If the compiler doesn't reliably detect changes, then it is a HUGE problem!

All the later stages can only depend on the results of the previous ones. To 
not trigger a necessary rebuild is WAY worse than to rebuild a few classes too 
much.

Exactly that is the reason why most people always use "mvn CLEAN install" all 
the time currently.

LieGrue,
strub


>________________________________
> From: Jason van Zyl <[email protected]>
>To: Maven Developers List <[email protected]>; Mark Struberg 
><[email protected]> 
>Sent: Thursday, August 30, 2012 3:11 PM
>Subject: Re: Removing unused code from maven-shared-utils
> 
>
>
>
>On Aug 30, 2012, at 9:07 AM, Mark Struberg wrote:
>
>If you are attempting to remove plexus-utils initially without a 
>>>
>>
>>binary compatible drop-in replacement with code from Apache then I'm -1. 
>>>
>>Please re-read my previous mail. We will NOT drop plexus-utils! We will just 
>>move maven core off it for most parts. 
>>
>>
>
>
>Hmm, what's the difference exactly?
>
>
>>
>>we've had incremental build support in m2e forever
>>>You don't see most of the 'incremental' problems in m2 because the recompile 
>>>already gets done by eclipse itself I assume. E.g. the maven-compiler-plugin 
>>>until recently failed to detect cross-class relations and did only recompile 
>>>the file which got changed. I changed that only recently.
>>
>>
>
>
>The compiler is not the problem, even for large code bases the compiler is 
>incredibly fast. Profile your average build and it's not the compilation 
>that's the bottleneck. But again I would argue you don't need to make drastic 
>changes in the core to make incremental builds work. Some, but they are few.
>
>
>>LieGrue,
>>strub
>>
>>
>>
>>________________________________
>>>
>>From: Jason van Zyl <[email protected]>
>>>
>>To: Maven Developers List <[email protected]>; Mark Struberg 
>><[email protected]> 
>>>
>>Sent: Thursday, August 30, 2012 2:49 PM
>>>
>>Subject: Re: Removing unused code from maven-shared-utils
>>>
>>
>>>
>>You need to do one thing at time and not conflate the replacement of the 
>>plexus-utils code with anything else you want to implement. Mixing the two 
>>will almost certainly lead to problems.
>>>
>>
>>>
>>If you are attempting to remove plexus-utils initially without a binary 
>>compatible drop-in replacement with code from Apache then I'm -1. I'm not 
>>going to change that position because I think you're needlessly causing 
>>problems. If you want to replace the code go for it, just do it safely.
>>>
>>
>>>
>>Incremental builds are a separate matter all together and can be done in a 
>>branch, we've had incremental build support in m2e forever and you don't need 
>>to massively change the core at all to do it. In fact it works right now with 
>>the core as it is. Put a complete proposal together (I've seen the initial 
>>ideas) because if you think you need to drastically change the core for 
>>incremental builds I believe you are incorrect.
>>>
>>
>>>
>>At any rate, one thing at a time. If you want to replace the code in 
>>plexus-utils with something else do that first as a separate, clear endeavor.
>>>
>>
>>>
>>On Aug 30, 2012, at 8:40 AM, Mark Struberg wrote:
>>>
>>
>>>
>>I fear my perspective is pretty different. 
>>>>
>>
>>>>
>>
>>>>
>>To _not_ produce a 1:1 drop in replacement gives us a few benefits. Actually 
>>it doesn't matter which version maven itself uses as this wont affect user 
>>builds. 
>>>>
>>
>>>>
>>BUT it would _heavily_ affect users if their old builds wont work anymore 
>>because we dropped some ancient signature. 
>>>>
>>
>>>>
>>And there is a well cast technical reason for this change: I need 
>>new/improved functionality for the incremental build.
>>>>
>>
>>>>
>>
>>>>
>>My proposal is to move maven itself to a much more progressive/improved 
>>version (thus off plexus-utils) but keep the old signatures and the original 
>>plexus-utils around for plugins which still uses it. 
>>>>
>>
>>>>
>>If a plugin likes to use the new maven-core-utils then this is perfectly 
>>possible as well of course.
>>>>
>>
>>>>
>>LieGrue,
>>>>
>>strub
>>>>
>>
>>>>
>>________________________________
>>>>>
>>From: Jason van Zyl <[email protected]>
>>>>>
>>To: Maven Developers List <[email protected]>; Mark Struberg 
>><[email protected]> 
>>>>>
>>Sent: Thursday, August 30, 2012 2:24 PM
>>>>>
>>Subject: Re: Removing unused code from maven-shared-utils
>>>>>
>>
>>>>>
>>
>>>>>
>>If you can create a drop-in replacement for the plexus-utils code go for it. 
>>I believe this exercise is of little value users and you run the real risk of 
>>introducing problems for no technical reason, but if you want to do the work 
>>that's your prerogative.
