Re: [zeromq-dev] Testing platform

Hi Mikko, mikko.koppa...@gmail.com said: > the idea came up on IRC about providing a testing platform for > language bindings (automated builds, executing tests etc) on different > architectures and compilers. I am going to prototype the idea soon > with a couple virtual machines and would like to

Re: [zeromq-dev] disabling epoll()

I'd happily do it in C as a proof of concept if you like. I'm not up to changing the current poller class. On 4 Oct 2010 15:39, "Martin Lucina" wrote: > p...@imatix.com said: >> On Mon, Oct 4, 2010 at 3:23 PM, Martin Lucina wrote: >> >> >> Could you explain how the build currently decides to use

Re: [zeromq-dev] [PATCH] detect (select poll epoll_ctl kqueue) at configure time

p...@imatix.com said: > Well, I had to use the correct subject line for a patch... :) Sure, I just confused by a reply in the original thread talking about a patch I couldn't see right away :-) FWIW you can reply to the original thread with a new subject, thus keeping track of the conversation an

Re: [zeromq-dev] Any progress on 64-bit Windows build

Hi Steven Well it is good news about the CRT stuff. About the MIB_IFROW::wszName field isn't it just a question of how the WCHAR define ends up. If you expect a MBCS value and a string is UNICODE a conversion function (mbstowcs_s) is available AFAIK even with the /Zc:wchar_t- value set

Re: [zeromq-dev] Any progress on 64-bit Windows build

2010/9/23 Mikael Helbo Kjær > I was unable to make compatible OpenPGM DLLs for my stuff (I needed them > without buffer security and without wchar_t as a native type to be > compatible with Qt). > > Security enhanced CRT usage is now configurable with CONFIG_HAVE_SECURITY_ENHANCED_CRT. Native wi

Re: [zeromq-dev] Any progress on 64-bit Windows build

That matches my experience with zeromq on another matter as well so I believe that to be true Regards, Mikael On 4 October 2010 15:04, Steven McCoy mailto:steven.mc...@miru.hk> > wrote: http://msdn.microsoft.com/en-us/library/ms684497(VS.85).aspx Wouldn't lack

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

On Mon, Oct 4, 2010 at 9:51 AM, Kamil Shakirov wrote: > On Mon, Oct 4, 2010 at 2:50 PM, Kamil Shakirov wrote: >> Added taskvent.lisp, taskwork.lisp and tasksink.lisp examples. Updated >> build scripts. >> -- >> --ska > > Licensed under the MIT/X11 license. Thanks, Kamil. Lisp translations are n

Re: [zeromq-dev] Any progress on 64-bit Windows build

2010/10/4 Mikael Helbo Kjær > About the MIB_IFROW::wszName field isn’t it just a question of how the > WCHAR define ends up. If you expect a MBCS value and a string is UNICODE a > conversion function (mbstowcs_s) is available AFAIK even with the > /Zc:wchar_t- value set in the compiler no matter

Re: [zeromq-dev] Any progress on 64-bit Windows build

On 4 October 2010 15:04, Steven McCoy wrote: > http://msdn.microsoft.com/en-us/library/ms684497(VS.85).aspx > > Wouldn't lack of wchar_t imply the absence of all the related APIs? > > Ok, MSDN contradicting itself I guess, http://msdn.microsoft.com/en-us/library/dh8che7s(VS.80).aspx

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

At Mon, 4 Oct 2010 14:50:40 +0700, Kamil Shakirov wrote: > *.fas* > +*.*fsl > diff --git a/examples/Common Lisp/build b/examples/Common Lisp/build > index 8b415fd..95aa989 100755 > --- a/examples/Common Lisp/build > +++ b/examples/Common Lisp/build > @@ -10,7 +10,7 @@ if [ /$1/ = /a

