You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@maven.apache.org by "fabrice.mercier1" <be...@yahoo.fr> on 2009/12/23 17:21:15 UTC
Re: dropping non-unique snapshots in Maven 3
Hi
Why do you think it is a bad practice ? I am currently trouble with the
choice of setting it to true or false.
Fabrice
Jason van Zyl-5 wrote:
>
> I think it can be in the same vein as the no versions for plugins. Not
> a very good idea and potentially harmful.
>
> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
> it's a bad practice in the majority of cases then we eliminate it.
> Over the life of 3.0 to 3.1 should be long enough. We can also start
> deprecating things like this in the 2.x line.
>
> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
>
>> I saw this on Benjamin's compat page yesterday and thought I'd throw
>> it up for discussion:
>> Non-unique Snapshot Deployments
>>
>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
>> repository has no effect in version 3.x, snapshot artifacts will
>> always be deployed using a timestamped version.
>>
>> I personally am not a fan of this feature as I've seen it cause more
>> problems than it solves. It also harkens back to a time when
>> Repository Managers didn't exist and seems like a work around to
>> cleaning up old snapshots. So my life would be easier if this didn't
>> exist.
>>
>> That said, I know of at least one instance where this is required
>> (will send a separate thread about that because it's a flaw in Maven
>> that should be addressed separately).
>>
>> With the focus on 3.0 being a drop in for M2, this clearly would cause
>> some heartache for users. Does anyone feel strongly enough that this
>> should go back? (and is willing to write the new code for it?)
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>
> Thanks,
>
> Jason
>
> ----------------------------------------------------------
> Jason van Zyl
> Founder, Apache Maven
> http://twitter.com/jvanzyl
> http://twitter.com/SonatypeNexus
> http://twitter.com/SonatypeM2E
> ----------------------------------------------------------
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
>
>
-----
http://old.nabble.com/file/u1297858/ft.gif
Architect
Almerys, activité santé d'Orange Business Services, TOULOUSE
http://www.orange-business.com http://www.orange-business.com
--
View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
Sent from the Maven Developers mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Brett Porter <br...@apache.org>.
On 12/01/2010, at 8:06 AM, Nigel Magnay wrote:
> How burnt-in to maven (and repo managers) is the snapshot resolution
> metadata ? Specifically, when I pull metadata like
>
> <?xml version="1.0" encoding="UTF-8"?>
> <metadata>
> <groupId>org.company</groupId>
> <artifactId>artifact</artifactId>
> <version>2.3-SNAPSHOT</version>
> <versioning>
> <snapshot>
> <timestamp>20081030.213329</timestamp>
> <buildNumber>5</buildNumber>
> </snapshot>
> <lastUpdated>20081030213342</lastUpdated>
> </versioning>
> </metadata>
>
> Does the resultant resolved snapshot *have* to be called
> artifact-${timestamp}-${buildnumber}.jar ?
ATM, yes.
>
> It would be useful if it could be configured to be artifact-${git
> sha1}.jar instead. Or even for it to be artifact-${uuid}, which is
> resolved in the <snapshot> section
>
> Even if it couldn't, an extensible tags element would be good; something like
> ...
> <snapshot>
> <timestamp>20081030.213329</timestamp>
> <buildNumber>5</buildNumber>
> <properties>
> <revision>df360a1e8df170892de84df8876aaf28e3815598</revision>
> <buildId>B2A7A1E7-FE24-4A81-B87C-88298C1D4CE5</buildId>
> <branch>master</branch>
> <hudson.project>Main-Build</hudson.project>
> <hudson.build.number>123</hudson.build.number>
> <flags>passed-integration, passed-regression</flags>
> ....
> <properties>
> </snapshot>
>
> Would that be difficult to add without breaking many things?
Not that hard in Maven itself, but the flow on to other tools would probably be significant. This is the same for any POM or metadata change where we need to consider backwards compat with Maven 2 and other tools. It's a good thing to consider for 3.1+ as we revisit repository metadata for things like classifiers, etc.
(In Archiva, we recently flipped to using a single ID in the metadata apart from the GAV which defaults to the whole filename. That way individual repository types and resolvers can construct from metadata and derive to metadata things like the type, classifier, artifact build version, etc however they like).
