JDK 7 build 27 is available at the openjdk.java.net website

[xiomara.jayasena]

The OpenJDK source is available at:
http://hg.openjdk.java.net/jdk7/jdk7
and the build 27 source is here:
http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328

The OpenJDK source binary plugs and Jtreg binary for the promoted JDK 7
build 27 are available under the openjdk http://openjdk.java.net website
under Source Code (direct link to bundles:
http://download.java.net/openjdk/jdk7)

Summary of changes:
http://download.java.net/jdk7/changes/jdk7-b27.html

-Xiomara

[Rob Ross-3]

What is the difference between the code in

(A) jdk7/jdk7
(B) jkd7/rev/56652b46f328?

I thought (A) was the "master" repository with the latest "build"  
code (not latest snapshot though) but I think I may just be confused.  
If (A) does not contain build 27 code, what codes does it contain?

Thanks,


Rob Ross, Lead Software Engineer
E! Networks

---------------------------------------------------
"Beware of he who would deny you access to information, for in his  
heart he dreams himself your master." -- Commissioner Pravin Lal



On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:

>
> The OpenJDK source is available at:
> http://hg.openjdk.java.net/jdk7/jdk7
> and the build 27 source is here:
> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>
> The OpenJDK source binary plugs and Jtreg binary for the promoted  
> JDK 7 build 27 are available under the openjdk http://
> openjdk.java.net website under Source Code (direct link to bundles:  
> http://download.java.net/openjdk/jdk7)
>
> Summary of changes:
> http://download.java.net/jdk7/changes/jdk7-b27.html
>
> -Xiomara
>

[xiomara.jayasena]

There is no difference in the code at this point.  Once integrations
start to flow into the master repositories then the two below will
defer, hence the reason to include the change set rev for this
particular build number.  In the future at any time if someone wants to
pull a specific build number then they would just pull the change set
rev for that build number or do the following:
 

hg fclone -r jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7 myjdk7


Regards,
-Xiomara


Rob Ross wrote:

> What is the difference between the code in
>
> (A) jdk7/jdk7
> (B) jkd7/rev/56652b46f328?
>
> I thought (A) was the "master" repository with the latest "build" code
> (not latest snapshot though) but I think I may just be confused. If
> (A) does not contain build 27 code, what codes does it contain?
>
> Thanks,
>
>
> Rob Ross, Lead Software Engineer
> E! Networks
>
> ---------------------------------------------------
> "Beware of he who would deny you access to information, for in his
> heart he dreams himself your master." -- Commissioner Pravin Lal
>
>
>
> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>
>>
>> The OpenJDK source is available at:
>> http://hg.openjdk.java.net/jdk7/jdk7
>> and the build 27 source is here:
>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>
>> The OpenJDK source binary plugs and Jtreg binary for the promoted JDK
>> 7 build 27 are available under the openjdk http://openjdk.java.net 
>> website under Source Code (direct link to bundles:
>> http://download.java.net/openjdk/jdk7)
>>
>> Summary of changes:
>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>
>> -Xiomara
>>
>

[Kelly O'Hair]

http://hg.openjdk.java.net/jdk7/jdk7  is the root of the repository
forest, but you can only see the top repository.
Using:
    hg fclone http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7
would get you the entire forest.

http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a reference to
a specific changeset in the top level repository. And this changeset
is the one that Xiomara created to define the jdk7-b27 tag so we
have a permanent record of what changesets were in jdk7 build 27.
Each repository in the forest will have a similar changeset, which just
creates this jdk7-b27 tag.

Using this jdk7-b27 tag, you could
      hg fclone -R jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7-b27
if you wanted to get the repository forest and the exact state of the sources
when Xiomara did the build for jdk7 Build 27. A handy feature.

-kto

Rob Ross wrote:

> What is the difference between the code in
>
> (A) jdk7/jdk7
> (B) jkd7/rev/56652b46f328?
>
> I thought (A) was the "master" repository with the latest "build" code
> (not latest snapshot though) but I think I may just be confused. If (A)
> does not contain build 27 code, what codes does it contain?
>
> Thanks,
>
>
> Rob Ross, Lead Software Engineer
> E! Networks
>
> ---------------------------------------------------
> "Beware of he who would deny you access to information, for in his heart
> he dreams himself your master." -- Commissioner Pravin Lal
>
>
>
> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>
>>
>> The OpenJDK source is available at:
>> http://hg.openjdk.java.net/jdk7/jdk7
>> and the build 27 source is here:
>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>
>> The OpenJDK source binary plugs and Jtreg binary for the promoted JDK
>> 7 build 27 are available under the openjdk http://openjdk.java.net 
>> website under Source Code (direct link to bundles:
>> http://download.java.net/openjdk/jdk7)
>>
>> Summary of changes:
>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>
>> -Xiomara
>>
>