>>>>>
>>
>>>>>
>>
>>>>>
>>So much time was spent to preserve the existing signatures and behaviour in 
>>the container swap and this is not something to be taken lightly, and you 
>>should do what has been done in the past.
>>>>>
>>
>>>>>
>>
>>>>>
>>I think the path to do this if you want to get rid of the plexus-utils 
>>reference is to reduce the foot print to the classes being used in the core 
>>and then drop in that reduced JAR. Create a replacement with the code you've 
>>copied, reimplemented, improved yet binary compatible and test that for a 
>>while. Binary compatible in all respects, package and signature. And from 
>>there then make any subsequent changes.
>>>>>
>>
>>>>>
>>
>>>>>
>>I honestly am not overly concerned with plugins because they can be fixed 
>>relatively quickly, but doing low value changes in the core for swapping out 
>>one set of code for another is seriously playing with fire.
>>>>>
>>
>>>>>
>>
>>>>>
>>If you initially are intending to change code in the core to accommodate this 
>>change I see no technical reason for it, is dangerous and I'm -1.
>>>>>
>>
>>>>>
>>
>>>>>
>>On Aug 30, 2012, at 8:03 AM, Mark Struberg wrote:
>>>>>
>>
>>>>>
>>While digging thru the plexus-utils usage I wonder whether we should place 
>>the maven-utils in maven-core or maven-shared.
>>>>>
>>
>>>>>>
>>maven-core doesn't yet have any maven-shared dependency it seems. Which means 
>>if we like to use them in maven-core as well we should relocate 
>>maven-shared-utils to the maven-3 core module.
>>>>>>
>>
>>>>>>
>>wdyt?
>>>>>>
>>
>>>>>>
>>LieGrue,
>>>>>>
>>strub
>>>>>>
>>
>>>>>>
>>
>>>>>>
>>
>>>>>>
>>
>>>>>>
>>----- Original Message -----
>>>>>>
>>
>>>>>>
>>From: Kristian Rosenvold <[email protected]>
>>>>>>
>>
>>>>>>>
>>To: Maven Developers List <[email protected]>
>>>>>>
>>
>>>>>>>
>>Cc: 
>>>>>>
>>
>>>>>>>
>>Sent: Thursday, August 30, 2012 10:48 AM
>>>>>>
>>
>>>>>>>
>>Subject: Re: Removing unused code from maven-shared-utils
>>>>>>
>>
>>>>>>>
>>
>>>>>>
>>
>>>>>>>
>>Well obviously given the current number of failing
>>>>>>
>>
>>>>>>>
>>tests,maven-shared-utils is going nowhere right now ;)
>>>>>>
>>
>>>>>>>
>>
>>>>>>
>>
>>>>>>>
>>I will do a test-migration and remove unused code in
>>>>>>
>>
>>>>>>>
>>org.apache.maven.shared.utils.io before we release.
>>>>>>
>>
>>>>>>>
>>
>>>>>>
>>
>>>>>>>
>>Kristian
>>>>>>
>>
>>>>>>>
>>
>>>>>>
>>
>>>>>>>
>>---------------------------------------------------------------------
>>>>>>
>>
>>>>>>>
>>To unsubscribe, e-mail: [email protected]
>>>>>>
>>
>>>>>>>
>>For additional commands, e-mail: [email protected]
>>>>>>
>>
>>>>>>>
>>
>>>>>>
>>
>>>>>>>
>>---------------------------------------------------------------------
>>>>>>
>>To unsubscribe, e-mail: [email protected]
>>>>>>
>>For additional commands, e-mail: [email protected]
>>>>>>
>>
>>>>>>
>>
>>>>>>
>>
>>>>>
>>Thanks,
>>>>>
>>
>>>>>
>>Jason
>>>>>
>>
>>>>>
>>----------------------------------------------------------
>>>>>
>>Jason van Zyl
>>>>>
>>Founder & CTO, Sonatype
>>>>>
>>Founder,  Apache Maven
>>>>>
>>http://twitter.com/jvanzyl
>>>>>
>>---------------------------------------------------------
>>>>>
>>
>>>>>
>>
>>>>>
>>What matters is not ideas, but the people who have them. Good people can fix 
>>bad ideas, but good ideas can't save bad people. 
>>>>>
>>
>>>>>
>>  -- Paul Graham
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>>
>>
>>>>
>>---------------------------------------------------------------------
>>>>
>>To unsubscribe, e-mail: [email protected]
>>>>
>>For additional commands, e-mail: [email protected]
>>>>
>>
>>>>
>>
>>>
>>Thanks,
>>>
>>
>>>
>>Jason
>>>
>>
>>>
>>----------------------------------------------------------
>>>
>>Jason van Zyl
>>>
>>Founder & CTO, Sonatype
>>>
>>Founder,  Apache Maven
>>>
>>http://twitter.com/jvanzyl
>>>
>>---------------------------------------------------------
>>>
>>
>>>
>>happiness is like a butterfly: the more you chase it, the more it will
>>>
>>elude you, but if you turn your attention to other things, it will come
>>>
>>and sit softly on your shoulder ...
>>>
>>
>>>
>>-- Thoreau 
>>>
>>
>>>
>>
>>>
>>
>>>
>>
>>>
>>
>>>
>>
>>>
>>
>>>
>>
>>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: [email protected]
>>For additional commands, e-mail: [email protected]
>>
>>
>
>Thanks,
>
>Jason
>
>----------------------------------------------------------
>Jason van Zyl
>Founder & CTO, Sonatype
>Founder,  Apache Maven
>http://twitter.com/jvanzyl
>---------------------------------------------------------
>
>
>Three people can keep a secret provided two of them are dead.
>
> -- Benjamin Franklin
>
>
>
>
>
>
>
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to