- Brett
--
Brett Porter
brett@apache.org
http://brettporter.wordpress.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Nigel Magnay <ni...@gmail.com>.
How burnt-in to maven (and repo managers) is the snapshot resolution
metadata ? Specifically, when I pull metadata like
<?xml version="1.0" encoding="UTF-8"?>
<metadata>
<groupId>org.company</groupId>
<artifactId>artifact</artifactId>
<version>2.3-SNAPSHOT</version>
<versioning>
<snapshot>
<timestamp>20081030.213329</timestamp>
<buildNumber>5</buildNumber>
</snapshot>
<lastUpdated>20081030213342</lastUpdated>
</versioning>
</metadata>
Does the resultant resolved snapshot *have* to be called
artifact-${timestamp}-${buildnumber}.jar ?
It would be useful if it could be configured to be artifact-${git
sha1}.jar instead. Or even for it to be artifact-${uuid}, which is
resolved in the <snapshot> section
Even if it couldn't, an extensible tags element would be good; something like
...
<snapshot>
<timestamp>20081030.213329</timestamp>
<buildNumber>5</buildNumber>
<properties>
<revision>df360a1e8df170892de84df8876aaf28e3815598</revision>
<buildId>B2A7A1E7-FE24-4A81-B87C-88298C1D4CE5</buildId>
<branch>master</branch>
<hudson.project>Main-Build</hudson.project>
<hudson.build.number>123</hudson.build.number>
<flags>passed-integration, passed-regression</flags>
....
<properties>
</snapshot>
Would that be difficult to add without breaking many things?
On Wed, Dec 23, 2009 at 11:55 PM, Brian Fox <br...@infinity.nu> wrote:
>
> On Wed, Dec 23, 2009 at 11:28 AM, Dan Tran <da...@gmail.com> wrote:
> > It is interesting that 3.0 apha x already drops this feature even thou
> > the discussion ( a while back ) asserted it would be deprecated in 3.0
> > and remove in 3.1
> >
>
> I would characterize it as this feature hasn't yet been re-implemented
> in 3.x. Since this area was rewritten from the ground up to simplify
> it, this was left out. AFAIK, it's just waiting for someone to scratch
> the itch and implement it.
>
> > However, as I have complained before ;-), this feature is crucially
> > needed for multi platforms ( native, c/c++) build of the same project.
> > I hope maven team could come up
> > with an alternative in 3.x
>
> Yes, this is a recognized problem. Enhancing the snapshot metadata is
> likely the best solution to this problem.
>
> >
> > -D
> >
> > On Wed, Dec 23, 2009 at 8:21 AM, fabrice.mercier1 <be...@yahoo.fr> wrote:
> >>
> >> Hi
> >>
> >> Why do you think it is a bad practice ? I am currently trouble with the
> >> choice of setting it to true or false.
> >>
> >> Fabrice
> >>
> >>
> >> Jason van Zyl-5 wrote:
> >>>
> >>> I think it can be in the same vein as the no versions for plugins. Not
> >>> a very good idea and potentially harmful.
> >>>
> >>> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
> >>> it's a bad practice in the majority of cases then we eliminate it.
> >>> Over the life of 3.0 to 3.1 should be long enough. We can also start
> >>> deprecating things like this in the 2.x line.
> >>>
> >>> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
> >>>
> >>>> I saw this on Benjamin's compat page yesterday and thought I'd throw
> >>>> it up for discussion:
> >>>> Non-unique Snapshot Deployments
> >>>>
> >>>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
> >>>> repository has no effect in version 3.x, snapshot artifacts will
> >>>> always be deployed using a timestamped version.
> >>>>
> >>>> I personally am not a fan of this feature as I've seen it cause more
> >>>> problems than it solves. It also harkens back to a time when
> >>>> Repository Managers didn't exist and seems like a work around to
> >>>> cleaning up old snapshots. So my life would be easier if this didn't
> >>>> exist.
> >>>>
> >>>> That said, I know of at least one instance where this is required
> >>>> (will send a separate thread about that because it's a flaw in Maven
> >>>> that should be addressed separately).
> >>>>
> >>>> With the focus on 3.0 being a drop in for M2, this clearly would cause
> >>>> some heartache for users. Does anyone feel strongly enough that this
> >>>> should go back? (and is willing to write the new code for it?)