Re: [zeromq-dev] Any progress on 64-bit Windows build

Hi Steven Hmm, I have no problem running code even without _NATIVE_WCHAR_T_DEFINED that uses the wchar_t type. I believe that there is a define like this in most relevant CRT files on windows (this was taken from a Windows SDK): #ifndef _WCHAR_T_DEFINED typedef unsigned short wchar_t; #

[zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

Added taskvent.lisp, taskwork.lisp and tasksink.lisp examples. Updated build scripts. -- --ska From e0632b7c40fe207df89dda40a5032cb2dbab293d Mon Sep 17 00:00:00 2001 From: Kamil Shakirov Date: Mon, 4 Oct 2010 14:39:41 +0700 Subject: [PATCH] Add taskvent.lisp, taskwork.lisp and tasksink.lisp exa

Re: [zeromq-dev] [PATCH] [CL binding] zmq:msg-data-as-string returns a Lisp string with a trailing null

At Fri, 01 Oct 2010 20:23:32 +0700, Kamil Shakirov wrote: > zmq:msg-data-as-string returns a Lisp string with a trailing null if used > with a > freshly created 0MQ message initialized by a Lisp string. I would expect it > to be > a normal Lisp string without trailing null by default so it could

Re: [zeromq-dev] disabling epoll()

On Mon, Oct 4, 2010 at 2:50 PM, Martin Lucina wrote: >> I'm submitting a patch that defines the correct macros.  You can use >> these in poller.cpp to decide which mechanism to use. > > We already do have these macros. What are you submitting, exactly? Where are those macros already? - Pieter _

Re: [zeromq-dev] Making language bindings work with both maint andmaster

At Sat, 02 Oct 2010 22:01:57 +0200, Martin Sustrik wrote: > > Why would this not work with some bindings? Are those not compiled using > > the C public API? > > Nope. The two I've mentioned redeclare the ABI of the library in a > native way. CFFI includes groveller which is able to generate lis

Re: [zeromq-dev] [PATCH] ZeroMQ CL binding fix for Mac OS X

At Thu, 30 Sep 2010 23:26:28 +0700, Kamil Shakirov wrote: > Fixed loading ZeroMQ library on Mac OS X (tested with Snow Leopard x86 > (10.6.4)). Applied. -- wbr, Vitaly ___ zeromq-dev mailing list zeromq-dev@lists.zeromq.org http://lists.zeromq.org/m

Re: [zeromq-dev] Making language bindings work with both maint andmaster

Hi all, > I think each language would have a preferred way to maintain such bindings, > and to tie the main library and the bindings like that may work for > one language, > but would cause severe pain for others. I'd vote for a much looser coupling > of between the main library and the bindings,

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

At Mon, 4 Oct 2010 17:41:40 +0700, Kamil Shakirov wrote: > What about other non-debian based OS distributions? Fedora, Gentoo and Archlinux are clc-capable as well :) > This is just temporary solution and could be changed with adding > more supported implementations. Ok. Just a hint that git wo

Re: [zeromq-dev] Any progress on 64-bit Windows build

Hi Steven, 2010/10/4 Steven McCoy : > Native wide character support is complicated, > MSVC forwards on _NATIVE_WCHAR_T_DEFINED > but without the conversion APIs data structures > that have defined WCHAR member variables are unusable? > I need a multi-byte string value of MIB_IFROW::wszName. I bel

Re: [zeromq-dev] Any progress on 64-bit Windows build

2010/10/4 Mikael Helbo Kjær > So that also explains why the code I write with wchar_t works. The Qt > thing is mostly an ABI issue I believe (they try to stay ABI and API > compatible across major versions and many many compilers/OSes). > > > The remaining issue is that Wine console & MinGW CRT

Re: [zeromq-dev] disabling epoll()

p...@imatix.com said: > On Mon, Oct 4, 2010 at 3:06 PM, Martin Lucina wrote: > > >> Where are those macros already? > > > > As Martin said, if you define ZMQ_FORCE_POLL or ZMQ_FORCE_SELECT during > > compilation of libzmq, this will force the zmq::poller_t and zmq_poll() to > > use a poll()- or s

