[... From pike@roxen.com mailing list, ]
Le lundi, 21 avr 2003, à 21:32 Europe/Paris, Martin Nilsson a écrit :
Marek Habersack wrote:
They are not up to date since there's no interest in them upstream and maintaining them is too much effort if it is to be futile.
The non-interest of the patches was not due to lack of interest in IPv6 support, but the fact that it didn't compile on many systems. But that isn't news to you...
I understand that, but you know DVB module in Pike 7.4/7.5 is only Linux specific, so "it didn't compile on many systems" is not really a good argument to the fact that IPv6 support was not accepted.
The idea is to start a IPv6 support on at least one Operating System, then trying to port on other depending of access on other system.
I really think that IPv6 (even if it is not complete) must be added on pike.
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
My 0,02€
/Xavier
I understand that, but you know DVB module in Pike 7.4/7.5 is only Linux specific, so "it didn't compile on many systems" is not really a good argument to the fact that IPv6 support was not accepted.
What's the DVB module?
IPv6 support ought to be on the same modules where IPv4 is today, that must be Stdio.File aka _files.Fd (or so?).
The idea is to start a IPv6 support on at least one Operating System, then trying to port on other depending of access on other system.
Does anyone has a primer on IPv6 and how to setup an IPv6 network? There isn't any chance that my ISP can talk IPv6, is there?
/ Mirar
Previous text:
2003-04-22 12:43: Subject: Re: [caudium-general] ipv6 support
[... From pike@roxen.com mailing list, ]
Le lundi, 21 avr 2003, à 21:32 Europe/Paris, Martin Nilsson a écrit :
Marek Habersack wrote:
They are not up to date since there's no interest in them upstream and maintaining them is too much effort if it is to be futile.
The non-interest of the patches was not due to lack of interest in IPv6 support, but the fact that it didn't compile on many systems. But that isn't news to you...
I understand that, but you know DVB module in Pike 7.4/7.5 is only Linux specific, so "it didn't compile on many systems" is not really a good argument to the fact that IPv6 support was not accepted.
The idea is to start a IPv6 support on at least one Operating System, then trying to port on other depending of access on other system.
I really think that IPv6 (even if it is not complete) must be added on pike.
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
My 0,02
/Xavier
/ Brevbäraren
On Tue, Apr 22, 2003 at 01:25:02PM +0200, Mirar @ Pike developers forum wrote:
There isn't any chance that my ISP can talk IPv6, is there?
not really, but there is no need. the general way today is to set up a tunnel from an existing ipv6 server to yours.
the best course of action would be to find out who in your vincinity is running ipv6 and then connect to that. as this person will him/herselfhave set up such a tunnel (s)he will know how the tunnel is made. i am pretty sure you should find someone at liu at least.
greetings, martin.
It's possible to set up more-or-less private tunnels of IPv6 over IPv4. We could try to set up our own virtual LAN. But we probably really need a good howto to do that in a reasonable time-frame.
/ Per Hedbor ()
Previous text:
2003-04-22 13:21: Subject: Re: [caudium-general] ipv6 support
I understand that, but you know DVB module in Pike 7.4/7.5 is only Linux specific, so "it didn't compile on many systems" is not really a good argument to the fact that IPv6 support was not accepted.
What's the DVB module?
IPv6 support ought to be on the same modules where IPv4 is today, that must be Stdio.File aka _files.Fd (or so?).
The idea is to start a IPv6 support on at least one Operating System, then trying to port on other depending of access on other system.
Does anyone has a primer on IPv6 and how to setup an IPv6 network? There isn't any chance that my ISP can talk IPv6, is there?
/ Mirar
Le mardi, 22 avr 2003, à 13:25 Europe/Paris, Mirar @ Pike developers forum a écrit :
I understand that, but you know DVB module in Pike 7.4/7.5 is only Linux specific, so "it didn't compile on many systems" is not really a good argument to the fact that IPv6 support was not accepted.
