| Re: [glassfish-dev] [External] : Re: How to build glassfish in less than an hour? |
Hi,
On Thu, Nov 4, 2021 at 3:50 PM Lukas Jungmann <lukas.jungmann@xxxxxxxxxx> wrote:
> by default the Embedded distro was severely neglected up to the point it
> didn't even work anymore.
building embedded by default is the right thing to do.
>
> There's a few other things worth considering. We now painstakingly build
> javatest.jar, since it's not available in Maven Central (or any other
> repo). An alternative would be to make sure this is available in Maven
> Central.
is javatest glassfish specific? Would it make sense to move it to its
own repo or at least let it have separate lifecycle instead?
Javatest is not glassfish specific, but some of the tests use it. Unfortunately it's not available in Maven Central, so GlassFish downloads an archive that contains it from Eclipse (the authorization TCK archive) and then grabs that zip file from it.
I've asked Scott (cc'ed) before for this to be moved to Maven Central. Since the original code is Oracle and used by the JDK, I figure this will not be solved within a short timeframe. So in the meantime we have to do alternative solutions.
I don't see any mention of Maven on
https://wiki.openjdk.java.net/display/CodeTools/JT+Harness and
https://github.com/openjdk/jtharness (javatest).
I'm not certain but https://bugs.openjdk.java.net/projects/CODETOOLS may be the issue tracker for https://github.com/openjdk/jtharness.
If someone wants to contribute a change to javatest, from my experience contributing changes to the JDK signature test tooling, you need to sign the contributor agreement before your pull request can be reviewed. So, if there are contributors that are feeling the pain caused by javatest not being available on Maven Central that are also OpenJDK contributors, you could contribute a javatest change to support Maven (perhaps a maven pom for pushing javatest releases to Maven Central).
https://github.com/jtulach/netbeans-apitest is a fork of the
Signature Test tooling that I ended up contributing to instead of
following the OpenJDK path, as a workaround for not being an
OpenJDK contributor.
Hope this helps,
Scott
What about having extra profiles the other way around - be fast (=dev
friendly) by default and have profiles for extra tests/checks instead?
Yes, this could be considered for sure, if the other GlassFish committers agree with that.
Kind regards,Arjan
> On Wed, Nov 3, 2021 at 10:35 PM Lukas Jungmann
> <lukas.jungmann@xxxxxxxxxx <mailto:lukas.jungmann@xxxxxxxxxx>> wrote:
>
> Hi,
>
> On 11/3/21 5:54 PM, arjan tijms wrote:
> > Hi,
> >
> > The easiest option, which we could opt to make a default (or some
> > variant of it), is the -Pfastest in combination with parallel
> building.
> >
> > e.g.
> >
> > mvn clean install -Pfastest -T10
>
> an average Joe (=me :-) ) reads the README at most (should he ever find
> it), asking him to read man pages or know the tools (especially -?
> option which can become very handy) is often too much...
>
> Would it make sense to put this info to the project's readme?
>
> thanks,
> --lukas
>
> >
> > iMac Pro 2017 (3 GHz 10-Core Intel Xeon W): *5:03*
> > MacBook Pro 2021 (M1 Max): 03:40 first time and largely
> empty .m2, *3:20
> > *second time.
> >
> > Some further performance improvements *could* be obtained by
> skippings
> > tests and javadoc, and by using a RAM DISK
> >
> > diskutil erasevolume HFS+ RAM_Disk $(hdiutil attach -nomount
> ram://1243840)
> >
> > mvn clean install -Dglassfish.root=/Volumes/RAM_Disk/
> > -DbuildDirectory='/Volumes/RAM_Disk/target/${project.artifactId}'
> > -DskipTests -Dmaven.test.skip=true
> -Dmaven.javadoc.skip=true -Pfastest -T10
> >
> > I've seen this reducing time on some systems by about 40%, but it of
> > course depends on how fast your storage is and how many cores
> there are.
> >
> > Kind regards,
> > Arjan Tijms
> >
> >
> >
> >
> >
> > On Wed, Nov 3, 2021 at 4:30 PM Scott Marlow <smarlow@xxxxxxxxxx
> <mailto:smarlow@xxxxxxxxxx>
> > <mailto:smarlow@xxxxxxxxxx <mailto:smarlow@xxxxxxxxxx>>> wrote:
> >
> >
> > On 11/1/21 1:50 PM, Lukas Jungmann wrote:
> >> Hi,
> >>
> >> I need to do some changes in GF codebase in order to test my
> >> work. Building GF with mvn -U -C clean install -Pstaging
> >> -Dcheckstyle.skip=true is taking me ~90 - ~135 mins today and it
> >> seems I'm not the only one seeing the build getting slower over
> >> past ~6 months[1]. What do I have to do in order to get the
> build
> >> time on my box to something reasonable again?
