are there people working on getting the rhwas 5 rpms builtup for centosplus 5? the gophp5 people are starting to have an impact and drupal users are wanting php 5.2, which i was surprised to see in the rhwas srpms directory. the users might even be willing to be testers since they're on a virtual server.
Joe Pruett wrote:
are there people working on getting the rhwas 5 rpms builtup for centosplus 5? the gophp5 people are starting to have an impact and drupal users are wanting php 5.2, which i was surprised to see in the rhwas srpms directory. the users might even be willing to be testers since they're on a virtual server.
Yes, been working on the rhwas stuff for centos5, and I was going to post an email on this list about that tomorrow.
Some of the components, like php5/mysql/httpd/mod_perl etc build with no issues - however, the Java bits are a bit tricky. And to be honest, I could use some help from people who know a bit more about the situation.
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
- KB
And, btw - gophp5 people are idiots. Feel free to tell them that. If only this much energy was spent on educating the script kiddies on proper coding methodologies and audit process etc, we wont have php as, imho, the largest security liability in the Linux server world today.
Karanbir Singh wrote:
Joe Pruett wrote:
are there people working on getting the rhwas 5 rpms builtup for centosplus 5? the gophp5 people are starting to have an impact and drupal users are wanting php 5.2, which i was surprised to see in the rhwas srpms directory. the users might even be willing to be testers since they're on a virtual server.
Yes, been working on the rhwas stuff for centos5, and I was going to post an email on this list about that tomorrow.
Some of the components, like php5/mysql/httpd/mod_perl etc build with no issues - however, the Java bits are a bit tricky. And to be honest, I could use some help from people who know a bit more about the situation.
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
I'm sure you can ship jre, always have "with your application." Wrt jdk, I'd check the licence with the latest, and if in doubt ask Sun or ship something to download it and install it from Sun's site.
John Summerfield wrote:
We cant build and ship out the java stuff without using some jre / jvm
- IcedTea is not quite there yet. And I dont think we can actually
ship any jdk due to licensing issues. Could we ?
I'm sure you can ship jre, always have "with your application." Wrt jdk, I'd check the licence with the latest, and if in doubt ask Sun or ship something to download it and install it from Sun's site.
I am relatively sure that putting jre in a repo didnt meet the criteria of 'with the app'. Sun had something called the DDL for distro's that wanted to ship their Java stack, however their terms didnt go down well with the CentOS developers and also they wanted various bits of paper signed etc. So, imho, Sun's stuff is quite far off the radar.
RedHat seem to rely on Bea stack, does anyone have a decent idea on what their terms might be ? Would someone like to actually knock on their door and ask ?
Karanbir Singh wrote:
John Summerfield wrote:
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
I'm sure you can ship jre, always have "with your application." Wrt jdk, I'd check the licence with the latest, and if in doubt ask Sun or ship something to download it and install it from Sun's site.
I am relatively sure that putting jre in a repo didnt meet the criteria of 'with the app'. Sun had something called the DDL for distro's that wanted to ship their Java stack, however their terms didnt go down well with the CentOS developers and also they wanted various bits of paper signed etc. So, imho, Sun's stuff is quite far off the radar.
What arrangement does Canonical have for Ubuntu?
RedHat seem to rely on Bea stack, does anyone have a decent idea on what their terms might be ? Would someone like to actually knock on their door and ask ?
John Summerfield wrote:
What arrangement does Canonical have for Ubuntu?
You'll need to go ask Canonical for that wont you ?
Karanbir Singh wrote:
John Summerfield wrote:
What arrangement does Canonical have for Ubuntu?
You'll need to go ask Canonical for that wont you ?
I would have thought you might be interested in a similar arrangement; it doesn't bother me particularly, but I'm trying to help solve _your_ problem. You said, "We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?"
If real Java can be shipped with an acceptable though imperfect licence then I'd think the problem's solved. And as Java licensing is moving recently, I suggest it's worth your time asking.
John Summerfield wrote:
What arrangement does Canonical have for Ubuntu?
You'll need to go ask Canonical for that wont you ?
I would have thought you might be interested in a similar arrangement;
Similar to what ? You havent actually pointed out any arrangement details that might be on offer.
If real Java can be shipped with an acceptable though imperfect licence then I'd think the problem's solved. And as Java licensing is moving recently, I suggest it's worth your time asking.
As I have pointed out a couple of times already on this thread, my aim of asking here was if anyone wanted to pursue this. I am not looking for opinions or ideas to chase up. If you are indeed interested in helping, then help, standing on the side lines and passing commentary is not something I am ( or most people on this list ) interested in.
Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
What arrangement does Canonical have for Ubuntu?
You'll need to go ask Canonical for that wont you ?
I would have thought you might be interested in a similar arrangement;
Similar to what ? You havent actually pointed out any arrangement details that might be on offer.
There is an arrangement in place, but I don't know the details. Probably a simple enquiry would find what's available now. It may well be more relaxed than what was on offer last year. IBM has been pressing Sun to open-source Java for years, I wouldn't expect any problem from IBM.
If real Java can be shipped with an acceptable though imperfect licence then I'd think the problem's solved. And as Java licensing is moving recently, I suggest it's worth your time asking.
As I have pointed out a couple of times already on this thread, my aim of asking here was if anyone wanted to pursue this. I am not looking for opinions or ideas to chase up. If you are indeed interested in helping, then help, standing on the side lines and passing commentary is not something I am ( or most people on this list ) interested in.
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
None of that goes to say I cannot offer useful advice.
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
William L. Maltby wrote:
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
I hate to get involved in this discussion, but it seems to me that the last thing that any of the core developers need at this point is more people coming in and heaping work on their plate. There are barely enough hours in the day for them to accomplish their existing tasks.
If you have something new you'd like to see done, that's great. It's always nice to know what's wanted, and how it might be accomplished... but unless you're ready to take it on yourself, chances are low that someone else on this mailing list is going to pick it up and run with it for you. About the only thing you *can* expect is Karanbir (or someone else) giving you a polite suggestion that you follow it up yourself, and come back when you have some work to show.
-Brandon
On Fri, 2008-03-28 at 16:50 -0700, Brandon Davidson wrote:
William L. Maltby wrote:
<snip>
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
I hate to get involved in this discussion, but it seems to me that the last thing that any of the core developers need at this point is more people coming in and heaping work on their plate. There are barely enough hours in the day for them to accomplish their existing tasks.
Well, not to disagree too strongly, but ...