[Kelly O'Hair]

Kelly O'Hair wrote:

> http://hg.openjdk.java.net/jdk7/jdk7  is the root of the repository
> forest, but you can only see the top repository.
> Using:
>    hg fclone http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7
> would get you the entire forest.
>
> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a reference to
> a specific changeset in the top level repository. And this changeset
> is the one that Xiomara created to define the jdk7-b27 tag so we
> have a permanent record of what changesets were in jdk7 build 27.
> Each repository in the forest will have a similar changeset, which just
> creates this jdk7-b27 tag.
>
> Using this jdk7-b27 tag, you could
>      hg fclone -R jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7 

                  ^^
                  -r
oops.

-kto

> yourjdk7-b27
> if you wanted to get the repository forest and the exact state of the
> sources
> when Xiomara did the build for jdk7 Build 27. A handy feature.
>
> -kto
>
> Rob Ross wrote:
>> What is the difference between the code in
>>
>> (A) jdk7/jdk7
>> (B) jkd7/rev/56652b46f328?
>>
>> I thought (A) was the "master" repository with the latest "build" code
>> (not latest snapshot though) but I think I may just be confused. If
>> (A) does not contain build 27 code, what codes does it contain?
>>
>> Thanks,
>>
>>
>> Rob Ross, Lead Software Engineer
>> E! Networks
>>
>> ---------------------------------------------------
>> "Beware of he who would deny you access to information, for in his
>> heart he dreams himself your master." -- Commissioner Pravin Lal
>>
>>
>>
>> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>>
>>>
>>> The OpenJDK source is available at:
>>> http://hg.openjdk.java.net/jdk7/jdk7
>>> and the build 27 source is here:
>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>>
>>> The OpenJDK source binary plugs and Jtreg binary for the promoted JDK
>>> 7 build 27 are available under the openjdk http://openjdk.java.net 
>>> website under Source Code (direct link to bundles:
>>> http://download.java.net/openjdk/jdk7)
>>>
>>> Summary of changes:
>>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>>
>>> -Xiomara
>>>
>>

[Rob Ross-3]

Sorry to ask again, but I'm trying to learn how this works :)

I had read through the archives regarding the repository process  
flow, and know there are several levels with gatekeeper repositories  
before the "master" repository, which I thought was at jdk7/jdk. I  
also thought that code does not get checked into this repository  
until it's gone through several gatekeeper repositories and been  
tested, reviewed, etc. So things would only get checked into master  
in well defined circumstances.

If my understanding of the above is not correct than what follows is  
also not correct. I understand about creating a changeset tag to be  
able to recreate that code in that moment in time, but I didn't  
understand where Xiomara said

"There is no difference in the code at this point.  Once integrations  
start to flow into the master repositories then the two below will  
defer, hence the reason to include the change set rev for this  
particular build number."

Wouldn't any changes that flow into the master repositories be based  
on code originally part of this tag, that had been tested, modified  
if necessary, and eventually approved? So any such changes in the  
master would be (excluding regressions) a positive development?

This is the opposite of how other SCM systems work, where the head/
top is usually in an unstable state due to active checkins, and you  
want to pull a stable release tag that has been created to mark a  
milestone during development - is this correct?


Rob Ross, Lead Software Engineer
E! Networks

---------------------------------------------------
"Beware of he who would deny you access to information, for in his  
heart he dreams himself your master." -- Commissioner Pravin Lal



On May 23, 2008, at 10:05 AM, Kelly O'Hair wrote:

>
>
> Kelly O'Hair wrote:
>> http://hg.openjdk.java.net/jdk7/jdk7  is the root of the repository
>> forest, but you can only see the top repository.
>> Using:
>>    hg fclone http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7
>> would get you the entire forest.
>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a  
>> reference to
>> a specific changeset in the top level repository. And this changeset
>> is the one that Xiomara created to define the jdk7-b27 tag so we
>> have a permanent record of what changesets were in jdk7 build 27.
>> Each repository in the forest will have a similar changeset, which  
>> just
>> creates this jdk7-b27 tag.
>> Using this jdk7-b27 tag, you could
>>      hg fclone -R jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7
>                  ^^
>                  -r
> oops.
>
> -kto
>
>> yourjdk7-b27
>> if you wanted to get the repository forest and the exact state of  
>> the sources
>> when Xiomara did the build for jdk7 Build 27. A handy feature.
>> -kto
>> Rob Ross wrote:
>>> What is the difference between the code in
>>>
>>> (A) jdk7/jdk7
>>> (B) jkd7/rev/56652b46f328?
>>>
>>> I thought (A) was the "master" repository with the latest "build"  
>>> code (not latest snapshot though) but I think I may just be  
>>> confused. If (A) does not contain build 27 code, what codes does  
>>> it contain?
>>>
>>> Thanks,
>>>
>>>
>>> Rob Ross, Lead Software Engineer
>>> E! Networks
>>>
>>> ---------------------------------------------------
>>> "Beware of he who would deny you access to information, for in  
>>> his heart he dreams himself your master." -- Commissioner Pravin Lal
>>>
>>>
>>>
>>> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>>>
>>>>
>>>> The OpenJDK source is available at:
>>>> http://hg.openjdk.java.net/jdk7/jdk7
>>>> and the build 27 source is here:
>>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>>>
>>>> The OpenJDK source binary plugs and Jtreg binary for the  
>>>> promoted JDK 7 build 27 are available under the openjdk http://
>>>> openjdk.java.net website under Source Code (direct link to  
>>>> bundles: http://download.java.net/openjdk/jdk7)
>>>>
>>>> Summary of changes:
>>>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>>>
>>>> -Xiomara
>>>>
>>>