> >
> > Skipping tests can reduce build time by adding:
> >
> > -DskipTests=true
> >
> >>
> >> thanks,
> >> --lukas
> >>
> >> [1]: https://github.com/eclipse-ee4j/glassfish/issues/23692
> <https://urldefense.com/v3/__https://github.com/eclipse-ee4j/glassfish/issues/23692__;!!ACWV5N9M2RV99hQ!bqoQ7nFyEX4ZBrsEroChHo8cUUvvhI-Hzoe5kmjm5ETbH5DxdFEEB2jFbDrceVApdwQ$>
> >>
> <https://urldefense.com/v3/__https://github.com/eclipse-ee4j/glassfish/issues/23692__;!!ACWV5N9M2RV99hQ!deqLtJCT5O9-J3FAY_hu6R_b1dxFOHMdQUXKDhURLwHAq-11Co4JJ-oW4gFYvEYBeg0$ <https://urldefense.com/v3/__https://github.com/eclipse-ee4j/glassfish/issues/23692__;!!ACWV5N9M2RV99hQ!deqLtJCT5O9-J3FAY_hu6R_b1dxFOHMdQUXKDhURLwHAq-11Co4JJ-oW4gFYvEYBeg0$>>
> >>
> >> _______________________________________________
> >> glassfish-dev mailing list
> >> glassfish-dev@xxxxxxxxxxx <mailto:glassfish-dev@xxxxxxxxxxx>
> <mailto:glassfish-dev@xxxxxxxxxxx <mailto:glassfish-dev@xxxxxxxxxxx>>
> >> To unsubscribe from this list, visit
> >> https://www.eclipse.org/mailman/listinfo/glassfish-dev
> <https://urldefense.com/v3/__https://www.eclipse.org/mailman/listinfo/glassfish-dev__;!!ACWV5N9M2RV99hQ!bqoQ7nFyEX4ZBrsEroChHo8cUUvvhI-Hzoe5kmjm5ETbH5DxdFEEB2jFbDrcEUZpVgg$>
> >>
> <https://urldefense.com/v3/__https://www.eclipse.org/mailman/listinfo/glassfish-dev__;!!ACWV5N9M2RV99hQ!deqLtJCT5O9-J3FAY_hu6R_b1dxFOHMdQUXKDhURLwHAq-11Co4JJ-oW4gFY4I2nT68$ <https://urldefense.com/v3/__https://www.eclipse.org/mailman/listinfo/glassfish-dev__;!!ACWV5N9M2RV99hQ!deqLtJCT5O9-J3FAY_hu6R_b1dxFOHMdQUXKDhURLwHAq-11Co4JJ-oW4gFY4I2nT68$>>
> >>
> >>
> > _______________________________________________
> > glassfish-dev mailing list
> > glassfish-dev@xxxxxxxxxxx <mailto:glassfish-dev@xxxxxxxxxxx>
> <mailto:glassfish-dev@xxxxxxxxxxx <mailto:glassfish-dev@xxxxxxxxxxx>>
> > To unsubscribe from this list, visit
> > https://www.eclipse.org/mailman/listinfo/glassfish-dev
> <https://urldefense.com/v3/__https://www.eclipse.org/mailman/listinfo/glassfish-dev__;!!ACWV5N9M2RV99hQ!bqoQ7nFyEX4ZBrsEroChHo8cUUvvhI-Hzoe5kmjm5ETbH5DxdFEEB2jFbDrcEUZpVgg$>
> >
> <https://urldefense.com/v3/__https://www.eclipse.org/mailman/listinfo/glassfish-dev__;!!ACWV5N9M2RV99hQ!deqLtJCT5O9-J3FAY_hu6R_b1dxFOHMdQUXKDhURLwHAq-11Co4JJ-oW4gFY4I2nT68$ <https://urldefense.com/v3/__https://www.eclipse.org/mailman/listinfo/glassfish-dev__;!!ACWV5N9M2RV99hQ!deqLtJCT5O9-J3FAY_hu6R_b1dxFOHMdQUXKDhURLwHAq-11Co4JJ-oW4gFY4I2nT68$>>
> >
>