It is from the success and vitality of the project that encourages folks to have a positive feeling that requests for additional features may be well received.
I feel that most folks know the project crew is fully loaded and most would be/are willing to help as they can. If they can't or, for whatever personal reason, don't choose to help do the work, does that mean that they should not request, suggest or participate in discussions?
I think not. The project folks know that not everyone will be able/willing to help in any way but they still cheerfully (usually) undertake the effort and do their best.
This is one of the things that make the project successful and will, in the long run, recruit more folks to help.
As to "heaping work on their plate", their is more than one alternative. The obvious one - folks contributing more - and a less obvious one. That is their is so little interest that no one requests anything, no one helps out, the project ends up bing guided only by only its internal view of what should be happening, etc.
No one "heaps work on their plate". *They* heap work on their plate in accordance with their appetite. As we are grateful for their efforts on our behalf, I believe that they are grateful for the satisfaction they get from having a large "smorgasbord" from which to fill their plates.
I'm not implying that a "pure user's" contribution is of the same scale as theirs. But it is a symbiotic relationship, if you will.
As a former long-time designer, developer, implementer, ... I believe that the team is basically motivated by things similar to what I see/saw in myself. I could be wrong. As my personal situation changes to permit it, I plan to try and contribute somewhere down the road. It will likely be quite minor, as to tangible items, but I will get some satisfaction from it and I will hope the community receives benefit from it.
If you have something new you'd like to see done, that's great. It's always nice to know what's wanted, and how it might be accomplished... but unless you're ready to take it on yourself, chances are low that someone else on this mailing list is going to pick it up and run with it for you.
I'm pretty sure that the long-time subscribers are aware of that. OTOH, with this project, because of the attitudes and efforts of the crew, ISTM that likelihood of a request being acted upon is higher than on some others that I used to follow.
So I disagree - I think chances are *relatively* *high*, if it is a request that seems to have overall application and meets other criteria that the project may have.
About the only thing you *can* expect is Karanbir (or someone else) giving you a polite suggestion that you follow it up yourself, and
^^^^^^
I hope no one reads this wrong. "Polite" is relative. There have been times that various crew members fail *my* definition of this and my tolerance is probably wider than most (self assessment, may be totally wrong). When folks are pressed (for time, resources or stressed about something) they may fail to take the time to add words or phrases that "soften" the tenor of their reply. Then it may sound undeservedly harsh, argumentative, aggressive or <pick your own PC descriptor here>.
Adding the inability of electronic communications to easily project your facial expressions, tone of voice, ... only exacerbates the potential for an adverse reaction to occur. Add in the *perception* and personality of the recipients (thin-skinned, thick-skinned, a southern U.S. genteel upbringing, ...) and the potential climbs even more.
As to Karanbir's response to the poster, it could have been a little less terse. I could fully understand the OP's response. For the OP's part, he could have been a little more cognizant of some of the things I mention above.
I've only seen one CentOS member that I believe engages in habitual "rudeness" (a subjective assessment - no discussion needed). I've never felt the need to call him on it, for various reasons related to things I discuss above.
come back when you have some work to show.
See above. "The law of unintended consequences" may apply. If there are no "pure users", what is the value of the effort? From *my* background, the answer is zero.
That does not mean that help should not be requested. Obviously it should.
Per Forrest Gump: "That's all I have to say about that". :-)
-Brandon
<snip sig stuff>
MHO
On Sat, Mar 29, 2008 at 12:18 AM, William L. Maltby CentOS4Bill@triad.rr.com wrote:
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
It is, but CentOS has a relatively small number of active contributors, compared to the work that needs to be done. Some of us have multi-page CentOS-related todo lists already. So, to move forward, people have to step up, and actively persue projects.
-- Daniel
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Daniel de Kok wrote: | On Sat, Mar 29, 2008 at 12:18 AM, William L. Maltby | CentOS4Bill@triad.rr.com wrote: |> On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote: |> > |> > Well, the whole point to CentOS is that anyone and everyone should be |> > able to help along. Dont need a tag / title.. |> |> And a good part of that is contributing ideas, pointers, support to |> others, etc. IMO. Not everyone has time or inclination to contribute in |> ways that the project might like. |> |> And that fits within the spirit of open source, no? | | It is, but CentOS has a relatively small number of active | contributors, compared to the work that needs to be done. Some of us | have multi-page CentOS-related todo lists already. So, to move | forward, people have to step up, and actively persue projects.
And where may we find the to-do list? If anything fits me I may offer to lend a hand. It might sound like a drop. But considere the ocean is just ~ a big bunch of drops.
I hope to take the sparc stuff for a spin this weekend.
Hugo.
- -- hvdkooij@vanderkooij.org http://hugo.vanderkooij.org/ PGP/GPG? Use: http://hugo.vanderkooij.org/0x58F19981.asc
A: Yes. >Q: Are you sure? >>A: Because it reverses the logical flow of conversation. >>>Q: Why is top posting frowned upon?
Bored? Click on http://spamornot.org/ and rate those images.
On Sat, Mar 29, 2008 at 3:53 PM, Hugo van der Kooij hvdkooij@vanderkooij.org wrote:
Daniel de Kok wrote: | On Sat, Mar 29, 2008 at 12:18 AM, William L. Maltby | CentOS4Bill@triad.rr.com wrote: |> On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote: |> > |> > Well, the whole point to CentOS is that anyone and everyone should be |> > able to help along. Dont need a tag / title.. |> |> And a good part of that is contributing ideas, pointers, support to |> others, etc. IMO. Not everyone has time or inclination to contribute in |> ways that the project might like. |> |> And that fits within the spirit of open source, no? | | It is, but CentOS has a relatively small number of active | contributors, compared to the work that needs to be done. Some of us | have multi-page CentOS-related todo lists already. So, to move | forward, people have to step up, and actively persue projects.
And where may we find the to-do list? If anything fits me I may offer to lend a hand. It might sound like a drop. But considere the ocean is just ~ a big bunch of drops.
There is a somewhat incomplete list on the wiki: http://wiki.centos.org/ToDo
You can also participate in a SIG that interests you.
Take care, Daniel
William L. Maltby wrote:
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
Thanks Bill.
In this particular case, Johnny more than anyone else speaks for CentOS, and I probably least of all.
CentOS has a problem with the Sun licence. On checking, it didn't seem so bad. Okay, so it maybe still isn't good enough. Perhaps one of the leaders ought tell Sun; in view of its decision to open-source Java, maybe Sun will move a little more. I don't think IBM will argue against more freedom, it's been taking shots at Sun over the matter for years.