[xiomara.jayasena]

Rob Ross wrote:

> Sorry to ask again, but I'm trying to learn how this works :)
>
> I had read through the archives regarding the repository process flow,
> and know there are several levels with gatekeeper repositories before
> the "master" repository, which I thought was at jdk7/jdk. I also
> thought that code does not get checked into this repository until it's
> gone through several gatekeeper repositories and been tested,
> reviewed, etc. So things would only get checked into master in well
> defined circumstances.
>
> If my understanding of the above is not correct than what follows is
> also not correct. I understand about creating a changeset tag to be
> able to recreate that code in that moment in time, but I didn't
> understand where Xiomara said
>
> "There is no difference in the code at this point.  Once integrations
> start to flow into the master repositories then the two below will
> defer, hence the reason to include the change set rev for this
> particular build number."

To clarify I was answering the question is there a difference between:

(A) jdk7/jdk7
(B) jkd7/rev/56652b46f328?

and the answer is no, there is no difference.  A clone of each of those
will yield the same source.

>
> Wouldn't any changes that flow into the master repositories be based
> on code originally part of this tag, that had been tested, modified if
> necessary, and eventually approved? So any such changes in the master
> would be (excluding regressions) a positive development?
>
> This is the opposite of how other SCM systems work, where the head/top
> is usually in an unstable state due to active checkins, and you want
> to pull a stable release tag that has been created to mark a milestone
> during development - is this correct?

Right, a stable release tag for this build is jdk7-b27, which
corresponds to changeset 56652b46f328.


The forest  will continue to evolve as checkins continue to go in, as
you mentioned above.

Regards,
-Xiomara

>
>
> Rob Ross, Lead Software Engineer
> E! Networks
>
> ---------------------------------------------------
> "Beware of he who would deny you access to information, for in his
> heart he dreams himself your master." -- Commissioner Pravin Lal
>
>
>
> On May 23, 2008, at 10:05 AM, Kelly O'Hair wrote:
>
>>
>>
>> Kelly O'Hair wrote:
>>> http://hg.openjdk.java.net/jdk7/jdk7  is the root of the repository
>>> forest, but you can only see the top repository.
>>> Using:
>>>    hg fclone http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7
>>> would get you the entire forest.
>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a
>>> reference to
>>> a specific changeset in the top level repository. And this changeset
>>> is the one that Xiomara created to define the jdk7-b27 tag so we
>>> have a permanent record of what changesets were in jdk7 build 27.
>>> Each repository in the forest will have a similar changeset, which just
>>> creates this jdk7-b27 tag.
>>> Using this jdk7-b27 tag, you could
>>>      hg fclone -R jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7
>>                  ^^
>>                  -r
>> oops.
>>
>> -kto
>>
>>> yourjdk7-b27
>>> if you wanted to get the repository forest and the exact state of
>>> the sources
>>> when Xiomara did the build for jdk7 Build 27. A handy feature.
>>> -kto
>>> Rob Ross wrote:
>>>> What is the difference between the code in
>>>>
>>>> (A) jdk7/jdk7
>>>> (B) jkd7/rev/56652b46f328?
>>>>
>>>> I thought (A) was the "master" repository with the latest "build"
>>>> code (not latest snapshot though) but I think I may just be
>>>> confused. If (A) does not contain build 27 code, what codes does it
>>>> contain?
>>>>
>>>> Thanks,
>>>>
>>>>
>>>> Rob Ross, Lead Software Engineer
>>>> E! Networks
>>>>
>>>> ---------------------------------------------------
>>>> "Beware of he who would deny you access to information, for in his
>>>> heart he dreams himself your master." -- Commissioner Pravin Lal
>>>>
>>>>
>>>>
>>>> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>>>>
>>>>>
>>>>> The OpenJDK source is available at:
>>>>> http://hg.openjdk.java.net/jdk7/jdk7
>>>>> and the build 27 source is here:
>>>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>>>>
>>>>> The OpenJDK source binary plugs and Jtreg binary for the promoted
>>>>> JDK 7 build 27 are available under the openjdk
>>>>> http://openjdk.java.net website under Source Code (direct link to
>>>>> bundles: http://download.java.net/openjdk/jdk7)
>>>>>
>>>>> Summary of changes:
>>>>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>>>>
>>>>> -Xiomara
>>>>>
>>>>
>