Re: [zeromq-dev] [PATCH] detect (select poll epoll_ctl kqueue) at configure time

On 10/04/2010 03:36 PM, Martin Lucina wrote: > I can kind of see where you're going with this; you want to detect the > polling mechanism in a generic way based off the availability of a system > call rather than based off a check for a particular OS. This is how things > would normally be done wi

Re: [zeromq-dev] disabling epoll()

On 10/04/2010 03:09 PM, Martin Lucina wrote: >> What I recall Malo told me long time ago is that autoconf is able to >> compile and run a small user-supplied test program and defined the macro >> depending on whether it succeeds or fails. > > Yes, of course. But the point is this is only of any he

Re: [zeromq-dev] disabling epoll()

Martin, sust...@250bpm.com said: > > I found some comments about how zmq will use epoll "when available", which > > makes me hopeful their is some configure switch to turn it off. Or is this > > a lost cause? > > Just define ZMQ_FORCE_POLL or ZMQ_FORCE_SELECT macro, depending on what > polling s

Re: [zeromq-dev] disabling epoll()

On Mon, Oct 4, 2010 at 3:06 PM, Martin Lucina wrote: >> Where are those macros already? > > As Martin said, if you define ZMQ_FORCE_POLL or ZMQ_FORCE_SELECT during > compilation of libzmq, this will force the zmq::poller_t and zmq_poll() to > use a poll()- or select()- based implementation irresp

Re: [zeromq-dev] disabling epoll()

p...@imatix.com said: > On Mon, Oct 4, 2010 at 3:23 PM, Martin Lucina wrote: > > >> Could you explain how the build currently decides to use e/poll() or > >> select()? > > > > This is determined at compile time by configure which sets the relevant > > ZMQ_HAVE_XXX to detect the OS and the rest i

Re: [zeromq-dev] disabling epoll()

p...@imatix.com said: > On Fri, Oct 1, 2010 at 8:44 PM, Martin Sustrik wrote: > > > 0MQ devs: Wouldn't it be better if we checked for the availability of > > different polling mechanisms during the configure phase? > > I'm submitting a patch that defines the correct macros. You can use > these

Re: [zeromq-dev] disabling epoll()

On Mon, Oct 4, 2010 at 3:19 PM, Martin Sustrik wrote: > On 10/04/2010 03:09 PM, Martin Lucina wrote: > > >> What I recall Malo told me long time ago is that autoconf is able to > >> compile and run a small user-supplied test program and defined the macro > >> depending on whether it succeeds or f

Re: [zeromq-dev] [PATCH] detect (select poll epoll_ctl kqueue) at configure time

Well, I had to use the correct subject line for a patch... :) On 4 Oct 2010 15:37, "Martin Lucina" wrote: > Pieter, > > sorry, I got confused, you started a new thread which is why I didn't > realise this was related to the "disabling epoll()" thread. > > p...@imatix.com said: >> index f8d2e5e..3b

Re: [zeromq-dev] build zeromq-2.0.9 with mingw

Hi steve, Here is my patch, (not using PGM), against 2.0.9. I don't understand why the configure script always failed if I uncomment configure.in:321 PKG_CHECK_MODULES([GLIB], [glib-2.0 gthread-2.0]). Anyway it can patch zeromq-2.0.9 smoothly and build completely. Thanks anyway for your suggestio

Re: [zeromq-dev] [PATCH] [Issue 90] 0MQ and OpenPGM on FreeBSD 8.0

steven.mc...@miru.hk said: > Just testing here, 0MQ master and OpenPGM 5.0.78 on FreeBSD 8.0 works fine as > expected.  A few patches required to build, > > freebsd80-openpgm.patch > > Cannot apply to master:  Apparently libcompat is broken on FreeBSD 8.0, it's > not build with -fPIC and libcompa