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
If the licence issue is resolved satisfactorily, _then_ there will be some work to do, and maybe someone with needs in that area will step up.
John Summerfield wrote:
William L. Maltby wrote:
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
Thanks Bill.
In this particular case, Johnny more than anyone else speaks for CentOS, and I probably least of all.
CentOS has a problem with the Sun licence. On checking, it didn't seem so bad. Okay, so it maybe still isn't good enough. Perhaps one of the leaders ought tell Sun; in view of its decision to open-source Java, maybe Sun will move a little more. I don't think IBM will argue against more freedom, it's been taking shots at Sun over the matter for years.
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
If the licence issue is resolved satisfactorily, _then_ there will be some work to do, and maybe someone with needs in that area will step up.
John,
Actually ... Sun DID do something about the license, they changed it completely for new versions of Java to GPL :D
The issue is, that the versions of Java that is GPL is 1.6.0 and newer ... and that CentOS-5 uses 1.4.2 level things.
The problem we have is with distributing the 1.4.2 version ... which Sun really can't change more than they have (or at least they seem unwilling to do so).
The GPL'ed version, 1.6.0, I am trying to get working on CentOS-5 in a sane way right now.
If we can make that work ... AND IF it will build the required java bits ... AND AGAIN IF we can make that work with the other gcj java bits already throughout CentOS, then we will be in business.
That is really the issue.
Thanks, Johnny Hughes
Johnny Hughes wrote:
John Summerfield wrote:
William L. Maltby wrote:
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
Thanks Bill.
In this particular case, Johnny more than anyone else speaks for CentOS, and I probably least of all.
CentOS has a problem with the Sun licence. On checking, it didn't seem so bad. Okay, so it maybe still isn't good enough. Perhaps one of the leaders ought tell Sun; in view of its decision to open-source Java, maybe Sun will move a little more. I don't think IBM will argue against more freedom, it's been taking shots at Sun over the matter for years.
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
If the licence issue is resolved satisfactorily, _then_ there will be some work to do, and maybe someone with needs in that area will step up.
John,
Actually ... Sun DID do something about the license, they changed it completely for new versions of Java to GPL :D
And that's better than I expected. These folk can probably help with legal advice: http://www.softwarefreedom.org/
I've just had a quick look at the licence for jdk 1.5.0 which I have on Scientific Linux 5, and two points look to me problematic 1. Indemnity. SFLC might have an opinion about that 2. ... software intended to replace any component(s) .... Sounds like a debating point to me, but IANAL. SFLC has some though.
The issue is, that the versions of Java that is GPL is 1.6.0 and newer ... and that CentOS-5 uses 1.4.2 level things.
The problem we have is with distributing the 1.4.2 version ... which Sun really can't change more than they have (or at least they seem unwilling to do so).
The GPL'ed version, 1.6.0, I am trying to get working on CentOS-5 in a sane way right now.
If we can make that work ... AND IF it will build the required java bits ... AND AGAIN IF we can make that work with the other gcj java bits already throughout CentOS, then we will be in business.
That is really the issue.
Does Jikes compile it satisfactorily?
Johnny Hughes wrote:
John Summerfield wrote:
William L. Maltby wrote:
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
Thanks Bill.
In this particular case, Johnny more than anyone else speaks for CentOS, and I probably least of all.
CentOS has a problem with the Sun licence. On checking, it didn't seem so bad. Okay, so it maybe still isn't good enough. Perhaps one of the leaders ought tell Sun; in view of its decision to open-source Java, maybe Sun will move a little more. I don't think IBM will argue against more freedom, it's been taking shots at Sun over the matter for years.
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
If the licence issue is resolved satisfactorily, _then_ there will be some work to do, and maybe someone with needs in that area will step up.
John,
Actually ... Sun DID do something about the license, they changed it completely for new versions of Java to GPL :D
The issue is, that the versions of Java that is GPL is 1.6.0 and newer ... and that CentOS-5 uses 1.4.2 level things.
The problem we have is with distributing the 1.4.2 version ... which Sun really can't change more than they have (or at least they seem unwilling to do so).
The GPL'ed version, 1.6.0, I am trying to get working on CentOS-5 in a sane way right now.
If we can make that work ... AND IF it will build the required java bits ... AND AGAIN IF we can make that work with the other gcj java bits already throughout CentOS, then we will be in business.
That is really the issue.
OK ... some progress on this
There are now java-1.6.0-openjdk packages for i386 in the testing repo:
http://dev.centos.org/centos/5/testing/i386/RPMS/
I am working on the x86_64 packages now.
The next step will be to see if we can use these to make the JAVA bits in RHWAS.
Thanks. Johnny Hughes
Johnny Hughes wrote:
Johnny Hughes wrote:
John Summerfield wrote:
William L. Maltby wrote:
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
Thanks Bill.
In this particular case, Johnny more than anyone else speaks for CentOS, and I probably least of all.
CentOS has a problem with the Sun licence. On checking, it didn't seem so bad. Okay, so it maybe still isn't good enough. Perhaps one of the leaders ought tell Sun; in view of its decision to open-source Java, maybe Sun will move a little more. I don't think IBM will argue against more freedom, it's been taking shots at Sun over the matter for years.
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
If the licence issue is resolved satisfactorily, _then_ there will be some work to do, and maybe someone with needs in that area will step up.
John,
Actually ... Sun DID do something about the license, they changed it completely for new versions of Java to GPL :D
The issue is, that the versions of Java that is GPL is 1.6.0 and newer ... and that CentOS-5 uses 1.4.2 level things.
The problem we have is with distributing the 1.4.2 version ... which Sun really can't change more than they have (or at least they seem unwilling to do so).
The GPL'ed version, 1.6.0, I am trying to get working on CentOS-5 in a sane way right now.
If we can make that work ... AND IF it will build the required java bits ... AND AGAIN IF we can make that work with the other gcj java bits already throughout CentOS, then we will be in business.
That is really the issue.
OK ... some progress on this
There are now java-1.6.0-openjdk packages for i386 in the testing repo:
http://dev.centos.org/centos/5/testing/i386/RPMS/
I am working on the x86_64 packages now.
The next step will be to see if we can use these to make the JAVA bits in RHWAS.
For anyone interested in trying the openjdk packages, the i386 and x86_64 java-1.6.0-openjdk packages are now posted in the testing repo.