[Kelly O'Hair]

The most stable sources (changesets) will be those tagged by Xiomara
and that represent the promoted build.

As various teams push changesets into the master area jdk7/jdk7
(http://hg.openjdk.java.net/jdk7/jdk7) these changesets should be
solid but they have not gone through the release engineering build
promotion process like the promoted build changesets.
Also, because of the timing of the pushes, it's rare but possible that
doing a forest pull from all the jdk7/jdk7 forest could get
some but not all of the changesets a team may have pushed into this
forest. Like I said, this is an extremely rare event.

Pushes of changesets into the jdk7/jdk7 forest only happens a few
times a day, based on integration slots listed here:
   http://openjdk.java.net/projects/jdk7/builds/
These time slots are dedicated to the team doing the integration,
allowing them reserved access to the jdk7/jdk7 forest for uninterrupted
pulls, merges, and the pushes.

Hope this helps.

-kto

Rob Ross wrote:

> Sorry to ask again, but I'm trying to learn how this works :)
>
> I had read through the archives regarding the repository process flow,
> and know there are several levels with gatekeeper repositories before
> the "master" repository, which I thought was at jdk7/jdk. I also thought
> that code does not get checked into this repository until it's gone
> through several gatekeeper repositories and been tested, reviewed, etc.
> So things would only get checked into master in well defined circumstances.
>
> If my understanding of the above is not correct than what follows is
> also not correct. I understand about creating a changeset tag to be able
> to recreate that code in that moment in time, but I didn't understand
> where Xiomara said
>
> "There is no difference in the code at this point.  Once integrations
> start to flow into the master repositories then the two below will
> defer, hence the reason to include the change set rev for this
> particular build number."
>
> Wouldn't any changes that flow into the master repositories be based on
> code originally part of this tag, that had been tested, modified if
> necessary, and eventually approved? So any such changes in the master
> would be (excluding regressions) a positive development?
>
> This is the opposite of how other SCM systems work, where the head/top
> is usually in an unstable state due to active checkins, and you want to
> pull a stable release tag that has been created to mark a milestone
> during development - is this correct?
>
>
> Rob Ross, Lead Software Engineer
> E! Networks
>
> ---------------------------------------------------
> "Beware of he who would deny you access to information, for in his heart
> he dreams himself your master." -- Commissioner Pravin Lal
>
>
>
> On May 23, 2008, at 10:05 AM, Kelly O'Hair wrote:
>
>>
>>
>> Kelly O'Hair wrote:
>>> http://hg.openjdk.java.net/jdk7/jdk7  is the root of the repository
>>> forest, but you can only see the top repository.
>>> Using:
>>>    hg fclone http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7
>>> would get you the entire forest.
>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a reference to
>>> a specific changeset in the top level repository. And this changeset
>>> is the one that Xiomara created to define the jdk7-b27 tag so we
>>> have a permanent record of what changesets were in jdk7 build 27.
>>> Each repository in the forest will have a similar changeset, which just
>>> creates this jdk7-b27 tag.
>>> Using this jdk7-b27 tag, you could
>>>      hg fclone -R jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7
>>                  ^^
>>                  -r
>> oops.
>>
>> -kto
>>
>>> yourjdk7-b27
>>> if you wanted to get the repository forest and the exact state of the
>>> sources
>>> when Xiomara did the build for jdk7 Build 27. A handy feature.
>>> -kto
>>> Rob Ross wrote:
>>>> What is the difference between the code in
>>>>
>>>> (A) jdk7/jdk7
>>>> (B) jkd7/rev/56652b46f328?
>>>>
>>>> I thought (A) was the "master" repository with the latest "build"
>>>> code (not latest snapshot though) but I think I may just be
>>>> confused. If (A) does not contain build 27 code, what codes does it
>>>> contain?
>>>>
>>>> Thanks,
>>>>
>>>>
>>>> Rob Ross, Lead Software Engineer
>>>> E! Networks
>>>>
>>>> ---------------------------------------------------
>>>> "Beware of he who would deny you access to information, for in his
>>>> heart he dreams himself your master." -- Commissioner Pravin Lal
>>>>
>>>>
>>>>
>>>> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>>>>
>>>>>
>>>>> The OpenJDK source is available at:
>>>>> http://hg.openjdk.java.net/jdk7/jdk7
>>>>> and the build 27 source is here:
>>>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>>>>
>>>>> The OpenJDK source binary plugs and Jtreg binary for the promoted
>>>>> JDK 7 build 27 are available under the openjdk
>>>>> http://openjdk.java.net website under Source Code (direct link to
>>>>> bundles: http://download.java.net/openjdk/jdk7)
>>>>>
>>>>> Summary of changes:
>>>>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>>>>
>>>>> -Xiomara
>>>>>
>>>>
>