Re: [zeromq-dev] [PATCH] detect (select poll epoll_ctl kqueue) at configure time

Pieter, sorry, I got confused, you started a new thread which is why I didn't realise this was related to the "disabling epoll()" thread. p...@imatix.com said: > index f8d2e5e..3b4fafb 100644 > --- a/configure.in > +++ b/configure.in > @@ -365,6 +365,7 @@ AC_SUBST(LIBZMQ_EXTRA_LDFLAGS) > # Check

Re: [zeromq-dev] Upgrade to OpenPGM 5.x (commit 00cd7d49c7f2b532b234)

steven.mc...@miru.hk said: > On 29 September 2010 00:43, Martin Lucina wrote: > > 1) > >         AC_DEFINE(ZMQ_HAVE_LINUX, 1, [Have Linux OS]) > +        AC_CHECK_LIB(rt, main) >         AC_CHECK_LIB(uuid, main, , > > What specifically introduces a dependency on librt? OpenP

Re: [zeromq-dev] Making language bindings work with both maint andmaster

I'm not sure why "master" and "maint" are featured here -- are we talking about the scenario where version X was released, but before version X + 1 is released and master still has the version number X ? If the above is correct, why don't zmq simply up the internal version number to X + 1 upon the

Re: [zeromq-dev] disabling epoll()

On Mon, Oct 4, 2010 at 3:23 PM, Martin Lucina wrote: >> Could you explain how the build currently decides to use e/poll() or >> select()? > > This is determined at compile time by configure which sets the relevant > ZMQ_HAVE_XXX to detect the OS and the rest is done in the ifdefs in > poller.hpp

Re: [zeromq-dev] build zeromq-2.0.9 with mingw

remove-tests.patch should have be named remove-tests~destructive.patch, Anyway feel free to implement a proper fix for the tests implementation. :D -- Steve-o On 3 October 2010 00:16, Steven McCoy wrote: > On 2 October 2010 13:11, Herry Sitepu wrote: > >> 2. make[2]: *** [libzmq_la-zmq.lo]

Re: [zeromq-dev] Making language bindings work with both maint andmaster

On Mon, Oct 4, 2010 at 7:29 AM, Daisuke Maki wrote: > Oh wait, or are you saying that currently there's no way to figure out the > version of zmq before compiling the bindings? > If that's the case I'd rather have a command line utility "zmq-config", > rather than a typedef so I can detect it even

Re: [zeromq-dev] Upgrade to OpenPGM 5.x (commit 00cd7d49c7f2b532b234)

sust...@250bpm.com said: > > Please also remove the check above the removed bits, thanks, since now that > > GLIB is not required by OpenPGM then pkg-config is not needed for the build > > at all. > > Well, you are a maintainer of the build system. Just do so. Done. -mato ___

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

On Mon, Oct 4, 2010 at 5:41 PM, Kamil Shakirov wrote: >>> diff --git a/examples/Common Lisp/build.lisp b/examples/Common >>> Lisp/build.lisp >>> index 5b7e89d..0eec089 100644 >>> --- a/examples/Common Lisp/build.lisp >>> +++ b/examples/Common Lisp/build.lisp >>> @@ -5,6 +5,12 @@ >>>  ;;; Kamil S

Re: [zeromq-dev] Making language bindings work with both maint andmaster

Yeah, I know, I let that slip by on purpose. I guess I didn't fool anyone ;) --d 2010/10/4 Mikko Koppanen : > On Mon, Oct 4, 2010 at 7:29 AM, Daisuke Maki wrote: >> Oh wait, or are you saying that currently there's no way to figure out the >> version of zmq before compiling the bindings? >> If t

[zeromq-dev] memory leak in PUB/SUB sockets

Hi Gonzalo, In answer to the Java bindings issue. I encountered the same problem as already mentioned in a precedent thread. I compiled the bindings without problems following README instructions. If I try the example in /perf: java -Djava.library.path=/virgoDev/zmq/v2r0p9/lib -classpath /virgoD

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

On Mon, Oct 4, 2010 at 2:50 PM, Kamil Shakirov wrote: > Added taskvent.lisp, taskwork.lisp and tasksink.lisp examples. Updated > build scripts. > > > -- > --ska > Licensed under the MIT/X11 license. -- --ska ___ zeromq-dev mailing list zeromq-dev@list

Re: [zeromq-dev] [PATCH] [CL binding] add missing 'pull' and 'push' socket types.