Thanks, Johnny Hughes
Johnny Hughes a écrit :
Johnny Hughes wrote:
John Summerfield wrote:
William L. Maltby wrote:
On Fri, 2008-03-28 at 20:29 +0000, Karanbir Singh wrote:
John Summerfield wrote:
Karanbir I find you very difficult to deal with, and I don't want that. Can we try to keep things calm?
I am always calm :D
I have no standing with the CentOS project, and I don't wish to become a developer. I know myself well enough to understand that it's not something I do well, even though a resume of my experience might suggest otherwise.
Well, the whole point to CentOS is that anyone and everyone should be able to help along. Dont need a tag / title..
And a good part of that is contributing ideas, pointers, support to others, etc. IMO. Not everyone has time or inclination to contribute in ways that the project might like.
And that fits within the spirit of open source, no?
Thanks Bill.
In this particular case, Johnny more than anyone else speaks for CentOS, and I probably least of all.
CentOS has a problem with the Sun licence. On checking, it didn't seem so bad. Okay, so it maybe still isn't good enough. Perhaps one of the leaders ought tell Sun; in view of its decision to open-source Java, maybe Sun will move a little more. I don't think IBM will argue against more freedom, it's been taking shots at Sun over the matter for years.
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
If the licence issue is resolved satisfactorily, _then_ there will be some work to do, and maybe someone with needs in that area will step up.
John,
Actually ... Sun DID do something about the license, they changed it completely for new versions of Java to GPL :D
The issue is, that the versions of Java that is GPL is 1.6.0 and newer ... and that CentOS-5 uses 1.4.2 level things.
The problem we have is with distributing the 1.4.2 version ... which Sun really can't change more than they have (or at least they seem unwilling to do so).
The GPL'ed version, 1.6.0, I am trying to get working on CentOS-5 in a sane way right now.
If we can make that work ... AND IF it will build the required java bits ... AND AGAIN IF we can make that work with the other gcj java bits already throughout CentOS, then we will be in business.
That is really the issue.
OK ... some progress on this
There are now java-1.6.0-openjdk packages for i386 in the testing repo:
http://dev.centos.org/centos/5/testing/i386/RPMS/
I am working on the x86_64 packages now.
The next step will be to see if we can use these to make the JAVA bits in RHWAS.
Thanks. Johnny Hughes
There is something that I don't have clearly understood, why couldn't you use icedtea java instead ?
Ralph Angenendt wrote:
Jean-Marc LIGER wrote:
There is something that I don't have clearly understood, why couldn't you use icedtea java instead ?
Because icedtea is Java 1.7, which isn't really there yet. This one is Java 1.6.
And even Java 1.6 is a strech, as the webstack is currently built using java-1.4.2 ... which we can not use to build since we can not distribute it.
Java 1.6 is the one we will need to use to get as close as possible to the correct version AND be able to release the RPMS themselves and the JRE required to run them.
Johnny Hughes wrote:
Ralph Angenendt wrote:
Jean-Marc LIGER wrote:
There is something that I don't have clearly understood, why couldn't you use icedtea java instead ?
Because icedtea is Java 1.7, which isn't really there yet. This one is Java 1.6.
And even Java 1.6 is a strech, as the webstack is currently built using java-1.4.2 ... which we can not use to build since we can not distribute it.
What's the problem with a jpackage nosrc rpm and instructions to download your own copy of the java binary from Sun and execute this command?
Les Mikesell wrote:
Johnny Hughes wrote:
Ralph Angenendt wrote:
Jean-Marc LIGER wrote:
There is something that I don't have clearly understood, why couldn't you use icedtea java instead ?
Because icedtea is Java 1.7, which isn't really there yet. This one is Java 1.6.
And even Java 1.6 is a strech, as the webstack is currently built using java-1.4.2 ... which we can not use to build since we can not distribute it.
What's the problem with a jpackage nosrc rpm and instructions to download your own copy of the java binary from Sun and execute this command?
Because CentOS does not distribute RPMS that we can't get repo closure on.
People expect that if they say yum install jbossas it will work and NOT require them to build stuff.
Also ... I can't build GPL stuff against NON-FREE stuff and distribute it as GPL :D
(Well or at least it is a questionable proactice)
AND since there is a FREE / GPL version available that is what we will use or we won't distribute it.
Johnny Hughes wrote:
There is something that I don't have clearly understood, why couldn't you use icedtea java instead ?
Because icedtea is Java 1.7, which isn't really there yet. This one is Java 1.6.
And even Java 1.6 is a strech, as the webstack is currently built using java-1.4.2 ... which we can not use to build since we can not distribute it.
What's the problem with a jpackage nosrc rpm and instructions to download your own copy of the java binary from Sun and execute this command?
Because CentOS does not distribute RPMS that we can't get repo closure on.
By popular demand? Dictated by some arbitrary policy?
People expect that if they say yum install jbossas it will work and NOT require them to build stuff.
I don't know anyone who currently expects that. Those expectations were dashed long ago in the fedora/centos world along with any expectations of java being usable at all. They used to expect to go spend a day reading the jpackage docs to find the right pieces, then grab the Sun package, rebuild the nosrc rpm, then use yum to install the jpackage apps. But current fedora and Centos 5.x include broken versions of what jpackage used to provide and jpackage doesn't even have documentation on how to use their packages with them even thought the nosrc rpm for sun would probably still work.
Also ... I can't build GPL stuff against NON-FREE stuff and distribute it as GPL :D
Aggregation isn't a problem.
(Well or at least it is a questionable proactice)
Nobody questioned it when Sun was the only possible JVM.
AND since there is a FREE / GPL version available that is what we will use or we won't distribute it.
But what about that expectation of working...
Les Mikesell wrote:
Johnny Hughes wrote:
There is something that I don't have clearly understood, why couldn't you use icedtea java instead ?
Because icedtea is Java 1.7, which isn't really there yet. This one is Java 1.6.
And even Java 1.6 is a strech, as the webstack is currently built using java-1.4.2 ... which we can not use to build since we can not distribute it.
What's the problem with a jpackage nosrc rpm and instructions to download your own copy of the java binary from Sun and execute this command?
Because CentOS does not distribute RPMS that we can't get repo closure on.
By popular demand? Dictated by some arbitrary policy?
Because I said so :D
People expect that if they say yum install jbossas it will work and NOT require them to build stuff.
