I’m wondering if in that case simply removing
<version>${project.royale.compiler.version}</version> from the child pom
wouldn’t cause mvn to pick up the parent pom version definition, but I’m not
messing with that now.
I hope to get us out of this limbo state soon. Thanks for fixing Hudson build
in the meantime.
From: Josh Tynjala<mailto:joshtynj...@bowlerhat.dev>
Sent: Saturday, June 19, 2021 2:40 AM
To: Apache Royale Development<mailto:dev@royale.apache.org>
Subject: Re: Hudson Nightly build as-js failure
I don't claim to be a Maven expert, but I think I understand how the
recursion is happening.
The main pom.xml has these values:
<version>0.9.8-SNAPSHOT</version>
<project.royale.compiler.version>${project.version}</project.royale.compiler.version>
And the child module distribution/jars/**/pom.xml files had this value
(these were the only poms in royale-asjs where the version was not
concrete, by the way):
<version>${project.royale.compiler.version}</version>
The parent's project.version is 0.9.8-SNAPSHOT, but project.version is
${project.royale.compiler.version} in the child modules.
When
<project.royale.compiler.version>${project.version}</project.royale.compiler.version>
is interpreted by the child module, it doesn't use the parent's
project.version. Instead, it uses its own, so it's essentially
becoming something like this:
<project.royale.compiler.version>${project.royale.compiler.version}</project.royale.compiler.version>
That's definitely recursion.
We didn't have recursion in the past because the parent pom.xml used to
have this:
<project.royale.compiler.version>0.9.8-SNAPSHOT</project.royale.compiler.version>
My original motivation for changing it to ${project.version} was because
creating a release branch left project.royale.compiler.version referencing
the old version, which put the new 0.9.9-SNAPSHOT in a bad state. Now,
${project.version} is a sensible default that keeps things updated
automatically, but anyone can also change it to a concrete version locally,
if they need to do that for any reason (either on the command line or by
modifying the main pom.xml on their computer).
--
Josh Tynjala
Bowler Hat LLC <https://bowlerhat.dev>
On Fri, Jun 18, 2021 at 11:57 AM Yishay Weiss <yishayj...@hotmail.com>
wrote:
> Did you understand why the version was considered to be recursive? I
> assumed it should be picked up from these lines in the main pom.xml
>
> <version>0.9.8-SNAPSHOT</version>
> …
>
> <project.royale.compiler.version>${project.version}</project.royale.compiler.version>
>
> Also, what was your original motivation for reverting
> b8d136a1644f5de2a14b9185c48637e1089313b4 ?
>
> FWIW, I’m fine with giving concrete version names, just curious.
>
>
> From: Josh Tynjala<mailto:joshtynj...@bowlerhat.dev>
> Sent: Friday, June 18, 2021 1:09 AM
> To: Apache Royale Development<mailto:dev@royale.apache.org>
> Subject: Re: Hudson Nightly build as-js failure
>
> I was able to reproduce the issue locally, and the changes in
> commit 87a9c7281458789e511a1178331b24dd71a0dcdd fixed it for me.
>
> Looking back at the history a bit, it seems that I essentially reverted
> commit b8d136a1644f5de2a14b9185c48637e1089313b4. That commit was perfectly
> valid at the time, but some recent tweaks to the Maven pom make it a less
> ideal solution now.
>
> --
> Josh Tynjala
> Bowler Hat LLC <https://bowlerhat.dev>
>
>
> On Thu, Jun 17, 2021 at 9:31 AM Josh Tynjala <joshtynj...@bowlerhat.dev>
> wrote:
>
> > You're welcome to revert my commits, if you think it will help. However,
> > I'm not sure that's the correct solution.
> >
> > We're in a weird state right now because royale-compiler and
> > royale-typedefs are both on 0.9.9-SNAPSHOT, and royale-asjs was just
> > reverted to 0.9.8-SNAPSHOT. With that in mind, it probably shouldn't be
> > surprising if builds are failing.
> >
> > --
> > Josh Tynjala
> > Bowler Hat LLC <https://bowlerhat.dev>
> >
> >
> > On Thu, Jun 17, 2021 at 5:41 AM Piotr Zarzycki <
> piotrzarzyck...@gmail.com>
> > wrote:
> >
> >> When I run Maven build locally I got errors related to this Josh's
> commit
> >> -> f752231aeb22a82d740dbb410d1d4deac84eb6e5
> >>
> >> I changed both lines from
> >> <project.royale.compiler.version>{project.version}
> >> </project.royale.compiler.version> to <project.royale.compiler.version
> >> >0.9.8-SNAPSHOT</project.royale.compiler.version> - and build started to
> >> running. It's in progress - I will let you know what's the results.
> >>
> >> czw., 17 cze 2021 o 14:18 Piotr Zarzycki <piotrzarzyck...@gmail.com>
> >> napisał(a):
> >>
> >> > Maybe revert issue and in one of the pom there are Git leftovers.
> >> >
> >> > czw., 17 cze 2021 o 13:45 Yishay Weiss <yishayj...@hotmail.com>
> >> > napisał(a):
> >> >
> >> >> We are getting this in log [1]
> >> >>
> >> >>
> >> >> [EnvInject] - Variables injected successfully.
> >> >>
> >> >> No emails were triggered.
> >> >>
> >> >> Parsing POMs
> >> >>
> >> >> SSH: Connecting from host [jenkins-win-he-de-6]
> >> >>
> >> >> SSH: Connecting with configuration [Nightlies] ...
> >> >>
> >> >> SSH: Disconnecting configuration [Nightlies] ...
> >> >>
> >> >> SSH: Transferred 2 file(s)
> >> >>
> >> >> Build step 'Send files or execute commands over SSH' changed build
> >> result
> >> >> to SUCCESS
> >> >>
> >> >> ERROR: Failed to parse POMs
> >> >>
> >> >> hudson.remoting.ProxyException:
> >> >> hudson.maven.MavenModuleSetBuild$MavenExecutionException
> >> >>
> >> >> Whereas successful build’s corresponding log [2] segment is
> >> >>
> >> >> [EnvInject] - Variables injected successfully.
> >> >> No emails were triggered.
> >> >> Parsing POMs
> >> >> Modules changed, recalculating dependency graph
> >> >>
> >> >> Any ideas?
> >> >>
> >> >>
> >> >>
> >> >> [1] Royale » Royale-asjs #1301 Console [Jenkins] (apache.org)<
> >> >>
> >>
> https://ci-builds.apache.org/job/Royale/job/Royale-asjs/lastFailedBuild/console
> >> >> >
> >> >> [2] Royale » Royale-asjs #1290 Console [Jenkins] (apache.org)<
> >> >>
> >>
> https://ci-builds.apache.org/job/Royale/job/Royale-asjs/1290/consoleFull>
> >> >>
> >> >
> >> >
> >> > --
> >> >
> >> > Piotr Zarzycki
> >> >
> >>
> >>
> >> --
> >>
> >> Piotr Zarzycki
> >>
> >
>
>