At Mon, 4 Oct 2010 03:56:24 +0700, Kamil Shakirov wrote: > Rename deprecated 'upstream' and 'downstream' socket types to 'pull' and > 'push'. Applied. -- wbr, Vitaly ___ zeromq-dev mailing list zeromq-dev@lists.zeromq.org http://lists.zeromq.org/mai

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

At Mon, 4 Oct 2010 17:59:27 +0700, Kamil Shakirov wrote: > > Sure, system definition would be better. Here I just followed the same > > 'compile' script way. Clozure CL worked here as well. > > > > Could it be better just build only one executable image from system > definition including all exam

Re: [zeromq-dev] disabling epoll()

p...@imatix.com said: > On Mon, Oct 4, 2010 at 2:50 PM, Martin Lucina wrote: > > >> I'm submitting a patch that defines the correct macros.  You can use > >> these in poller.cpp to decide which mechanism to use. > > > > We already do have these macros. What are you submitting, exactly? > > Where

Re: [zeromq-dev] disabling epoll()

On 10/04/2010 02:49 PM, Martin Lucina wrote: >> 0MQ devs: Wouldn't it be better if we checked for the availability of >> different polling mechanisms during the configure phase? > > Not possible. configure can only determine what your target C library is, > it knows nothing about your target kerne

Re: [zeromq-dev] disabling epoll()

sust...@250bpm.com said: > On 10/04/2010 02:49 PM, Martin Lucina wrote: > > >> 0MQ devs: Wouldn't it be better if we checked for the availability of > >> different polling mechanisms during the configure phase? > > > > Not possible. configure can only determine what your target C library is, > > i

Re: [zeromq-dev] Making language bindings work with both maintandmaster

> With version id it would be the following (major,three digits for > minor,three digits for patch) > > #if ZMQ_VERSION_ID >= 2001001 > > Where as with ZMQ_MAJOR/MINOR/PATCH it would be: > > #if ZMQ_MAJOR > 2 || (ZMQ_MAJOR == 2 && ZMQ_MINOR >= 1 && ZMQ_PATCH >= 1) > > I guess this is a matter o

Re: [zeromq-dev] Website changes

Hi Pieter, p...@imatix.com said: > I've made a number of cleanups to the zeromq.org website. > > * The left nav menu is cleaner and the longer lists are now collapsible It's not at all clear visually which lists are collapsible; e.g. "Main" is red but not collapsible, "ØMQ Community" is black bu

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

Hi Vitaly, On Mon, Oct 4, 2010 at 5:10 PM, Vitaly Mayatskikh wrote: > At Mon, 4 Oct 2010 14:50:40 +0700, Kamil Shakirov wrote: > >>  *.fas* >> +*.*fsl >> diff --git a/examples/Common Lisp/build b/examples/Common Lisp/build >> index 8b415fd..95aa989 100755 >> --- a/examples/Common Lisp/build >> ++

Re: [zeromq-dev] Types for options in [gs]etsockopt()

> > In short, a simple (but perhaps incorrect) proposal would be to have all > > of them be unsigned 64 bits, using uint64_t. > > Totally agree. > > > When we reach consensus, I can do the changes to the code. > > This is already on the roadmap for 3.0: http://www.zeromq.org/docs:3-0 > > Since

Re: [zeromq-dev] Types for options in [gs]etsockopt()

On 10/04/2010 04:55 PM, gonzalo diethelm wrote: >> Since it's an API change it would have to be done for 3.0 unless >> there's consensus from binding maintainers that this counts as a >> 'minor change' in which case it could go into 2.1 now. >> >> I'd see this going in 2.2.0. It will cause warnin