> >>>>
> >>>> ---------------------------------------------------------------------
> >>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> >>>> For additional commands, e-mail: dev-help@maven.apache.org
> >>>>
> >>>
> >>> Thanks,
> >>>
> >>> Jason
> >>>
> >>> ----------------------------------------------------------
> >>> Jason van Zyl
> >>> Founder, Apache Maven
> >>> http://twitter.com/jvanzyl
> >>> http://twitter.com/SonatypeNexus
> >>> http://twitter.com/SonatypeM2E
> >>> ----------------------------------------------------------
> >>>
> >>>
> >>> ---------------------------------------------------------------------
> >>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> >>> For additional commands, e-mail: dev-help@maven.apache.org
> >>>
> >>>
> >>>
> >>
> >>
> >> -----
> >> http://old.nabble.com/file/u1297858/ft.gif
> >> Architect
> >> Almerys, activité santé d'Orange Business Services, TOULOUSE
> >> http://www.orange-business.com http://www.orange-business.com
> >> --
> >> View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
> >> Sent from the Maven Developers mailing list archive at Nabble.com.
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> >> For additional commands, e-mail: dev-help@maven.apache.org
> >>
> >>
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> > For additional commands, e-mail: dev-help@maven.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Brian Fox <br...@infinity.nu>.
On Wed, Dec 23, 2009 at 11:28 AM, Dan Tran <da...@gmail.com> wrote:
> It is interesting that 3.0 apha x already drops this feature even thou
> the discussion ( a while back ) asserted it would be deprecated in 3.0
> and remove in 3.1
>
I would characterize it as this feature hasn't yet been re-implemented
in 3.x. Since this area was rewritten from the ground up to simplify
it, this was left out. AFAIK, it's just waiting for someone to scratch
the itch and implement it.
> However, as I have complained before ;-), this feature is crucially
> needed for multi platforms ( native, c/c++) build of the same project.
> I hope maven team could come up
> with an alternative in 3.x
Yes, this is a recognized problem. Enhancing the snapshot metadata is
likely the best solution to this problem.
>
> -D
>
> On Wed, Dec 23, 2009 at 8:21 AM, fabrice.mercier1 <be...@yahoo.fr> wrote:
>>
>> Hi
>>
>> Why do you think it is a bad practice ? I am currently trouble with the
>> choice of setting it to true or false.
>>
>> Fabrice
>>
>>
>> Jason van Zyl-5 wrote:
>>>
>>> I think it can be in the same vein as the no versions for plugins. Not
>>> a very good idea and potentially harmful.
>>>
>>> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
>>> it's a bad practice in the majority of cases then we eliminate it.
>>> Over the life of 3.0 to 3.1 should be long enough. We can also start
>>> deprecating things like this in the 2.x line.
>>>
>>> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
>>>
>>>> I saw this on Benjamin's compat page yesterday and thought I'd throw
>>>> it up for discussion:
>>>> Non-unique Snapshot Deployments
>>>>
>>>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
>>>> repository has no effect in version 3.x, snapshot artifacts will
>>>> always be deployed using a timestamped version.
>>>>
>>>> I personally am not a fan of this feature as I've seen it cause more
>>>> problems than it solves. It also harkens back to a time when
>>>> Repository Managers didn't exist and seems like a work around to
>>>> cleaning up old snapshots. So my life would be easier if this didn't
>>>> exist.
>>>>
>>>> That said, I know of at least one instance where this is required
>>>> (will send a separate thread about that because it's a flaw in Maven
>>>> that should be addressed separately).
>>>>
>>>> With the focus on 3.0 being a drop in for M2, this clearly would cause
>>>> some heartache for users. Does anyone feel strongly enough that this
>>>> should go back? (and is willing to write the new code for it?)
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>>
>>>
>>> Thanks,
>>>
>>> Jason
>>>
>>> ----------------------------------------------------------
>>> Jason van Zyl
>>> Founder, Apache Maven
>>> http://twitter.com/jvanzyl
>>> http://twitter.com/SonatypeNexus
>>> http://twitter.com/SonatypeM2E
>>> ----------------------------------------------------------
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>
>>>
>>>
>>
>>
>> -----
>> http://old.nabble.com/file/u1297858/ft.gif
>> Architect
>> Almerys, activité santé d'Orange Business Services, TOULOUSE
>> http://www.orange-business.com http://www.orange-business.com
>> --
>> View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
>> Sent from the Maven Developers mailing list archive at Nabble.com.