[Rob Ross-3]

Yes this helps.

So it turns out this process is similar to other SCM processes after  
all. The "top" level represents the latest (in this case pre-vetted)  
changes, but tagged changesets (repositories?) actually represent  
stable/milestone builds.


Rob Ross, Lead Software Engineer
E! Networks

---------------------------------------------------
"Beware of he who would deny you access to information, for in his  
heart he dreams himself your master." -- Commissioner Pravin Lal



On May 23, 2008, at 12:10 PM, Kelly O'Hair wrote:

> The most stable sources (changesets) will be those tagged by Xiomara
> and that represent the promoted build.
>
> As various teams push changesets into the master area jdk7/jdk7
> (http://hg.openjdk.java.net/jdk7/jdk7) these changesets should be
> solid but they have not gone through the release engineering build
> promotion process like the promoted build changesets.
> Also, because of the timing of the pushes, it's rare but possible that
> doing a forest pull from all the jdk7/jdk7 forest could get
> some but not all of the changesets a team may have pushed into this
> forest. Like I said, this is an extremely rare event.
>
> Pushes of changesets into the jdk7/jdk7 forest only happens a few
> times a day, based on integration slots listed here:
>   http://openjdk.java.net/projects/jdk7/builds/
> These time slots are dedicated to the team doing the integration,
> allowing them reserved access to the jdk7/jdk7 forest for  
> uninterrupted
> pulls, merges, and the pushes.
>
> Hope this helps.
>
> -kto
>
> Rob Ross wrote:
>> Sorry to ask again, but I'm trying to learn how this works :)
>> I had read through the archives regarding the repository process  
>> flow, and know there are several levels with gatekeeper  
>> repositories before the "master" repository, which I thought was  
>> at jdk7/jdk. I also thought that code does not get checked into  
>> this repository until it's gone through several gatekeeper  
>> repositories and been tested, reviewed, etc. So things would only  
>> get checked into master in well defined circumstances.
>> If my understanding of the above is not correct than what follows  
>> is also not correct. I understand about creating a changeset tag  
>> to be able to recreate that code in that moment in time, but I  
>> didn't understand where Xiomara said
>> "There is no difference in the code at this point.  Once  
>> integrations start to flow into the master repositories then the  
>> two below will defer, hence the reason to include the change set  
>> rev for this particular build number."
>> Wouldn't any changes that flow into the master repositories be  
>> based on code originally part of this tag, that had been tested,  
>> modified if necessary, and eventually approved? So any such  
>> changes in the master would be (excluding regressions) a positive  
>> development?
>> This is the opposite of how other SCM systems work, where the head/
>> top is usually in an unstable state due to active checkins, and  
>> you want to pull a stable release tag that has been created to  
>> mark a milestone during development - is this correct?
>> Rob Ross, Lead Software Engineer
>> E! Networks
>> ---------------------------------------------------
>> "Beware of he who would deny you access to information, for in his  
>> heart he dreams himself your master." -- Commissioner Pravin Lal
>> On May 23, 2008, at 10:05 AM, Kelly O'Hair wrote:
>>>
>>>
>>> Kelly O'Hair wrote:
>>>> http://hg.openjdk.java.net/jdk7/jdk7  is the root of the repository
>>>> forest, but you can only see the top repository.
>>>> Using:
>>>>    hg fclone http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7
>>>> would get you the entire forest.
>>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a  
>>>> reference to
>>>> a specific changeset in the top level repository. And this  
>>>> changeset
>>>> is the one that Xiomara created to define the jdk7-b27 tag so we
>>>> have a permanent record of what changesets were in jdk7 build 27.
>>>> Each repository in the forest will have a similar changeset,  
>>>> which just
>>>> creates this jdk7-b27 tag.
>>>> Using this jdk7-b27 tag, you could
>>>>      hg fclone -R jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7
>>>                  ^^
>>>                  -r
>>> oops.
>>>
>>> -kto
>>>
>>>> yourjdk7-b27
>>>> if you wanted to get the repository forest and the exact state  
>>>> of the sources
>>>> when Xiomara did the build for jdk7 Build 27. A handy feature.
>>>> -kto
>>>> Rob Ross wrote:
>>>>> What is the difference between the code in
>>>>>
>>>>> (A) jdk7/jdk7
>>>>> (B) jkd7/rev/56652b46f328?
>>>>>
>>>>> I thought (A) was the "master" repository with the latest  
>>>>> "build" code (not latest snapshot though) but I think I may  
>>>>> just be confused. If (A) does not contain build 27 code, what  
>>>>> codes does it contain?
>>>>>
>>>>> Thanks,
>>>>>
>>>>>
>>>>> Rob Ross, Lead Software Engineer
>>>>> E! Networks
>>>>>
>>>>> ---------------------------------------------------
>>>>> "Beware of he who would deny you access to information, for in  
>>>>> his heart he dreams himself your master." -- Commissioner  
>>>>> Pravin Lal
>>>>>
>>>>>
>>>>>
>>>>> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>>>>>
>>>>>>
>>>>>> The OpenJDK source is available at:
>>>>>> http://hg.openjdk.java.net/jdk7/jdk7
>>>>>> and the build 27 source is here:
>>>>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>>>>>
>>>>>> The OpenJDK source binary plugs and Jtreg binary for the  
>>>>>> promoted JDK 7 build 27 are available under the openjdk http://
>>>>>> openjdk.java.net website under Source Code (direct link to  
>>>>>> bundles: http://download.java.net/openjdk/jdk7)
>>>>>>
>>>>>> Summary of changes:
>>>>>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>>>>>
>>>>>> -Xiomara
>>>>>>
>>>>>