I don't know anyone who currently expects that. Those expectations were dashed long ago in the fedora/centos world along with any expectations of java being usable at all. They used to expect to go spend a day reading the jpackage docs to find the right pieces, then grab the Sun package, rebuild the nosrc rpm, then use yum to install the jpackage apps. But current fedora and Centos 5.x include broken versions of what jpackage used to provide and jpackage doesn't even have documentation on how to use their packages with them even thought the nosrc rpm for sun would probably still work.
Also ... I can't build GPL stuff against NON-FREE stuff and distribute it as GPL :D
Aggregation isn't a problem.
It isn't aggregation IF it is a derivative work.
(Well or at least it is a questionable proactice)
Nobody questioned it when Sun was the only possible JVM.
Sure it was ... you don't see any of that stuff in Fedora do you? Or Gentoo or Debian. If it was such a good idea, they all would have had it in there several years ago.
AND since there is a FREE / GPL version available that is what we will use or we won't distribute it.
But what about that expectation of working...
If it doesn't work it doesn't ship ... and we are right back where we started.
The bottom line is if you require that, buy a RHWAS and jboss subscription ... at least if we can't make a redistributable one work.
Johnny Hughes wrote:
Aggregation isn't a problem.
It isn't aggregation IF it is a derivative work.
It can't be a derivative if an alternative could possibly work (regardless of which you ship).
(Well or at least it is a questionable proactice)
Nobody questioned it when Sun was the only possible JVM.
Sure it was ... you don't see any of that stuff in Fedora do you?
I see it in RHEL. So shipping some GPL'd stuff too can't be a problem.
Or Gentoo or Debian. If it was such a good idea, they all would have had it in there several years ago.
If those guys had all the good ideas there would be no need for ubuntu.
If it doesn't work it doesn't ship ... and we are right back where we started.
The bottom line is if you require that, buy a RHWAS and jboss subscription ... at least if we can't make a redistributable one work.
OK, I guess it's the thought that counts.
On Sat, 29 Mar 2008, John Summerfield wrote:
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
... If it should be done ** at all ** within CentOS. 'has to be done' inplies obligation to do it; there is no such obligation absent sufficient support. How shall those who do it insure against the liability that Sun document purports to impose? How shall that insurance be paid for? If a given person wants to take that risk, uninsured, that is their perogative.
Shall the individuals who make up the CentOS project core put their personal assets at risk, for free and without compensation, to meet someone's 'expectation' for which they have not paid? I think not.
And it is just not the case that the CentOS 'has to', nor indeed _can_ be all things to all people. The mice can vote to 'bell the cat' all they wish, but that does not do it until some mouse does it.
This week, in another part of FOSS, Fedora finds itself in a trap, fueled in part by 'what if' extreme remarks by some who post here, to be 'more friendly' to users who will not learn the Unix ways of PATH. If Fedora proceeds that way, it is at risk of being not Unix0like any more. It would lose parts of the aspects which make Fedora a proving ground for Enterprise distributions. Gentoo, and Ubuntu can do that just fine already. If it goes that way, Fedora commits suicide.
Part of an Enterprise approach is growing up, putting aside childish ways, and being 'businesslike'. I like CentOS as a boring and reliable environment in which to live professionally. I will continue to argue strongly against risking its continued existence.
-- Russ herrold
On Sat, 29 Mar 2008, R P Herrold wrote:
And it is just not the case that the CentOS 'has to', nor indeed _can_ be all things to all people. The mice can vote to 'bell the cat' all they wish, but that does not do it until some mouse does it.
I see Johnny's post intervening. There was this movie (and it appears a book) when I was a child: The Mouse That Roared. http://en.wikipedia.org/wiki/The_Mouse_That_Roared
I nominate hughes as Emperor of the Duchy of Grand Fenwick ;) Rent the movie for a fun evening, if you've not seen it.
- Russ herrold
R P Herrold wrote:
This week, in another part of FOSS, Fedora finds itself in a trap, fueled in part by 'what if' extreme remarks by some who post here, to be 'more friendly' to users who will not learn the Unix ways of PATH. If Fedora proceeds that way, it is at risk of being not Unix0like any more. It would lose parts of the aspects which make Fedora a proving ground for Enterprise distributions. Gentoo, and Ubuntu can do that just fine already. If it goes that way, Fedora commits suicide.
Do you think a unix-like OS distribution is only viable if it propagates the flavor wars with incompatible differences with other distributions? I guess that does follow a long tradition...
On Sat, 29 Mar 2008, Les Mikesell wrote:
R P Herrold wrote:
This week, in another part of FOSS, Fedora finds itself in a trap, fueled in part by 'what if' extreme remarks by some who post here, to be 'more friendly' to users who will not learn the Unix ways of PATH.
Do you think a unix-like OS distribution is only viable if it propagates the flavor wars with incompatible differences with other distributions?
Yes under the principle of not breaking reasonable expectation (see: the first Gancarz essay 'The UNIX Philosophy' for a good reprise of it).
Not that I concur in the combustible qualification later in that sentence.
We are in the present place we are and the fragmentation wars are long over; the ATT which attacked Berkeley is also long gone and unmourned. No tears can wash out the ink the moving finger has left. The landscape shaped by how we got here are in no way 'incompatible' to a user in the tool builders culture of 'nix; BSD and SysV live in the 'ps' and 'find' I rely upon every day, and each brings a strength to those who live in the shell.
Two lines of code permit a person ignorant of PATH changes to reach their ill-advised desire; altering the operating system, so that the minimal tools needed for recovery from the single '/' partition would be a massive and pervasive changed, and a serious loss. It is a fool's errand to do so; it has proved a debating society's pigsty to wallow in.
Discarding culture, ignoring history, and faddishly taking away strengths to 'gain share' is the way of vendors and those with an agenda to grind; and not the way of those who live in the Unix culture who need to work in a long lived stable environment.
- Russ herrold
R P Herrold wrote:
Two lines of code permit a person ignorant of PATH changes to reach their ill-advised desire;
But giving root a different and confusingly different environment was a late and unique branch in history. And unnecessary.
altering the operating system, so that the minimal tools needed for recovery from the single '/' partition would be a massive and pervasive changed, and a serious loss. It is a fool's errand to do so; it has proved a debating society's pigsty to wallow in.
When small disk drives cost $10,000 and most machines could only be booted from the vendor-suppied device, there was a reason to care if you could boot from a tiny partition. That reason is long gone but...
That's not the change being discussed. It is more about combining /sbin with /bin and /usr/sbin with /usr/bin - or simply giving everyone the same PATH. There is no sensible reason that a user should wonder why he can't run ifconfig to get his IP address, or that root shouldn't be able to find fdisk if he used 'su' instead of 'su -' to get there.