Re: [zeromq-dev] Making language bindings work with both maint and master

Hi, sust...@250bpm.com said: > We have to think of a way to make language bindings work with both maint > and master. I don't really see a problem with actual *work* involved all that much. Given that all the official bindings are designed to be really thin wrappers and that 0MQ 2.0.x is in deep

Re: [zeromq-dev] Making language bindings work with both maint andmaster

sust...@250bpm.com said: > Hi all, > > > I think each language would have a preferred way to maintain such bindings, > > and to tie the main library and the bindings like that may work for > > one language, > > but would cause severe pain for others. I'd vote for a much looser coupling > > of betw

Re: [zeromq-dev] memory leak in PUB/SUB sockets

> In answer to the Java bindings issue. I encountered the same problem as > already mentioned in a precedent thread. > I compiled the bindings without problems following README instructions. > If I try the example in /perf: > java -Djava.library.path=/virgoDev/zmq/v2r0p9/lib -classpath > /virgoDev/

Re: [zeromq-dev] Website changes

The colors depend only on whether or not the text is a link. What colors would you suggest for collapsing lists? I will try adding a visual indicator. Feel free to propose improved css. -Pieter On 4 Oct 2010 16:34, "Martin Lucina" wrote: > Hi Pieter, > > p...@imatix.com said: >> I've made a numbe

Re: [zeromq-dev] Making language bindings work with both maint andmaster

m...@kotelna.sk said: > The only thing I'm not sure of is what is the best kind of version macro(s) > to define... I'll have a think about it and look at the various options; > will get back to this tomorrow. Having thought about this a bit more and looked at the various proposals on this thread I

Re: [zeromq-dev] memory leak in PUB/SUB sockets

gonzalo diethelm wrote: >> In answer to the Java bindings issue. I encountered the same problem >> > as > >> already mentioned in a precedent thread. >> I compiled the bindings without problems following README >> > instructions. > >> If I try the example in /perf: >> java -Djava.li

[zeromq-dev] Changes to dev build process of pyzmq

Hi, This might affect a number of you who are using pyzmq from github. The .c files of pyzm are autogenerated by Cython. We used to include these autogenerated .c files in the github repo, but because they change often and in significant ways, the pyzmq repo was growing faster than we want. From

Re: [zeromq-dev] memory leak in PUB/SUB sockets

> > Wouldn't this point to the fact that you are not finding the shared > > library (native code) that underlies the Java binding? Where are both > > your libzmq.so and libjzmq.so libraries located? Are both those > > locations in your LD_LIBRARY_PATH or in any place that can be found by > > the dy

Re: [zeromq-dev] Making language bindings work with both maint andmaster

At Mon, 4 Oct 2010 17:28:47 +0200, Martin Lucina wrote: > Having thought about this a bit more and looked at the various proposals on > this thread I'd like to propose the following syntax for version macros: > > 1) zmq.h would define ZMQ_VERSION to be some magic (opaque) constant that > incorpor

Re: [zeromq-dev] [PATCH] detect (select poll epoll_ctl kqueue) at configure time

Mato, I wasn't sure that would work. Email clients have various ways of detecting threads. Anyhow, I'm happy to make a PoC for runtime detection of poll/select using these HAVE_FCTNAME macros and some system dependencies, but it'd be in C, not C++. In any case it'd require some abstraction arou

Re: [zeromq-dev] Making language bindings work with bothmaint andmaster

> 1) zmq.h would define ZMQ_VERSION to be some magic (opaque) constant that > incorporates the major, minor and patch versions; the constants should be > arithmetically comparable; i.e. something like what Gonzalo proposed. > > 2) it would also define a ZMQ_MAKE_VERSION(major,minor,patch) macro th