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Jason van Zyl <ja...@maven.org>.
The metadata is missing lots of information, among which is the classifier.
On 2009-12-23, at 2:35 PM, Paul Benedict wrote:
> I thought it was discovered Maven 2/3 does not incorporate the
> artifact type in the snapshot metadata. So if you deployed artifact X
> type POM and artifact X type JAR, the latest snapshot metadata
> dictates the type. Wasn't an enhancement already requested?
>
> On Wed, Dec 23, 2009 at 1:23 PM, Jason van Zyl <ja...@maven.org> wrote:
>>
>> On 2009-12-23, at 11:28 AM, Dan Tran wrote:
>>
>>> It is interesting that 3.0 apha x already drops this feature even thou
>>> the discussion ( a while back ) asserted it would be deprecated in 3.0
>>> and remove in 3.1
>>>
>>
>> If you care that much and commit to supporting it put it back.
>>
>>> However, as I have complained before ;-), this feature is crucially
>>> needed for multi platforms ( native, c/c++) build of the same project.
>>> I hope maven team could come up
>>> with an alternative in 3.x
>>>
>>> -D
>>>
>>> On Wed, Dec 23, 2009 at 8:21 AM, fabrice.mercier1 <be...@yahoo.fr> wrote:
>>>>
>>>> Hi
>>>>
>>>> Why do you think it is a bad practice ? I am currently trouble with the
>>>> choice of setting it to true or false.
>>>>
>>>> Fabrice
>>>>
>>>>
>>>> Jason van Zyl-5 wrote:
>>>>>
>>>>> I think it can be in the same vein as the no versions for plugins. Not
>>>>> a very good idea and potentially harmful.
>>>>>
>>>>> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
>>>>> it's a bad practice in the majority of cases then we eliminate it.
>>>>> Over the life of 3.0 to 3.1 should be long enough. We can also start
>>>>> deprecating things like this in the 2.x line.
>>>>>
>>>>> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
>>>>>
>>>>>> I saw this on Benjamin's compat page yesterday and thought I'd throw
>>>>>> it up for discussion:
>>>>>> Non-unique Snapshot Deployments
>>>>>>
>>>>>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
>>>>>> repository has no effect in version 3.x, snapshot artifacts will
>>>>>> always be deployed using a timestamped version.
>>>>>>
>>>>>> I personally am not a fan of this feature as I've seen it cause more
>>>>>> problems than it solves. It also harkens back to a time when
>>>>>> Repository Managers didn't exist and seems like a work around to
>>>>>> cleaning up old snapshots. So my life would be easier if this didn't
>>>>>> exist.
>>>>>>
>>>>>> That said, I know of at least one instance where this is required
>>>>>> (will send a separate thread about that because it's a flaw in Maven
>>>>>> that should be addressed separately).
>>>>>>
>>>>>> With the focus on 3.0 being a drop in for M2, this clearly would cause
>>>>>> some heartache for users. Does anyone feel strongly enough that this
>>>>>> should go back? (and is willing to write the new code for it?)
>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>>>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>>>>
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Jason
>>>>>
>>>>> ----------------------------------------------------------
>>>>> Jason van Zyl
>>>>> Founder, Apache Maven
>>>>> http://twitter.com/jvanzyl
>>>>> http://twitter.com/SonatypeNexus
>>>>> http://twitter.com/SonatypeM2E
>>>>> ----------------------------------------------------------
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> -----
>>>> http://old.nabble.com/file/u1297858/ft.gif
>>>> Architect
>>>> Almerys, activité santé d'Orange Business Services, TOULOUSE
>>>> http://www.orange-business.com http://www.orange-business.com
>>>> --
>>>> View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
>>>> Sent from the Maven Developers mailing list archive at Nabble.com.