What's the DVB module?
Something about Video / DVD under Linux. I don't know what it is since I don't run Linux.
IPv6 support ought to be on the same modules where IPv4 is today, that must be Stdio.File aka _files.Fd (or so?).
I'd really like... The patch from Marek is, I think, like this.
The idea is to start a IPv6 support on at least one Operating System, then trying to port on other depending of access on other system.
Does anyone has a primer on IPv6 and how to setup an IPv6 network? There isn't any chance that my ISP can talk IPv6, is there?
You can have a Free tunnel (IPv6 over IPv4) tunne with a /64 netmask using http://www.freenet6.net/. It is really free and the most easiest way to have a IPv6 connectivity for developping and starting to understand how IPv6 works.
/Xavier
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
Incompleteness was not an issue. Pike didn't compile at all on about half of the systems in AutoBuild (which was the verification system at the time). It is of course fully acceptable that while merging a patch into the development tree the builds are red for a few days. Marek however said that he had no intention in trying to fix his checkin, so it was reverted after half a week.
/ Martin Nilsson (har bott i google)
Previous text:
2003-04-22 12:43: Subject: Re: [caudium-general] ipv6 support
[... From pike@roxen.com mailing list, ]
Le lundi, 21 avr 2003, à 21:32 Europe/Paris, Martin Nilsson a écrit :
Marek Habersack wrote:
They are not up to date since there's no interest in them upstream and maintaining them is too much effort if it is to be futile.
The non-interest of the patches was not due to lack of interest in IPv6 support, but the fact that it didn't compile on many systems. But that isn't news to you...
I understand that, but you know DVB module in Pike 7.4/7.5 is only Linux specific, so "it didn't compile on many systems" is not really a good argument to the fact that IPv6 support was not accepted.
The idea is to start a IPv6 support on at least one Operating System, then trying to port on other depending of access on other system.
I really think that IPv6 (even if it is not complete) must be added on pike.
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
My 0,02
/Xavier
/ Brevbäraren
There were som issues with finding ipv6 drivers for Win* IIRC.
/ Peter Lundqvist (disjunkt)
Previous text:
2003-04-22 13:32: Subject: Re: [caudium-general] ipv6 support
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
Incompleteness was not an issue. Pike didn't compile at all on about half of the systems in AutoBuild (which was the verification system at the time). It is of course fully acceptable that while merging a patch into the development tree the builds are red for a few days. Marek however said that he had no intention in trying to fix his checkin, so it was reverted after half a week.
/ Martin Nilsson (har bott i google)
According to URL:http://www.microsoft.com/windowsserver2003/technologies/ipv6/default.mspx XP should already contain what is needed, so it doesn't seem to be an issue anymore. Maybe you have to do something slightly creative to avoid a dependency on the "getaddrinfo" symbol on systems that doesn't have it, but GetProcAddress can typically be used to handle such eventualities.
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
Previous text:
2003-04-22 13:48: Subject: Re: [caudium-general] ipv6 support
There were som issues with finding ipv6 drivers for Win* IIRC.
/ Peter Lundqvist (disjunkt)
On Tue, Apr 22, 2003 at 01:35:01PM +0200, Martin Nilsson (har bott i google) @ Pike (-) developers forum scribbled:
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
Incompleteness was not an issue. Pike didn't compile at all on about half of the systems in AutoBuild (which was the verification system at the time). It is of course fully acceptable that while merging a patch into the development tree the builds are red for a few days. Marek however said that he had no intention in trying to fix his checkin, so it was reverted after half a week.
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
marek
Would you like to do another attempt? Since you have the patches almost ready, it would be least work. Adding the necessary configure tests shouldn't be that much work today, with the support of the build logs in xenofarm.