[Kelly O'Hair]

The major difference is that the individual developer pushes changesets
to their team area, so the absolute freshest changesets, are down in
the team forests. But those fresh changesets will likely not be tested
as well as the ones living up in jdk7/jdk7.

The chance that bad changesets will get into the jdk7/jdk7 forest is
extremely slim, not impossible of course.

-kto

Rob Ross wrote:

> Yes this helps.
>
> So it turns out this process is similar to other SCM processes after
> all. The "top" level represents the latest (in this case pre-vetted)
> changes, but tagged changesets (repositories?) actually represent
> stable/milestone builds.
>
>
> Rob Ross, Lead Software Engineer
> E! Networks
>
> ---------------------------------------------------
> "Beware of he who would deny you access to information, for in his heart
> he dreams himself your master." -- Commissioner Pravin Lal
>
>
>
> On May 23, 2008, at 12:10 PM, Kelly O'Hair wrote:
>
>> The most stable sources (changesets) will be those tagged by Xiomara
>> and that represent the promoted build.
>>
>> As various teams push changesets into the master area jdk7/jdk7
>> (http://hg.openjdk.java.net/jdk7/jdk7) these changesets should be
>> solid but they have not gone through the release engineering build
>> promotion process like the promoted build changesets.
>> Also, because of the timing of the pushes, it's rare but possible that
>> doing a forest pull from all the jdk7/jdk7 forest could get
>> some but not all of the changesets a team may have pushed into this
>> forest. Like I said, this is an extremely rare event.
>>
>> Pushes of changesets into the jdk7/jdk7 forest only happens a few
>> times a day, based on integration slots listed here:
>>   http://openjdk.java.net/projects/jdk7/builds/
>> These time slots are dedicated to the team doing the integration,
>> allowing them reserved access to the jdk7/jdk7 forest for uninterrupted
>> pulls, merges, and the pushes.
>>
>> Hope this helps.
>>
>> -kto
>>
>> Rob Ross wrote:
>>> Sorry to ask again, but I'm trying to learn how this works :)
>>> I had read through the archives regarding the repository process
>>> flow, and know there are several levels with gatekeeper repositories
>>> before the "master" repository, which I thought was at jdk7/jdk. I
>>> also thought that code does not get checked into this repository
>>> until it's gone through several gatekeeper repositories and been
>>> tested, reviewed, etc. So things would only get checked into master
>>> in well defined circumstances.
>>> If my understanding of the above is not correct than what follows is
>>> also not correct. I understand about creating a changeset tag to be
>>> able to recreate that code in that moment in time, but I didn't
>>> understand where Xiomara said
>>> "There is no difference in the code at this point.  Once integrations
>>> start to flow into the master repositories then the two below will
>>> defer, hence the reason to include the change set rev for this
>>> particular build number."
>>> Wouldn't any changes that flow into the master repositories be based
>>> on code originally part of this tag, that had been tested, modified
>>> if necessary, and eventually approved? So any such changes in the
>>> master would be (excluding regressions) a positive development?
>>> This is the opposite of how other SCM systems work, where the
>>> head/top is usually in an unstable state due to active checkins, and
>>> you want to pull a stable release tag that has been created to mark a
>>> milestone during development - is this correct?
>>> Rob Ross, Lead Software Engineer
>>> E! Networks
>>> ---------------------------------------------------
>>> "Beware of he who would deny you access to information, for in his
>>> heart he dreams himself your master." -- Commissioner Pravin Lal
>>> On May 23, 2008, at 10:05 AM, Kelly O'Hair wrote:
>>>>
>>>>
>>>> Kelly O'Hair wrote:
>>>>> http://hg.openjdk.java.net/jdk7/jdk7  is the root of the repository
>>>>> forest, but you can only see the top repository.
>>>>> Using:
>>>>>    hg fclone http://hg.openjdk.java.net/jdk7/jdk7 yourjdk7
>>>>> would get you the entire forest.
>>>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a
>>>>> reference to
>>>>> a specific changeset in the top level repository. And this changeset
>>>>> is the one that Xiomara created to define the jdk7-b27 tag so we
>>>>> have a permanent record of what changesets were in jdk7 build 27.
>>>>> Each repository in the forest will have a similar changeset, which
>>>>> just
>>>>> creates this jdk7-b27 tag.
>>>>> Using this jdk7-b27 tag, you could
>>>>>      hg fclone -R jdk7-b27 http://hg.openjdk.java.net/jdk7/jdk7
>>>>                  ^^
>>>>                  -r
>>>> oops.
>>>>
>>>> -kto
>>>>
>>>>> yourjdk7-b27
>>>>> if you wanted to get the repository forest and the exact state of
>>>>> the sources
>>>>> when Xiomara did the build for jdk7 Build 27. A handy feature.
>>>>> -kto
>>>>> Rob Ross wrote:
>>>>>> What is the difference between the code in
>>>>>>
>>>>>> (A) jdk7/jdk7
>>>>>> (B) jkd7/rev/56652b46f328?
>>>>>>
>>>>>> I thought (A) was the "master" repository with the latest "build"
>>>>>> code (not latest snapshot though) but I think I may just be
>>>>>> confused. If (A) does not contain build 27 code, what codes does
>>>>>> it contain?
>>>>>>
>>>>>> Thanks,
>>>>>>
>>>>>>
>>>>>> Rob Ross, Lead Software Engineer
>>>>>> E! Networks
>>>>>>
>>>>>> ---------------------------------------------------
>>>>>> "Beware of he who would deny you access to information, for in his
>>>>>> heart he dreams himself your master." -- Commissioner Pravin Lal
>>>>>>
>>>>>>
>>>>>>
>>>>>> On May 23, 2008, at 8:04 AM, Xiomara Jayasena wrote:
>>>>>>
>>>>>>>
>>>>>>> The OpenJDK source is available at:
>>>>>>> http://hg.openjdk.java.net/jdk7/jdk7
>>>>>>> and the build 27 source is here:
>>>>>>> http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328
>>>>>>>
>>>>>>> The OpenJDK source binary plugs and Jtreg binary for the promoted
>>>>>>> JDK 7 build 27 are available under the openjdk
>>>>>>> http://openjdk.java.net website under Source Code (direct link to
>>>>>>> bundles: http://download.java.net/openjdk/jdk7)
>>>>>>>
>>>>>>> Summary of changes:
>>>>>>> http://download.java.net/jdk7/changes/jdk7-b27.html
>>>>>>>
>>>>>>> -Xiomara
>>>>>>>
>>>>>>
>

[bradford.wetmore]

I was reading between the lines here just to make sure the right answers
are coming out.

> To clarify I was answering the question is there a difference between:
>
> (A) jdk7/jdk7
> (B) jkd7/rev/56652b46f328?
>
> and the answer is no, there is no difference.  A clone of each of those
> will yield the same source.

Assuming that 56652b46f328 is still the topmost rev when you clone...

 > The "top" level represents the latest (in this case pre-vetted)
 > changes, but tagged changesets (repositories?) actually represent
 > stable/milestone builds.

Yes.

Warning: hg technical content ahead.

The build tag changesets that Xiomara is creating are just like any
other changeset that go into any repository:  the only file modification
that is normally being made here is to the .hgtags file, which is at the
top directory of every repository's tree (<topdir>/.hgtags).  This file
is just a "list of version pointers".  For example, the top most one in
jdk7/jdk used to contain:

     cfeea66a3fa8ca3686a7cfa2d0ce8ab0169f168d jdk7-b24
     cbc8ad9dd0e085a607427ea35411990982f19a36 jdk7-b25
     9410f77cc30c604d1caf7c9fe3a57fa19e1acbe8 jdk7-b26

The b27 tag change added the following line:

     11b4dc9f2be3523ef989a0db8459eb56b3045c3a jdk7-b27

(Note, this version uses the "long" version of the hg changeset id.  The
short version is just the first 12 chars)

For most hg users, you should stop reading here.