Discarding culture, ignoring history, and faddishly taking away strengths to 'gain share' is the way of vendors and those with an agenda to grind; and not the way of those who live in the Unix culture who need to work in a long lived stable environment.
You should look at the whole history before saying that. You'll find that adding the /sbin and /usr/sbin directories was the faddish move in Solaris, probably because they didn't trust their dynamic-linked programs and put static-linked utilities there to help recover from possible problems. I'm not sure who had the bright idea of supplying different environments to root compared to other users, but I'd bet it didn't come from the original and elegantly simple unix versions, and it doesn't play well with the current best practice advice to only switch to root when needed. Aside from the PATH nonsense, consider what happens when someone is accustomed to the aliases only in root's environment accidentally does 'su' instead of 'su -', and wonders why rm didn't ask if he really wanted to do that.
If you want people to learn to use unix usefully, you have to expose its simplicity consistently, not disguise it differently under different circumstances.
On Sat, Mar 29, 2008 at 1:35 PM, Les Mikesell lesmikesell@gmail.com wrote:
R P Herrold wrote:
If you want people to learn to use unix usefully, you have to expose its simplicity consistently, not disguise it differently under different circumstances.
CentOS-devel mailing list CentOS-devel@centos.org
Errr... I'm afraid we are deviating from the original topic and this is no longer a CentOS-specific issue. Unlike the IRC, we do not have "CentOS-social" mailing list. Therefore, I would like to invite everyone interested in discussing the current subject to the CentOS forum at:
http://www.centos.org/modules/newbb/viewforum.php?forum=14
Akemi / toracat
Les Mikesell wrote:
R P Herrold wrote:
Two lines of code permit a person ignorant of PATH changes to reach their ill-advised desire;
But giving root a different and confusingly different environment was a late and unique branch in history. And unnecessary.
altering the operating system, so that the minimal tools needed for recovery from the single '/' partition would be a massive and pervasive changed, and a serious loss. It is a fool's errand to do so; it has proved a debating society's pigsty to wallow in.
When small disk drives cost $10,000 and most machines could only be booted from the vendor-suppied device, there was a reason to care if you could boot from a tiny partition. That reason is long gone but...
That's not the change being discussed. It is more about combining /sbin with /bin and /usr/sbin with /usr/bin - or simply giving everyone the same PATH. There is no sensible reason that a user should wonder why he can't run ifconfig to get his IP address, or that root shouldn't be able to find fdisk if he used 'su' instead of 'su -' to get there.
Discarding culture, ignoring history, and faddishly taking away strengths to 'gain share' is the way of vendors and those with an agenda to grind; and not the way of those who live in the Unix culture who need to work in a long lived stable environment.
You should look at the whole history before saying that. You'll find that adding the /sbin and /usr/sbin directories was the faddish move in Solaris, probably because they didn't trust their dynamic-linked programs and put static-linked utilities there to help recover from possible problems. I'm not sure who had the bright idea of supplying different environments to root compared to other users, but I'd bet it didn't come from the original and elegantly simple unix versions, and it doesn't play well with the current best practice advice to only switch to root when needed. Aside from the PATH nonsense, consider what happens when someone is accustomed to the aliases only in root's environment accidentally does 'su' instead of 'su -', and wonders why rm didn't ask if he really wanted to do that.
If you want people to learn to use unix usefully, you have to expose its simplicity consistently, not disguise it differently under different circumstances.
What the hell does this have to do with anything remotely relating to CentOS.
There is an FHS, almost all UNIX variants use it ... there is a reason for root to have a different path than normal users.
This is the document: http://www.pathname.com/fhs/pub/fhs-2.3.html
If you want the standards changed, there are working groups to get it changed ... this is NOT one of those working groups.
Please take traffic concerning how screwed up the FHS is and how it is an old standard for an old time to the appropriate place.
If the FHS changes and removes /sbin and/or if the recommended paths change, then will change UNIX wide. I can guarantee that if that happens, CentOS will also change. I can also guarantee that as long as things stay the way they are, it is silly to discuss these changes on this thread.
Thanks, Johnny Hughes
R P Herrold wrote:
On Sat, 29 Mar 2008, John Summerfield wrote:
I'm the last person who should be negotiating on CentOS's behalf, _this_ has to be done by an acknowledged leader, whether it's Johnny, Karanbir or someone else. Someone with a sound knowledge of the issues, someone Sun will see as someone representing the project.
... If it should be done ** at all ** within CentOS. 'has to be done' inplies obligation to do it; there is no such obligation absent sufficient support. How shall those who do it insure against the liability that Sun document purports to impose? How shall that insurance be paid for? If a given person wants to take that risk, uninsured, that is their perogative.
Shall the individuals who make up the CentOS project core put their personal assets at risk, for free and without compensation, to meet someone's 'expectation' for which they have not paid? I think not.
And it is just not the case that the CentOS 'has to', nor indeed _can_ be all things to all people. The mice can vote to 'bell the cat' all they wish, but that does not do it until some mouse does it.
This week, in another part of FOSS, Fedora finds itself in a trap, fueled in part by 'what if' extreme remarks by some who post here, to be 'more friendly' to users who will not learn the Unix ways of PATH. If Fedora proceeds that way, it is at risk of being not Unix0like any more. It would lose parts of the aspects which make Fedora a proving ground for Enterprise distributions. Gentoo, and Ubuntu can do that just fine already. If it goes that way, Fedora commits suicide.
Part of an Enterprise approach is growing up, putting aside childish ways, and being 'businesslike'. I like CentOS as a boring and reliable environment in which to live professionally. I will continue to argue strongly against risking its continued existence.
I don't see that any of that has anything to do with the text you quoted.
Karanbir Singh wrote:
John Summerfield wrote:
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
I'm sure you can ship jre, always have "with your application." Wrt jdk, I'd check the licence with the latest, and if in doubt ask Sun or ship something to download it and install it from Sun's site.
I am relatively sure that putting jre in a repo didnt meet the criteria of 'with the app'. Sun had something called the DDL for distro's that wanted to ship their Java stack, however their terms didnt go down well with the CentOS developers and also they wanted various bits of paper signed etc. So, imho, Sun's stuff is quite far off the radar.
RedHat seem to rely on Bea stack, does anyone have a decent idea on what their terms might be ? Would someone like to actually knock on their door and ask ?