/ Mirar
Previous text:
2003-04-23 14:03: Subject: Re: [caudium-general] ipv6 support
On Tue, Apr 22, 2003 at 01:35:01PM +0200, Martin Nilsson (har bott i google) @ Pike (-) developers forum scribbled:
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
Incompleteness was not an issue. Pike didn't compile at all on about half of the systems in AutoBuild (which was the verification system at the time). It is of course fully acceptable that while merging a patch into the development tree the builds are red for a few days. Marek however said that he had no intention in trying to fix his checkin, so it was reverted after half a week.
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
marek
/ Brevbäraren
On Wed, Apr 23, 2003 at 02:15:02PM +0200, Mirar @ Pike developers forum scribbled:
Would you like to do another attempt? Since you have the patches
I'm rediffing the patch right now. Will try to post it tonight.
almost ready, it would be least work. Adding the necessary configure tests shouldn't be that much work today, with the support of the build logs in xenofarm.
Yes, I suppose it shouldn't be hard.
marek
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
Well, you have access to the same tools as all other Pike developers (CVS, pikefarm and MSDN).
/ Martin Nilsson (har bott i google)
Previous text:
2003-04-23 14:03: Subject: Re: [caudium-general] ipv6 support
On Tue, Apr 22, 2003 at 01:35:01PM +0200, Martin Nilsson (har bott i google) @ Pike (-) developers forum scribbled:
Even if Marek work is incomplete, it is really a good start to finish it on other OS...
Incompleteness was not an issue. Pike didn't compile at all on about half of the systems in AutoBuild (which was the verification system at the time). It is of course fully acceptable that while merging a patch into the development tree the builds are red for a few days. Marek however said that he had no intention in trying to fix his checkin, so it was reverted after half a week.
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
marek
/ Brevbäraren
On Wed, Apr 23, 2003 at 02:25:02PM +0200, Martin Nilsson (har bott i google) @ Pike (-) developers forum scribbled:
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
Well, you have access to the same tools as all other Pike developers (CVS, pikefarm and MSDN).
Maybe you have missed the 'had' above. Besides, I have no interest in working on Windows in any way - but that's what the 'community development' model is for, isn't it? Somebody with the tools, the interest and experience with NT will, hopefully, fix where I failed.
marek
Well, before pikefarm there was autobuild, so all the tools needed to create a working configure scripts was already there. I have never said nor can recall anyone else having said that it should work on all systems. It must not break already working functionality on any system though.
/ Martin Nilsson (har bott i google)
Previous text:
2003-04-23 14:37: Subject: Re: [caudium-general] ipv6 support
On Wed, Apr 23, 2003 at 02:25:02PM +0200, Martin Nilsson (har bott i google) @ Pike (-) developers forum scribbled:
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
Well, you have access to the same tools as all other Pike developers (CVS, pikefarm and MSDN).
Maybe you have missed the 'had' above. Besides, I have no interest in working on Windows in any way - but that's what the 'community development' model is for, isn't it? Somebody with the tools, the interest and experience with NT will, hopefully, fix where I failed.
marek
/ Brevbäraren
On Wed, Apr 23, 2003 at 02:45:04PM +0200, Martin Nilsson (har bott i google) @ Pike (-) developers forum scribbled:
Well, before pikefarm there was autobuild, so all the tools needed to create a working configure scripts was already there. I have never said nor can recall anyone else having said that it should work on all systems. It must not break already working functionality on any system though.
And which system was that?
marek
But if you check in code that breaks on a platform that you can't fix it for, make a configuretest to make sure that your code doesn't break the build on that platform!
/ Marcus Agehall (Trådlös)
Previous text:
2003-04-23 14:37: Subject: Re: [caudium-general] ipv6 support
On Wed, Apr 23, 2003 at 02:25:02PM +0200, Martin Nilsson (har bott i google) @ Pike (-) developers forum scribbled:
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
Well, you have access to the same tools as all other Pike developers (CVS, pikefarm and MSDN).