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>>
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>
>>
>> Thanks,
>>
>> Jason
>>
>> ----------------------------------------------------------
>> Jason van Zyl
>> Founder, Apache Maven
>> http://twitter.com/jvanzyl
>> ----------------------------------------------------------
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
Thanks,
Jason
----------------------------------------------------------
Jason van Zyl
Founder, Apache Maven
http://twitter.com/jvanzyl
----------------------------------------------------------
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Paul Benedict <pb...@apache.org>.
I thought it was discovered Maven 2/3 does not incorporate the
artifact type in the snapshot metadata. So if you deployed artifact X
type POM and artifact X type JAR, the latest snapshot metadata
dictates the type. Wasn't an enhancement already requested?
On Wed, Dec 23, 2009 at 1:23 PM, Jason van Zyl <ja...@maven.org> wrote:
>
> On 2009-12-23, at 11:28 AM, Dan Tran wrote:
>
>> It is interesting that 3.0 apha x already drops this feature even thou
>> the discussion ( a while back ) asserted it would be deprecated in 3.0
>> and remove in 3.1
>>
>
> If you care that much and commit to supporting it put it back.
>
>> However, as I have complained before ;-), this feature is crucially
>> needed for multi platforms ( native, c/c++) build of the same project.
>> I hope maven team could come up
>> with an alternative in 3.x
>>
>> -D
>>
>> On Wed, Dec 23, 2009 at 8:21 AM, fabrice.mercier1 <be...@yahoo.fr> wrote:
>>>
>>> Hi
>>>
>>> Why do you think it is a bad practice ? I am currently trouble with the
>>> choice of setting it to true or false.
>>>
>>> Fabrice
>>>
>>>
>>> Jason van Zyl-5 wrote:
>>>>
>>>> I think it can be in the same vein as the no versions for plugins. Not
>>>> a very good idea and potentially harmful.
>>>>
>>>> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
>>>> it's a bad practice in the majority of cases then we eliminate it.
>>>> Over the life of 3.0 to 3.1 should be long enough. We can also start
>>>> deprecating things like this in the 2.x line.
>>>>
>>>> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
>>>>
>>>>> I saw this on Benjamin's compat page yesterday and thought I'd throw
>>>>> it up for discussion:
>>>>> Non-unique Snapshot Deployments
>>>>>
>>>>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
>>>>> repository has no effect in version 3.x, snapshot artifacts will
>>>>> always be deployed using a timestamped version.
>>>>>
>>>>> I personally am not a fan of this feature as I've seen it cause more
>>>>> problems than it solves. It also harkens back to a time when
>>>>> Repository Managers didn't exist and seems like a work around to
>>>>> cleaning up old snapshots. So my life would be easier if this didn't
>>>>> exist.
>>>>>
>>>>> That said, I know of at least one instance where this is required
>>>>> (will send a separate thread about that because it's a flaw in Maven
>>>>> that should be addressed separately).
>>>>>
>>>>> With the focus on 3.0 being a drop in for M2, this clearly would cause
>>>>> some heartache for users. Does anyone feel strongly enough that this
>>>>> should go back? (and is willing to write the new code for it?)
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>>>
>>>>
>>>> Thanks,
>>>>
>>>> Jason
>>>>
>>>> ----------------------------------------------------------
>>>> Jason van Zyl
>>>> Founder, Apache Maven
>>>> http://twitter.com/jvanzyl
>>>> http://twitter.com/SonatypeNexus
>>>> http://twitter.com/SonatypeM2E
>>>> ----------------------------------------------------------
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>>
>>>>
>>>>
>>>
>>>
>>> -----
>>> http://old.nabble.com/file/u1297858/ft.gif
>>> Architect
>>> Almerys, activité santé d'Orange Business Services, TOULOUSE
>>> http://www.orange-business.com http://www.orange-business.com
>>> --
>>> View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
>>> Sent from the Maven Developers mailing list archive at Nabble.com.
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>
> Thanks,
>
> Jason
>
> ----------------------------------------------------------
> Jason van Zyl
> Founder, Apache Maven
> http://twitter.com/jvanzyl
> ----------------------------------------------------------
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Jason van Zyl <ja...@maven.org>.
On 2009-12-23, at 11:28 AM, Dan Tran wrote:
> It is interesting that 3.0 apha x already drops this feature even thou
> the discussion ( a while back ) asserted it would be deprecated in 3.0
> and remove in 3.1
>
If you care that much and commit to supporting it put it back.