Or there's the OpenNMS guys that are probably using the 'distributing with application' approach. But, if you set up the yum repository here http://www.opennms.org/index.php/Installation:Yum you can 'yum install opennms' which is a really nice network management/monitoring tool. Or you can just 'yum install jdk'and get a working Sun Java 1.5.
There's also the jpackage nosrc approach if you want to stay on the safe side. You pick up your own non-rpm copy of the jdk from sun and it builds a packaged version with all the 'provides' and symlinks to keep rpm and the alternatives system happy. I've never understood why that hasn't been in the stock repo from the start, especially for 5.x where there isn't a jpackage repo that works.
On Thu, Mar 27, 2008 at 2:02 AM, Karanbir Singh mail-lists@karan.org wrote:
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
I know that Dag used to ship jre with the mozilla plugin. And that would be a huge improvement, given that IcedTea-plguin simply doesn't work for 3/4 of the things I want java in my browser for.
For all I know, things could be much different for jdk - but something like "install the jdk rpm that you get from Sun" and then assume it exists in /usr/java/latest is IMNSHO better than nothing.
regards, Bent
Bent Terp wrote:
For all I know, things could be much different for jdk - but something like "install the jdk rpm that you get from Sun" and then assume it exists in /usr/java/latest is IMNSHO better than nothing.
ok, so thats one way of doing this - if I used $x's java stack, would every user also need $x's stack to run that ? If so, I'd guess the BEA stack is what we'd need to end up using
Karanbir Singh wrote:
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
openjdk6 is available from Fedora 9 - but as their wiki page says:
"Sun has licensed the OpenJDK trademark for use in Fedora."
http://fedoraproject.org/wiki/Docs/Beats/Java
So I'm not sure if that one can be used without problems. And it is not "Java compatible" yet (probably meaning that it hasn't been certified by Sun).
Cheers,
Ralph
Ralph Angenendt wrote:
So I'm not sure if that one can be used without problems. And it is not "Java compatible" yet (probably meaning that it hasn't been certified by Sun).
Right, and I dont quite think its there yet
On Thu, 2008-03-27 at 11:21 +0000, Karanbir Singh wrote:
Ralph Angenendt wrote:
So I'm not sure if that one can be used without problems. And it is not "Java compatible" yet (probably meaning that it hasn't been certified by Sun).
Right, and I dont quite think its there yet
IcedTea certainly seems closer to production than OpenJDK, which is ironic considering that IcedTea is the *next* version.
On Wed, Mar 26, 2008 at 7:02 PM, Karanbir Singh mail-lists@karan.org wrote:
Joe Pruett wrote:
are there people working on getting the rhwas 5 rpms builtup for centosplus 5? the gophp5 people are starting to have an impact and drupal users are wanting php 5.2, which i was surprised to see in the rhwas srpms directory. the users might even be willing to be testers since they're on a virtual server.
Yes, been working on the rhwas stuff for centos5, and I was going to post an email on this list about that tomorrow.
Some of the components, like php5/mysql/httpd/mod_perl etc build with no issues
- however, the Java bits are a bit tricky. And to be honest, I could use some
help from people who know a bit more about the situation.
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
Java items require a contract and permission to ship from each of the vendors (IBM, BEA, Sun). Part of this comes from the overlying Sun contract that IBM/BEA have with Sun to use the Java trademark, and some come from usual business practices. With the application itself licensing is also involved, but that seems to be limited in some ways. Blech... every time I have had to deal with this in the past... the lawyers have to work it out first.
Stephen John Smoogen wrote:
Java items require a contract and permission to ship from each of the vendors (IBM, BEA, Sun). Part of this comes from the overlying Sun contract that IBM/BEA have with Sun to use the Java trademark, and some come from usual business practices. With the application itself licensing is also involved, but that seems to be limited in some ways. Blech... every time I have had to deal with this in the past... the lawyers have to work it out first.
Considering we dont have any lawyers, what would the next step be ?
On Thu, Mar 27, 2008 at 8:17 AM, Karanbir Singh mail-lists@karan.org wrote:
Stephen John Smoogen wrote:
Java items require a contract and permission to ship from each of the vendors (IBM, BEA, Sun). Part of this comes from the overlying Sun contract that IBM/BEA have with Sun to use the Java trademark, and some come from usual business practices. With the application itself licensing is also involved, but that seems to be limited in some ways. Blech... every time I have had to deal with this in the past... the lawyers have to work it out first.
Considering we dont have any lawyers, what would the next step be ?
It would be probably contacting Sun, IBM, BEA to find out what is required for us to ship their RPM with our stuff. They would probably send a long legalese document that some 'official' body has to sign so that they can revoke our shipping their packages if we violate something like their patents, trademark, time of day, etc. Then we get a lawyer to evaluate that document (or just wish and pray and sign it), and decide if whatever CentOS organization can sign away its second kin and third cousins once removed to ship the packages.
Which all probably relates to whether there is an organization that Sun, IBM, BEA, etc consider that they can make a business contract with.
Anyway thats my guess but I am running on little sleep.
Karanbir Singh wrote:
Stephen John Smoogen wrote:
Java items require a contract and permission to ship from each of the vendors (IBM, BEA, Sun). Part of this comes from the overlying Sun contract that IBM/BEA have with Sun to use the Java trademark, and some come from usual business practices. With the application itself licensing is also involved, but that seems to be limited in some ways. Blech... every time I have had to deal with this in the past... the lawyers have to work it out first.
Considering we dont have any lawyers, what would the next step be ?
I'm sending a bcc to Monica Pawlin at Sun. She has some involvement with Java, I'm hoping she can refer this to someone at Sun who matters.
We have corresponded wrt Java and Blackboys (they're trees) before, and she told be about the Sun/Ubuntu announcement.
Monica, I think you don't have a working email address for me, you certainly can't reach this one. summer js.id.au will do.
Joe Pruett wrote:
are there people working on getting the rhwas 5 rpms builtup for centosplus 5? the gophp5 people are starting to have an impact and drupal users are wanting php 5.2, which i was surprised to see in the rhwas srpms directory. the users might even be willing to be testers since they're on a virtual server.
since i started this and was really looking for php support, would it be possible to release all the other packages from rhwas that did compile easily? i was able to build just the php 5.2 rpm myself and it seems to install fine, so it seems like the rest of the non-java stuff should be fine as well. holding it all up for the java licensing issue seems like it could be a long wait...
Joe Pruett wrote:
since i started this and was really looking for php support, would it be possible to release all the other packages from rhwas that did compile easily? i was able to build just the php 5.2 rpm myself and it seems to install fine, so it seems like the rest of the non-java stuff should be fine as well. holding it all up for the java licensing issue seems like it could be a long wait...