Maybe you have missed the 'had' above. Besides, I have no interest in working on Windows in any way - but that's what the 'community development' model is for, isn't it? Somebody with the tools, the interest and experience with NT will, hopefully, fix where I failed.
marek
/ Brevbäraren
On Wed, Apr 23, 2003 at 11:40:01PM +0200, Marcus Agehall (Trådlös) @ Pike (-) developers forum scribbled:
But if you check in code that breaks on a platform that you can't fix it for, make a configuretest to make sure that your code doesn't break the build on that platform!
And if I don't know why the test breaks or what to test for on the given platform? If you are a Windows coder, then when I tell you to test for the different sizes of sockaddr structures specific for different socket families or whether /dev/epoll is available and if so, whether it is Linux or Solaris, what will you do? You will call me a bully. And you will be right.
marek
p.s. no need to yell, really.
If you change something that works on FOO but breaks on BAR and you can't fix it, make sure that your configuretest *only* allows your code to be compiled on FOO-machines. That is how I would do it.
/ Marcus Agehall (Trådlös)
Previous text:
2003-04-24 00:12: Subject: Re: [caudium-general] ipv6 support
On Wed, Apr 23, 2003 at 11:40:01PM +0200, Marcus Agehall (Trådlös) @ Pike (-) developers forum scribbled:
But if you check in code that breaks on a platform that you can't fix it for, make a configuretest to make sure that your code doesn't break the build on that platform!
And if I don't know why the test breaks or what to test for on the given platform? If you are a Windows coder, then when I tell you to test for the different sizes of sockaddr structures specific for different socket families or whether /dev/epoll is available and if so, whether it is Linux or Solaris, what will you do? You will call me a bully. And you will be right.
marek
p.s. no need to yell, really.
/ Brevbäraren
On Thu, Apr 24, 2003 at 12:30:05AM +0200, Marcus Agehall (Trådlös) @ Pike (-) developers forum scribbled:
If you change something that works on FOO but breaks on BAR and you can't fix it, make sure that your configuretest *only* allows your code to be compiled on FOO-machines. That is how I would do it.
You're welcome to do it then. Once again, I'm sorry I did anything. I should have sit down on my butt and not waste time on writing even a line of code. Next time I will think 5 times before saying a word.
marek
Could you please tell me what, in the following line of reasoning, that you feel is fucked up:
1. It should be possible to compile Pike. 2. Thus changes in the Pike source code should not break its abilities to compile (on a system on which it already compiles). 3. Thus changes that breaks Pikes abilities to compile should be fixed or reverted. 4. Thus if the person that made the changes that breaks Pikes abilities to compile is not willing to fix them (i.e. take on responsibility for making such a fix, e.g. by asking someone who knows anything about the system on which the compiling ability of Pike ceased) the changes should be reverted.
/ Martin Nilsson (har bott i google)
Previous text:
2003-04-24 00:37: Subject: Re: [caudium-general] ipv6 support
On Thu, Apr 24, 2003 at 12:30:05AM +0200, Marcus Agehall (Trådlös) @ Pike (-) developers forum scribbled:
If you change something that works on FOO but breaks on BAR and you can't fix it, make sure that your configuretest *only* allows your code to be compiled on FOO-machines. That is how I would do it.
You're welcome to do it then. Once again, I'm sorry I did anything. I should have sit down on my butt and not waste time on writing even a line of code. Next time I will think 5 times before saying a word.
marek
/ Brevbäraren
On Thu, 24 Apr 2003, Martin Nilsson (har bott i google) @ Pike (-) developers forum wrote:
Please do not feed the troll. This will only lead to Marek and maybe others not contributing to Pike and other to feel angry.
For my point of view it looks everybody has made mistakes in this and everybody has some good reasons. Maybe some more than others, that's not the problem. I quite agree with the lack of communication problem, you sometimes have to speak with "soft" words when speaking to humans even if there are coders and are wrong from your point of view. Maybe some more than others. Just keep cool and try to understand the other poeple point of view. If everybody does some compromises it can be fixed.