Re: [zeromq-dev] memory leak in PUB/SUB sockets

gonzalo diethelm wrote: >> In answer to the Java bindings issue. I encountered the same problem >> > as > >> already mentioned in a precedent thread. >> I compiled the bindings without problems following README >> > instructions. > >> If I try the example in /perf: >> java -Djava.li

Re: [zeromq-dev] memory leak in PUB/SUB sockets

> Well, does this apply to you? This is rather weird, I compile & use the > Java binding on a daily basis (on Windows XP) and I have seen people on > this list using it from other platforms, and they don't report having > this problem. Are you in any way modifying the Java binding prior to its > >

Re: [zeromq-dev] Making language bindings work withbothmaint andmaster

> Yes, this is even nicer (I think). In short: > > // These are the ONLY numbers you have to manually maintain: > #define ZMQ_VERSION_MAJOR 2 > #define ZMQ_VERSION_MINOR 0 > #define ZMQ_VERSION_PATCH 1 Something else: these numbers should be defined in a single place. From a cursory look at the c

Re: [zeromq-dev] [PATCH] zguide: taskvent.lisp, taskwork.lisp and tasksink.lisp

On Mon, Oct 4, 2010 at 6:02 PM, Vitaly Mayatskikh wrote: > At Mon, 4 Oct 2010 17:59:27 +0700, Kamil Shakirov wrote: > >> > Sure, system definition would be better. Here I just followed the same >> > 'compile' script way. Clozure CL worked here as well. >> > >> >> Could it be better just build only

Re: [zeromq-dev] Changes to dev build process of pyzmq

On Mon, Oct 4, 2010 at 08:34, Brian Granger wrote: > Hi, > > This might affect a number of you who are using pyzmq from github. > The .c files of pyzm are autogenerated by Cython. We used to include > these autogenerated .c files in the github repo, but because they > change often and in signifi

Re: [zeromq-dev] using 0MQ in a MMORPG online game.

kasicass said the following on 9/3/2010 8:56 AM: > Aha, your solution is great. I will try. > > I've use libevent for it right now, and it works perfectly. > You may want to look at Jenkins Softwares' RakNet http://www.jenkinssoftware.com/ - he's done much of the work necessary for the client-f

[zeromq-dev] [PATCH] Add VERSION macros

Hi all, Here's the patch to maint to allow to determine 0MQ version on compile-time. Review appreciated. Thanks! Martin >From 8ae5fbf68f9ed0594a5f3e78d3c4227f76e0acac Mon Sep 17 00:00:00 2001 From: Martin Sustrik Date: Mon, 4 Oct 2010 21:14:07 +0200 Subject: [PATCH] add VERSION macros --- bu

[zeromq-dev] 0MQ crash when peer disconnects

Hi Martin, I've found a reproducible crash caused by closing a peer connection. I've posted a test case. This affects 2.0.10 and 2.1.0 afaics. http://github.com/zeromq/zeromq2/issues/issue/94 Cheers Pieter ___ zeromq-dev mailing list zeromq-dev@lists.

Re: [zeromq-dev] multiple bind with one socket

Hi Martin, > > So we have 11000 bytes transferred in both cases. > > > > What's the flow in the above reasoning? > > Err, you caught me being a doink.. :) I'm mixing up some of the > historic systems I've used and not sticking to the raw tcp stream portion > of > this. It's tough to keep

Re: [zeromq-dev] [PATCH] Add VERSION macros

> Here's the patch to maint to allow to determine 0MQ version on compile- > time. > > Review appreciated. Looks good to me. A few remarks: 0. You don't really need the PACKAGE_VERSION_XXX macros? They looked (to this noob) like a standard thing... 1. I would change "Place any MSVC-specific defi

Re: [zeromq-dev] using 0MQ in a MMORPG online game.