 > http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a reference
 > to a specific changeset in the top level repository. And this
 > changeset is the one that Xiomara created to define the jdk7-b27 tag
 > so we have a permanent record of what changesets were in jdk7 build
 > 27.

Refer to:

changeset:   13:56652b46f328
tag:         tip
user:        xdono
date:        Thu May 22 09:37:37 2008 -0700
files:       .hgtags
description:
Added tag jdk7-b27 for changeset 11b4dc9f2be3

changeset:   12:11b4dc9f2be3
tag:         jdk7-b27
parent:      11:0f440f3321f5
parent:      10:613dea62de17
user:        xdono
date:        Tue May 13 11:31:45 2008 -0700
description:
Merge

Note that "clone -r ver" doesn't not actually bring the *FULL*
repository down, only up to the version you asked for.  If you asked for
"-r jdk7-b27", it will only bring down revision 11b4dc9f2be3, not
56652b46f328 which is the version that actually modified the .hgtags
files.  That was surprising to me at first.  A "hg clone -r jdk7-b27"
followed by a "hg in" will show there is still a change (56652b46f328)
incoming.

Hope this helps, and not confuses!

Brad

[Weijun Wang]

IMHO, It's like after a child is born, the parents go to an authority  
to do a birth registration. Two actions. the second one tags the  
first one. The first one is the important one (and will be remembered  
by the family).

Max

On May 27, 2008, at 6:36 AM, Brad Wetmore wrote:

> I was reading between the lines here just to make sure the right  
> answers are coming out.
>
>> To clarify I was answering the question is there a difference  
>> between:
>> (A) jdk7/jdk7
>> (B) jkd7/rev/56652b46f328?
>> and the answer is no, there is no difference.  A clone of each of  
>> those will yield the same source.
>
> Assuming that 56652b46f328 is still the topmost rev when you clone...
>
> > The "top" level represents the latest (in this case pre-vetted)
> > changes, but tagged changesets (repositories?) actually represent
> > stable/milestone builds.
>
> Yes.
>
> Warning: hg technical content ahead.
>
> The build tag changesets that Xiomara is creating are just like any  
> other changeset that go into any repository:  the only file  
> modification that is normally being made here is to the .hgtags  
> file, which is at the top directory of every repository's tree  
> (<topdir>/.hgtags).  This file is just a "list of version  
> pointers".  For example, the top most one in jdk7/jdk used to contain:
>
>     cfeea66a3fa8ca3686a7cfa2d0ce8ab0169f168d jdk7-b24
>     cbc8ad9dd0e085a607427ea35411990982f19a36 jdk7-b25
>     9410f77cc30c604d1caf7c9fe3a57fa19e1acbe8 jdk7-b26
>
> The b27 tag change added the following line:
>
>     11b4dc9f2be3523ef989a0db8459eb56b3045c3a jdk7-b27
>
> (Note, this version uses the "long" version of the hg changeset  
> id.  The short version is just the first 12 chars)
>
> For most hg users, you should stop reading here.
>
> > http://hg.openjdk.java.net/jdk7/jdk7/rev/56652b46f328  is a  
> reference
> > to a specific changeset in the top level repository. And this
> > changeset is the one that Xiomara created to define the jdk7-b27 tag
> > so we have a permanent record of what changesets were in jdk7 build
> > 27.
>
> Refer to:
>
> changeset:   13:56652b46f328
> tag:         tip
> user:        xdono
> date:        Thu May 22 09:37:37 2008 -0700
> files:       .hgtags
> description:
> Added tag jdk7-b27 for changeset 11b4dc9f2be3
>
> changeset:   12:11b4dc9f2be3
> tag:         jdk7-b27
> parent:      11:0f440f3321f5
> parent:      10:613dea62de17
> user:        xdono
> date:        Tue May 13 11:31:45 2008 -0700
> description:
> Merge
>
> Note that "clone -r ver" doesn't not actually bring the *FULL*  
> repository down, only up to the version you asked for.  If you  
> asked for "-r jdk7-b27", it will only bring down revision  
> 11b4dc9f2be3, not 56652b46f328 which is the version that actually  
> modified the .hgtags files.  That was surprising to me at first.  A  
> "hg clone -r jdk7-b27" followed by a "hg in" will show there is  
> still a change (56652b46f328) incoming.
>
> Hope this helps, and not confuses!
>
> Brad