Well, there is nothing stopping just a few rpms from making the repos - in CentOS 4, Johnny and I decided that we would actually NOT build the java stuff at all, the legal situation and the fog around the whole java issue was way too dense at the time. However, I am under the (wrong?) impression that some of those issues may have cleared by now. And would like to get the entire Jboss and java stuff in rhwas done for c5plus.
What I am sort of hoping for is that someone who has $clue > 0 w.r.t Java and associated issues, steps forward to try and help with this issue.
Karanbir Singh wrote:
Joe Pruett wrote:
are there people working on getting the rhwas 5 rpms builtup for centosplus 5? the gophp5 people are starting to have an impact and drupal users are wanting php 5.2, which i was surprised to see in the rhwas srpms directory. the users might even be willing to be testers since they're on a virtual server.
Yes, been working on the rhwas stuff for centos5, and I was going to post an email on this list about that tomorrow.
Some of the components, like php5/mysql/httpd/mod_perl etc build with no issues - however, the Java bits are a bit tricky. And to be honest, I could use some help from people who know a bit more about the situation.
We cant build and ship out the java stuff without using some jre / jvm - IcedTea is not quite there yet. And I dont think we can actually ship any jdk due to licensing issues. Could we ?
- KB
And, btw - gophp5 people are idiots. Feel free to tell them that. If only this much energy was spent on educating the script kiddies on proper coding methodologies and audit process etc, we wont have php as, imho, the largest security liability in the Linux server world today.
My take on this is that we should build and ship the non java bits of this now ... and wait for openjdk to be able to produce the java bits.
The whole issue is that to ship Sun Java, we have to indemnify Sun and assume responsibility for things that I don't have the money for. Mark Shuttleworth has a bit more money than me, so we will need to wait for OpenJDK to really be Java to distribute those things I think. All the other Javas are really copies of Sun Java too.
Thanks, Johnny Hughes
Johnny Hughes wrote: <snip>
http://download.java.net/dlj/jdk6/README.html
Doing more research, the license for distributing java FOR THE LATEST SDK seems usable by CentOS.
The new 1.1 version of the JDL is much better than the older license:
http://download.java.net/dlj/jdk6/LICENSE
Whether or not we can use this version of Java to work with the rhwas remains to be seen, but if it will work, we may be in business :D
Thanks, Johnny Hughes
Johnny Hughes wrote:
Johnny Hughes wrote:
<snip>
http://download.java.net/dlj/jdk6/README.html
Doing more research, the license for distributing java FOR THE LATEST SDK seems usable by CentOS.
The new 1.1 version of the JDL is much better than the older license:
http://download.java.net/dlj/jdk6/LICENSE
Whether or not we can use this version of Java to work with the rhwas remains to be seen, but if it will work, we may be in business :D
Hmmm .. the indemnity clause is still there, though a bit lighter that the one I had major issues with.
And openjdk is close ... I think we should wait since we will be supporting it for 5 more years, whatever we do. Open GPL'ed software is a much better approach I think.
Sorry for getting everyone's hopes up for 30 minutes
on 3-28-2008 5:36 AM Johnny Hughes spake the following:
Johnny Hughes wrote:
Johnny Hughes wrote:
<snip>
http://download.java.net/dlj/jdk6/README.html
Doing more research, the license for distributing java FOR THE LATEST SDK seems usable by CentOS.
The new 1.1 version of the JDL is much better than the older license:
http://download.java.net/dlj/jdk6/LICENSE
Whether or not we can use this version of Java to work with the rhwas remains to be seen, but if it will work, we may be in business :D
Hmmm .. the indemnity clause is still there, though a bit lighter that the one I had major issues with.
And openjdk is close ... I think we should wait since we will be supporting it for 5 more years, whatever we do. Open GPL'ed software is a much better approach I think.
Sorry for getting everyone's hopes up for 30 minutes
Is there any hopes in getting everything but java ready and provide a howto or script that will get java going, like maybe a nosrc rpm?
Scott Silva wrote:
Is there any hopes in getting everything but java ready and provide a howto or script that will get java going, like maybe a nosrc rpm?
That is definitely an option, do you want to look at that and propose a nosrc ?
Karanbir Singh wrote:
Scott Silva wrote:
Is there any hopes in getting everything but java ready and provide a howto or script that will get java going, like maybe a nosrc rpm?
That is definitely an option, do you want to look at that and propose a nosrc ?
I'm pretty sure the jpackage version for 4.x will still work for 5.x even though they don't have a yum repo for 5.x.
Or, you might ask the opennms guys if you could mirror their yum repo that has a working jdk rpm...
Johnny Hughes wrote:
And openjdk is close ... I think we should wait since we will be supporting it for 5 more years, whatever we do. Open GPL'ed software is a much better approach I think.
if it meets the requirements, absolutely!
to take a slightly different tack on this issue, would it make sense to put the rhwas rpms into a different repo than centosplus? since the rhwas stuff does come from upstream, it is a little less dangerous than the other things that go into centosplus.
Joe Pruett wrote:
to take a slightly different tack on this issue, would it make sense to put the rhwas rpms into a different repo than centosplus? since the rhwas stuff does come from upstream, it is a little less dangerous than the other things that go into centosplus.
the plan is to facilitate something that allows people to opt into the rhwas only or entire centos plus.
However, the rpms will be available in centosplus as well, to make sure there is an update path always available.
On Fri, 28 Mar 2008, Johnny Hughes wrote:
Johnny Hughes wrote:
Whether or not we can use this version of Java to work with the rhwas remains to be seen, but if it will work, we may be in business :D
Hmmm .. the indemnity clause is still there, though a bit lighter that the one I had major issues with.
Sorry for getting everyone's hopes up for 30 minutes
drat; double drat, as in testing neither openjava, nor the IBM implementations will run a closed source Java jar set I run daily.
Write once; run Sun's version
still is the rule of the day for me.
- Russ herrold
-
Akemi Yagi -
Bent Terp -
Brandon Davidson -
Daniel de Kok -
Hugo van der Kooij -
Ignacio Vazquez-Abrams -
Jean-Marc LIGER -
Joe Pruett -
John Summerfield -
Johnny Hughes -
Karanbir Singh -
Les Mikesell -
R P Herrold -
Ralph Angenendt -
Scott Silva -
Stephen John Smoogen -
William L. Maltby