I used to work with the author and can say that RakNet is pretty damned nice and he's quite clever. I'll just point out that RakNet previously did not scale as nicely as you would hope for, but it is definitely closer than most. I've not looked at it in quite a while though, so maybe he's fixed t

Re: [zeromq-dev] multiple bind with one socket

Kelly, > So, I went ahead and spent a bit of time doing an experiment and > making sure I was not mixing up old items with 0mq.. As mentioned, one of > my tests was pushing about 200mb's a second which is unacceptable. I > switched off the no_delay options on the sockets and that cut the b

Re: [zeromq-dev] [PATCH] Add VERSION macros

> Looks good to me. A few remarks: And one more thing: it would be great if these changes were applied to both maint and master, so that the bindings could be modified right away to more intelligently check for the library version of 0MQ. If this is the plan, what exactly will be the version numbe

Re: [zeromq-dev] multiple bind with one socket

Hi Martin, > Kelly, > > > So, I went ahead and spent a bit of time doing an experiment and > > making sure I was not mixing up old items with 0mq.. As mentioned, one > of > > my tests was pushing about 200mb's a second which is unacceptable. I > > switched off the no_delay options on the so

[zeromq-dev] Implementing failover in client side

Hi, I'm trying to understand the proper way to implement failover in a ZMQ requester (not just load balancing). Let me expose the scenario I want to achieve: 1) An application processes SIP requests (VoIP) in a single process. 2) The application has a ZMQ REQ socket connected to N servers (ZMQ R

Re: [zeromq-dev] multiple bind with one socket

Martin, Btw, typo in the data, sorry just noticed it: Initial conversion: 180mb/s rms, 22mb/s deviation. 42ms latency rms, 17ms deviation. Nagle enabled: 151mb/s rms, 35mb/s deviation. >>> 43ms latency rms, 14ms deviation. Single connection: 72mb/s rms, 5mb/s deviation. 31ms la

Re: [zeromq-dev] odd zmqstream.py behavior

Hi, Min and I looked at this today and this is a bug in pyzmq. When a callback in the eventloop raises, we are currently closing the socket, which we should not do. We are still figuring out the error handling symantics, but we should have a fix in place soon in pyzmq master. Cheers, Brian On

[zeromq-dev] 0MQ and Firewalls

Is there a officially defined position with regards to 0MQ's ability to work in a client/server architecture where the client might be behind a firewall/NAT-router? If so, where can I find this documented? In this scenario, what is the support for: 1. TCP transport REQ/REP 2. Multi-

Re: [zeromq-dev] Implementing failover in client side

Hi Brian, I've not yet covered reliable request-reply / failover in the user guide yet. 0MQ doesn't do it directly. I'd probably look at using XREQ rather than REQ, so you can keep the socket open, and possibly even XREP so you can explicitly address the different servers. Look at the custom lo

Re: [zeromq-dev] 0MQ and Firewalls

On Tue, Oct 5, 2010 at 4:01 AM, John McLaughlin - PTIR wrote: > Is there a officially defined position with regards to 0MQ’s ability to work > in a client/server architecture where the client might be behind a > firewall/NAT-router?  If so, where can I find this documented? The connectivity guar

Re: [zeromq-dev] 0MQ and Firewalls

On 5 October 2010 10:01, John McLaughlin - PTIR wrote: > My assumptions are: > >1. Further, I assume #5 above, PGM (server publishing) would not work >through a firewall. > > PGM can work fine through a configured firewall, if you really would want to is another question. The firewall w

Re: [zeromq-dev] [PATCH] Add VERSION macros

Hi Gonzalo, > 0. You don't really need the PACKAGE_VERSION_XXX macros? They looked (to > this noob) like a standard thing... Dunno. They were generated by hand. > 1. I would change "Place any MSVC-specific definitions here" into "Place > any Windows-specific definitions here". The file in quest

Re: [zeromq-dev] [PATCH] Add VERSION macros

On 10/04/2010 10:35 PM, gonzalo diethelm wrote: >> Looks good to me. A few remarks: > > And one more thing: it would be great if these changes were applied to > both maint and master, so that the bindings could be modified right away > to more intelligently check for the library version of 0MQ. If