0.5) * delta; return 1.0 / (1.0 + x * x); })(iota(1, n + 1))); Runtime.terminate(); return pi; } causes an immediate segfault (with LDC and DMD. I am assuming that the problem is the lack of initialization of the std.parallelism module and hence the use of taskPool is causing a problem. I am betting I am missing something very simple about module initialization, and that this is not actually a bug. Anyone any proposals?Btw have you looked at Colvin's prettypyd ? It's a nicer way to wrap things. Just pdef!() before functions, aggregates and fields to wrap them. For demos, I should also think that showing Python code in one Jupyter cell calling D code in another is a pretty nice way to show interop. Just need to install the pyd Magic. Your D code can import dub libraries too.
Sep 19 2015
On Saturday, 19 September 2015 at 10:45:22 UTC, Russel Winder
wrote:
Calling D from Python. I have two functions in D, compiled to a
shared object on Linux using LDC (but I get same problem using
DMD).
The sequential code:
extern(C)
double sequential(const int n, const double delta) {
Runtime.initialize();
const pi = 4.0 * delta * reduce!(
(double t, int i){ immutable x = (i - 0.5) * delta;
return t + 1.0 / (1.0 + x * x); })(
0.0, iota(1, n + 1));
Runtime.terminate();
return pi;
}
works entirely fine. However the "parallel" code:
extern(C)
double parallel(const int n, const double delta) {
Runtime.initialize();
const pi = 4.0 * delta * taskPool.reduce!"a + b"(
map!((int i){ immutable x = (i - 0.5) * delta; return
1.0 / (1.0 + x * x); })(iota(1, n + 1)));
Runtime.terminate();
return pi;
}
causes an immediate segfault (with LDC and DMD. I am assuming
that the problem is the lack of initialization of the
std.parallelism module and hence the use of taskPool is causing
a problem. I am betting I am missing something very simple
about module initialization, and that this is not actually a
bug.
Anyone any proposals?
Isn't it simply that when doing some work asynchronously, as with
task pool, work is being done in other threads than the main. All
the while, you're killing the Runtime. The fact that when
removing Runtime.terminate makes the code work might support
that explanation.
Sep 20 2015
On Sun, 2015-09-20 at 07:49 +0000, rom via Digitalmars-d-learn wrote:=20[=E2=80=A6][=E2=80=A6]works entirely fine. However the "parallel" code: =20 extern(C) double parallel(const int n, const double delta) { Runtime.initialize(); const pi =3D 4.0 * delta * taskPool.reduce!"a + b"( map!((int i){ immutable x =3D (i - 0.5) * delta; return=20 1.0 / (1.0 + x * x); })(iota(1, n + 1))); Runtime.terminate(); return pi; } =20=20 Isn't it simply that when doing some work asynchronously, as with=20 task pool, work is being done in other threads than the main. All=20 the while, you're killing the Runtime. The fact that when=20 removing Runtime.terminate makes the code work might support=20 that explanation.I am not sure that can be the case per se as reduce only uses threads internally, the task pool operations should be complete by the time pi is initialized. On the other hand there might be a lock contention between a quiescent default task pool and the terminate code. The code is now moved on a bit from the above, and it works sufficient for the task at hand. This isn't production code so it doesn't have to be quite as good and correct as maybe it should be. =20 --=20 Russel. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D Dr Russel Winder t: +44 20 7585 2200 voip: sip:russel.winder ekiga.n= et 41 Buckmaster Road m: +44 7770 465 077 xmpp: russel winder.org.uk London SW11 1EN, UK w: www.russel.org.uk skype: russel_winder
Sep 22 2015