> However, as I have complained before ;-), this feature is crucially
> needed for multi platforms ( native, c/c++) build of the same project.
> I hope maven team could come up
> with an alternative in 3.x
>
> -D
>
> On Wed, Dec 23, 2009 at 8:21 AM, fabrice.mercier1 <be...@yahoo.fr> wrote:
>>
>> Hi
>>
>> Why do you think it is a bad practice ? I am currently trouble with the
>> choice of setting it to true or false.
>>
>> Fabrice
>>
>>
>> Jason van Zyl-5 wrote:
>>>
>>> I think it can be in the same vein as the no versions for plugins. Not
>>> a very good idea and potentially harmful.
>>>
>>> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
>>> it's a bad practice in the majority of cases then we eliminate it.
>>> Over the life of 3.0 to 3.1 should be long enough. We can also start
>>> deprecating things like this in the 2.x line.
>>>
>>> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
>>>
>>>> I saw this on Benjamin's compat page yesterday and thought I'd throw
>>>> it up for discussion:
>>>> Non-unique Snapshot Deployments
>>>>
>>>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
>>>> repository has no effect in version 3.x, snapshot artifacts will
>>>> always be deployed using a timestamped version.
>>>>
>>>> I personally am not a fan of this feature as I've seen it cause more
>>>> problems than it solves. It also harkens back to a time when
>>>> Repository Managers didn't exist and seems like a work around to
>>>> cleaning up old snapshots. So my life would be easier if this didn't
>>>> exist.
>>>>
>>>> That said, I know of at least one instance where this is required
>>>> (will send a separate thread about that because it's a flaw in Maven
>>>> that should be addressed separately).
>>>>
>>>> With the focus on 3.0 being a drop in for M2, this clearly would cause
>>>> some heartache for users. Does anyone feel strongly enough that this
>>>> should go back? (and is willing to write the new code for it?)
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>>
>>>
>>> Thanks,
>>>
>>> Jason
>>>
>>> ----------------------------------------------------------
>>> Jason van Zyl
>>> Founder, Apache Maven
>>> http://twitter.com/jvanzyl
>>> http://twitter.com/SonatypeNexus
>>> http://twitter.com/SonatypeM2E
>>> ----------------------------------------------------------
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>
>>>
>>>
>>
>>
>> -----
>> http://old.nabble.com/file/u1297858/ft.gif
>> Architect
>> Almerys, activité santé d'Orange Business Services, TOULOUSE
>> http://www.orange-business.com http://www.orange-business.com
>> --
>> View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
>> Sent from the Maven Developers mailing list archive at Nabble.com.
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
Thanks,
Jason
----------------------------------------------------------
Jason van Zyl
Founder, Apache Maven
http://twitter.com/jvanzyl
----------------------------------------------------------
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Dan Tran <da...@gmail.com>.
It is interesting that 3.0 apha x already drops this feature even thou
the discussion ( a while back ) asserted it would be deprecated in 3.0
and remove in 3.1
However, as I have complained before ;-), this feature is crucially
needed for multi platforms ( native, c/c++) build of the same project.
I hope maven team could come up
with an alternative in 3.x
-D
On Wed, Dec 23, 2009 at 8:21 AM, fabrice.mercier1 <be...@yahoo.fr> wrote:
>
> Hi
>
> Why do you think it is a bad practice ? I am currently trouble with the
> choice of setting it to true or false.
>
> Fabrice
>
>
> Jason van Zyl-5 wrote:
>>
>> I think it can be in the same vein as the no versions for plugins. Not
>> a very good idea and potentially harmful.
>>
>> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
>> it's a bad practice in the majority of cases then we eliminate it.
>> Over the life of 3.0 to 3.1 should be long enough. We can also start
>> deprecating things like this in the 2.x line.
>>
>> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
>>
>>> I saw this on Benjamin's compat page yesterday and thought I'd throw
>>> it up for discussion:
>>> Non-unique Snapshot Deployments
>>>
>>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
>>> repository has no effect in version 3.x, snapshot artifacts will
>>> always be deployed using a timestamped version.