-- David Gourdelier
Could you please tell me what, in the following line of reasoning, that you feel is fucked up:
- It should be possible to compile Pike.
- Thus changes in the Pike source code should not break its abilities to compile (on a system on which it already compiles).
- Thus changes that breaks Pikes abilities to compile should be fixed or reverted.
- Thus if the person that made the changes that breaks Pikes abilities to compile is not willing to fix them (i.e. take on responsibility for making such a fix, e.g. by asking someone who knows anything about the system on which the compiling ability of Pike ceased) the changes should be reverted.
/ Martin Nilsson (har bott i google)
Previous text:
2003-04-24 00:37: Subject: Re: [caudium-general] ipv6 support
On Thu, Apr 24, 2003 at 12:30:05AM +0200, Marcus Agehall (Trådlös) @ Pike (-) developers forum scribbled:
If you change something that works on FOO but breaks on BAR and you can't fix it, make sure that your configuretest *only* allows your code to be compiled on FOO-machines. That is how I would do it.
You're welcome to do it then. Once again, I'm sorry I did anything. I should have sit down on my butt and not waste time on writing even a line of code. Next time I will think 5 times before saying a word.
marek
Le mercredi, 23 avr 2003, à 23:40 Europe/Paris, Marcus Agehall (Trådlös) @ Pike (-) developers forum a écrit :
But if you check in code that breaks on a platform that you can't fix it for, make a configuretest to make sure that your code doesn't break the build on that platform!
Humm... Interressing... But how to make a make configuretest on a platform we don't any access (eg WinNT or whatever exotic OS ?).
AFAIK 7.5 branch is -devel version, eg version we can break for a small time when we try to add some new feature in the core... Or I didn't understood what is a devel version for ?
My 0.02€...
/Xavier
On Thu, Apr 24, 2003 at 12:16:29AM +0200, Xavier Beaudouin scribbled:
Le mercredi, 23 avr 2003, à 23:40 Europe/Paris, Marcus Agehall (Trådlös) @ Pike (-) developers forum a écrit :
But if you check in code that breaks on a platform that you can't fix it for, make a configuretest to make sure that your code doesn't break the build on that platform!
Humm... Interressing... But how to make a make configuretest on a platform we don't any access (eg WinNT or whatever exotic OS ?).
AFAIK 7.5 branch is -devel version, eg version we can break for a small time when we try to add some new feature in the core... Or I didn't understood what is a devel version for ?
And I, on the other hand, assume that a community effort is where people work in common to achieve a certain goal. If that definition is right, then it is reasonable to assume that others will help with the stuff you cannot do instead of stepping up and pointing fingers or yelling. But maybe both Xavier and I are wrong (quite possibly so).
marek
We have sort of ended up in a feedback loop of bad vibes, here. Last time ipv6 was hacked at, we ended up nowhere, probably as a result of lack of communication. We don't have the logs to prove what went wrong how, but let's try again, pursue whatever problems show up in pikefarm and see where it gets us.
/ Johan Sundström (folkskådare)
Previous text:
2003-04-24 00:20: Subject: Re: [caudium-general] ipv6 support
On Thu, Apr 24, 2003 at 12:16:29AM +0200, Xavier Beaudouin scribbled:
Le mercredi, 23 avr 2003, à 23:40 Europe/Paris, Marcus Agehall (Trådlös) @ Pike (-) developers forum a écrit :
But if you check in code that breaks on a platform that you can't fix it for, make a configuretest to make sure that your code doesn't break the build on that platform!
Humm... Interressing... But how to make a make configuretest on a platform we don't any access (eg WinNT or whatever exotic OS ?).
AFAIK 7.5 branch is -devel version, eg version we can break for a small time when we try to add some new feature in the core... Or I didn't understood what is a devel version for ?
