Andre I think that solution strikes the right balance without being a lot of effort.
As mentioned Travis-ci's docs say that caching stuff like .m2 doesn't help but in my experience you usually still get a little bump so it's worth it as long as the ci isn't lying about the results. Removing ~/.m2/org/apache/nifi should minimize that, so good. > On Mar 25, 2017, at 6:48 PM, Andre <andre-li...@fucs.org> wrote: > > Aldrin, > > While we may not get much granularity we can certainly hack our way by "rm > -rf ~/.m2/repository/whatever_we_want_to_delete" prior to calling maven ? > > I appreciate caching is sometime a pain, but given our build currently > takes almost 40 minutes and travis-ci.org jobs will timeout at 50 minutes, > I suspect the time saved by not having to download dependencies comes very > handy. > > I would suggest that we keep caching on (the thing that happened with > jB(C|c)rypt is not that usual after all...) but remove what we are testing > (i.e. ~/.m2/repository/org/nifi ) prior to build, what do you think? > > Cheers > > >> On Sun, Mar 26, 2017 at 3:38 AM, Aldrin Piri <aldrinp...@gmail.com> wrote: >> >> Awesome, thanks for fixing it up, Bryan. >> >> I don't think we can get that kind of granularity with Travis, >> unfortunately. However, the last time was because an artifact changed its >> name (or more specifically, casing). >> >> Not sure removing caching is the best option, but seems like the the >> optimization may not provide as much value in build speedup as the >> consistent, cleanroom environment for builds. Just a thought. >> >>> On Sat, Mar 25, 2017 at 11:18 AM, Bryan Bende <bbe...@gmail.com> wrote: >>> >>> Thanks Aldrin, I pushed the commit. >>> >>> As far as travis, I am not familiar with how it works, but can you >>> specify what to cache? >>> >>> In this case we didn't need a completely clean .m2, we just needed a >>> clean .m2/org/apache/nifi. >>> >>> On Sat, Mar 25, 2017 at 11:10 AM, Aldrin Piri <aldrinp...@gmail.com> >>> wrote: >>>> Please just push to correct. Simple fixes are fine in my book. >>>> >>>> Does it make sense to potentially scrap caching in Travis? >>>> >>>> This is another time we have missed something like this that no caching >>>> would have prevented. Additionally, given the large footprint of the >>>> repository download it seems as though its benefit may be marginal as >> per >>>> https://docs.travis-ci.com/user/caching/#How-does-caching-work%3F >>>>> On Sat, Mar 25, 2017 at 11:04 Bryan Bende <bbe...@gmail.com> wrote: >>>>> >>>>> Andre, >>>>> >>>>> Thanks for bringing this up. >>>>> >>>>> The standard prioritizers moved from the standard bundle to the >>>>> framework bundle, and sure enough the parent was still set as standard >>>>> bundle. We had never built with a clean repo and were getting lucky, >>>>> so I am glad you found this. >>>>> >>>>> In >>>>> nifi-nar-bundles/nifi-framework-bundle/nifi-framework/nifi-standard- >>> prioritizers/pom.xml >>>>> >>>>> The parent should be <artifactId>nifi-framework</artifactId> and not >>>>> <artifactId>nifi-standard-bundle</artifactId>. >>>>> >>>>> If no one objects I can push up this change, not sure if we need a >>>>> formal PR for a one line change for something thats already broken?? >>>>> >>>>> Thanks, >>>>> >>>>> Bryan >>>>> >>>>>> On Sat, Mar 25, 2017 at 9:33 AM, Andre <andre-li...@fucs.org> wrote: >>>>>> dev, >>>>>> >>>>>> Is anyone else having issues building master from "clean" (i.e. rm >> -rf >>>>>> ~/.m2/repositories/org/apache/nifi) after commit d90cf84 ? >>>>>> >>>>>> My attempts to build currently yield: >>>>>> >>>>>> $ mvn -T2.0C -DskipTests=true -Pdir-only clean install >>>>>> [INFO] Scanning for projects... >>>>>> [ERROR] [ERROR] Some problems were encountered while processing the >>> POMs: >>>>>> [WARNING] 'parent.relativePath' of POM >>>>>> org.apache.nifi:nifi-standard-prioritizers:[unknown-version] >>>>>> >>>>> (/home/afucs/nifi/nifi-nar-bundles/nifi-framework-bundle/ >>> nifi-framework/nifi-standard-prioritizers/pom.xml) >>>>>> points at org.apache.nifi:nifi-framework instead of >>>>>> org.apache.nifi:nifi-standard-bundle, please verify your project >>>>> structure >>>>>> @ line 17, column 13 >>>>>> [FATAL] Non-resolvable parent POM for >>>>>> org.apache.nifi:nifi-standard-prioritizers:[unknown-version]: Could >>> not >>>>>> find artifact org.apache.nifi:nifi-standard- >> bundle:pom:1.2.0-SNAPSHOT >>> and >>>>>> 'parent.relativePath' points at wrong local POM @ line 17, column 13 >>>>>> @ >>>>>> [ERROR] The build could not read 1 project -> [Help 1] >>>>>> [ERROR] >>>>>> [ERROR] The project >>>>>> org.apache.nifi:nifi-standard-prioritizers:[unknown-version] >>>>>> >>>>> (/home/afucs/nifi/nifi-nar-bundles/nifi-framework-bundle/ >>> nifi-framework/nifi-standard-prioritizers/pom.xml) >>>>>> has 1 error >>>>>> [ERROR] Non-resolvable parent POM for >>>>>> org.apache.nifi:nifi-standard-prioritizers:[unknown-version]: Could >>> not >>>>>> find artifact org.apache.nifi:nifi-standard- >> bundle:pom:1.2.0-SNAPSHOT >>> and >>>>>> 'parent.relativePath' points at wrong local POM @ line 17, column 13 >>> -> >>>>>> [Help 2] >>>>>> [ERROR] >>>>>> [ERROR] To see the full stack trace of the errors, re-run Maven with >>> the >>>>> -e >>>>>> switch. >>>>>> [ERROR] Re-run Maven using the -X switch to enable full debug >> logging. >>>>>> [ERROR] >>>>>> [ERROR] For more information about the errors and possible >> solutions, >>>>>> please read the following articles: >>>>>> [ERROR] [Help 1] >>>>>> >>>>> http://cwiki.apache.org/confluence/display/MAVEN/ >>> ProjectBuildingException >>>>>> [ERROR] [Help 2] >>>>>> >>>>> http://cwiki.apache.org/confluence/display/MAVEN/ >>> UnresolvableModelException >>>>>> >>>>>> >>>>>> However, when I interactively rebase the branch and dro pthe above >>>>>> mentioned commit, build starts working again. >>>>>> >>>>>> >>>>>> Cheers >>>>> >>> >>