>>>
>>> I personally am not a fan of this feature as I've seen it cause more
>>> problems than it solves. It also harkens back to a time when
>>> Repository Managers didn't exist and seems like a work around to
>>> cleaning up old snapshots. So my life would be easier if this didn't
>>> exist.
>>>
>>> That said, I know of at least one instance where this is required
>>> (will send a separate thread about that because it's a flaw in Maven
>>> that should be addressed separately).
>>>
>>> With the focus on 3.0 being a drop in for M2, this clearly would cause
>>> some heartache for users. Does anyone feel strongly enough that this
>>> should go back? (and is willing to write the new code for it?)
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>
>>
>> Thanks,
>>
>> Jason
>>
>> ----------------------------------------------------------
>> Jason van Zyl
>> Founder, Apache Maven
>> http://twitter.com/jvanzyl
>> http://twitter.com/SonatypeNexus
>> http://twitter.com/SonatypeM2E
>> ----------------------------------------------------------
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>>
>>
>
>
> -----
> http://old.nabble.com/file/u1297858/ft.gif
> Architect
> Almerys, activité santé d'Orange Business Services, TOULOUSE
> http://www.orange-business.com http://www.orange-business.com
> --
> View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
> Sent from the Maven Developers mailing list archive at Nabble.com.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: dropping non-unique snapshots in Maven 3
Posted by Jason van Zyl <ja...@maven.org>.
On 2009-12-23, at 11:21 AM, fabrice.mercier1 wrote:
>
> Hi
>
> Why do you think it is a bad practice ? I am currently trouble with the
> choice of setting it to true or false.
>
You cannot uniquely identify one build from the next and just many other potentially bad practices.
> Fabrice
>
>
> Jason van Zyl-5 wrote:
>>
>> I think it can be in the same vein as the no versions for plugins. Not
>> a very good idea and potentially harmful.
>>
>> We put it in, deprecate it over 3.0 and it becomes taboo in 3.1. If
>> it's a bad practice in the majority of cases then we eliminate it.
>> Over the life of 3.0 to 3.1 should be long enough. We can also start
>> deprecating things like this in the 2.x line.
>>
>> On 2009-09-04, at 4:49 PM, Brian Fox wrote:
>>
>>> I saw this on Benjamin's compat page yesterday and thought I'd throw
>>> it up for discussion:
>>> Non-unique Snapshot Deployments
>>>
>>> The setting <uniqueVersion>false</uniqueVersion> for a distribution
>>> repository has no effect in version 3.x, snapshot artifacts will
>>> always be deployed using a timestamped version.
>>>
>>> I personally am not a fan of this feature as I've seen it cause more
>>> problems than it solves. It also harkens back to a time when
>>> Repository Managers didn't exist and seems like a work around to
>>> cleaning up old snapshots. So my life would be easier if this didn't
>>> exist.
>>>
>>> That said, I know of at least one instance where this is required
>>> (will send a separate thread about that because it's a flaw in Maven
>>> that should be addressed separately).
>>>
>>> With the focus on 3.0 being a drop in for M2, this clearly would cause
>>> some heartache for users. Does anyone feel strongly enough that this
>>> should go back? (and is willing to write the new code for it?)
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: dev-help@maven.apache.org
>>>
>>
>> Thanks,
>>
>> Jason
>>
>> ----------------------------------------------------------
>> Jason van Zyl
>> Founder, Apache Maven
>> http://twitter.com/jvanzyl
>> http://twitter.com/SonatypeNexus
>> http://twitter.com/SonatypeM2E
>> ----------------------------------------------------------
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
>> For additional commands, e-mail: dev-help@maven.apache.org
>>
>>
>>
>
>
> -----
> http://old.nabble.com/file/u1297858/ft.gif
> Architect
> Almerys, activité santé d'Orange Business Services, TOULOUSE
> http://www.orange-business.com http://www.orange-business.com
> --
> View this message in context: http://old.nabble.com/dropping-non-unique-snapshots-in-Maven-3-tp25295809p26904205.html
> Sent from the Maven Developers mailing list archive at Nabble.com.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
Thanks,
Jason
----------------------------------------------------------
Jason van Zyl
Founder, Apache Maven
http://twitter.com/jvanzyl
----------------------------------------------------------
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org