And I, on the other hand, assume that a community effort is where people work in common to achieve a certain goal. If that definition is right, then it is reasonable to assume that others will help with the stuff you cannot do instead of stepping up and pointing fingers or yelling. But maybe both Xavier and I are wrong (quite possibly so).
marek
/ Brevbäraren
On Thu, Apr 24, 2003 at 12:35:03AM +0200, Johan Sundström (folkskådare) @ Pike (-) developers forum scribbled:
We have sort of ended up in a feedback loop of bad vibes, here. Last time ipv6 was hacked at, we ended up nowhere, probably as a result of lack of communication. We don't have the logs to prove what went wrong how, but let's try again, pursue whatever problems show up in pikefarm and see where it gets us.
I can't agree with you more, Johan, but it's so fucking hard (pardon my french) to communicate with some folks around here in simple matters, that it's pretty much discouraging. Yelling, whining and bitching about as simple a matter as adding a few stupid tests to configure to HELP somebody damn it! Seeing response like that it just makes you wanna puke and go elsewhere. Last time it was like that: "The code doesn't compile on Windows, it fails even to configure. It was removed. Fix it and we'll accept it." And that's fucking it - now the history is repeating. Some people over here are so dense than even concrete is as soft as clay compared to them. Granted, what I do or did is nothing compared to most of the people around here, but another sure thing is that the community doesn't encourage external contributions. Take the fork() flamewar sometime ago - it was a huge waste of time and it was enough to post an explanation like the one Mirar posted to pike@roxen.com (I think) two or three weeks ago - several sentences, concise, precise. Sufficient. And what happens now? Nilsson butts in and plays a teacher preaching what and how should be done, then Agehall joins him to sing unisono - too bad their posts contribute nothing substantial but whitenoise and fuel for flames. Maybe it's time to create a mailing list for them? pike-bitchers@pike.ida.liu.se would sound quite cool - and you can count on me joining it. After all, I love flamewars.
marek
p.s. yes, I'm venting off and call me a flame maker if you want, I couldn't care less.
Possibly the same way as we do (since we don't have access to many of the machines either): Read the error messages. Search the Internet. Read the docs. Check in new code. Iterate.
/ Martin Nilsson (har bott i google)
Previous text:
2003-04-24 00:16: Subject: Re: [caudium-general] ipv6 support
Le mercredi, 23 avr 2003, à 23:40 Europe/Paris, Marcus Agehall (Trådlös) @ Pike (-) developers forum a écrit :
But if you check in code that breaks on a platform that you can't fix it for, make a configuretest to make sure that your code doesn't break the build on that platform!
Humm... Interressing... But how to make a make configuretest on a platform we don't any access (eg WinNT or whatever exotic OS ?).
AFAIK 7.5 branch is -devel version, eg version we can break for a small time when we try to add some new feature in the core... Or I didn't understood what is a devel version for ?
My 0.02...
/Xavier
/ Brevbäraren
I have access to MSDN? How?
/ Mirar
Previous text:
2003-04-23 14:21: Subject: Re: [caudium-general] ipv6 support
For another time. I had no means (and, therefore, no intention) to fix it on Windows NT.
Well, you have access to the same tools as all other Pike developers (CVS, pikefarm and MSDN).
/ Martin Nilsson (har bott i google)
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
Previous text:
2003-04-23 14:42: Subject: Re: [caudium-general] ipv6 support
I have access to MSDN? How?
/ Mirar
pike-devel@lists.lysator.liu.se
-
David Gourdelier -
Johan Sundstr�m (folksk�dare) ＠ Pike (-) developers forum -
Marcus Agehall (Tr�dl�s) ＠ Pike (-) developers forum
-
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum
-
Marek Habersack -
Martin Baehr -
Martin Nilsson (har bott i google) ＠ Pike (-) developers forum
-
Mirar ＠ Pike developers forum
-
Per Hedbor () ＠ Pike (-) developers forum
-
Peter Lundqvist (disjunkt) ＠ Pike (-) developers forum
-
Xavier Beaudouin