Thanks Taher,
I can confirm it worked twice on BuildBot, so I guess it's OK :)
https://ci.apache.org/builders/ofbiz-trunk/builds/1210
I have still the tiny testFopMacroLibrary issue on Win7 but I don't know yet if
it's due to my Windows config or general:
Response contentType isn't good : text/html;charset=UTF-8
expected:<[application/pdf];charset=UTF-8> but was:<[text/html];charset=UTF-8>
I did not dig in it for now
Jacques
Le 08/08/2016 à 13:25, Taher Alkhateeb a écrit :
I just bumped shiro to 1.3.0 from 1.2.5, all tests pass and everything runs
on my computer. I hope this resolves it. Waiting to see.
On Mon, Aug 8, 2016 at 12:44 PM, Jacques Le Roux <
jacques.le.r...@les7arts.com> wrote:
Hi Taher,
I just tried, same issue https://ci.apache.org/builders
/ofbiz-trunk/builds/1208/steps/shell/logs/stdio
I'll see later now (being off for 4 days)
Jacques
Le 08/08/2016 à 10:11, Taher Alkhateeb a écrit :
Hi Jacques,
I think they are different. Your code seems to apply encoding only on the
javaCompile task while my code applies a global jvm property that spans
the
whole project.
Anyway if that does not work then that leaves us with either replacing the
dependency or fixing the jar classpath construction logic. Let's see.
Taher Alkhateeb
On Sunday, 7 August 2016, Jacques Le Roux <jacques.le.r...@les7arts.com>
wrote:
I already did compileJava.options.encoding = 'UTF-8'
at http://svn.apache.org/viewvc?view=revision&revision=1755451
Would be below different?
BTW what to you mean with VM it seems we have no issues with the trunk
demo on OFBIZ-VM, but it's cached there (we use the wrapper)
By VM do you mean BuildBot? If yes, then yes try it
Jacques
Le 07/08/2016 à 21:58, Taher Alkhateeb a écrit :
I suspect the dependencies refactoring I wrote might have exposed
libraries
with this encoding issue.
Jacques is it possible to experiment on the VM? We can try the below
workaround in build.gradle and see if that resolves issues for now
System.setProperty("file.encoding", "UTF-8")
On Aug 7, 2016 8:42 PM, "Jacques Le Roux" <jacques.le.r...@les7arts.com
wrote:
In case it helps I have set --stacktrace in BuildBot
See https://ci.apache.org/builders/ofbiz-trunk/builds/1207/
steps/shell/logs/stdio
Jacques
Le 07/08/2016 à 20:45, Jacques Le Roux a écrit :
As it was random, I double checked forcing a BuildBot build and it did
not work, still the same random issue https://ci.apache.org/builders
/ofbiz-trunk/builds/1206
I read at http://www.odi.ch/weblog/posting.php?posting=585 (and
somewhere else on stackoverflow I don't clearly remember) that Java
seems
to support both UTF8 and UTF-8
But it seems to depend on libs (facepalm)
https://docs.oracle.com/javase
/8/docs/technotes/guides/intl/encoding.doc.html
Even found something clearer: https://en.wikipedia.org/wiki/
UTF-8#Modified_UTF-8 did not digg further and I'm in vacation for 4
days
tomorrow morning :)
I though can't see why the same encoding name (UTF8) would work one
time
and not the other :-o
Jacques
Le 07/08/2016 à 20:25, Jacques Le Roux a écrit :
I think I explained this already several times. Buildbot starts anew
each time on one of the 3 slaves OFBiz uses. So AFAIK there is noway
to
cache things.
We could ask about other projects practices but I suppose I'd have
been
noticed when I asked for Gradle installation
https://issues.apache.org/jira/servicedesk/customer/portal/
1/INFRA-12235
Anyway I fixed it with r1755451
Jacques
Le 07/08/2016 à 11:58, Taher Alkhateeb a écrit :
Actually this is a good question Scott.
Jacques why aren't the libs cached? are we deleting ".gradle" in
buildbot
upon each new build?
It is hard to investigate second hand. Is there a way to check what
is
going on exactly on the VM? The inability to repeat on my computer
makes it
hard to pinpoint the problem.
On Sunday, 7 August 2016, Scott Gray <scott.g...@hotwaxsystems.com>
wrote:
Surely there's some solution for the jars not being cached? I wonder
how
the other projects deal with this
Regards
Scott
On 7/08/2016 03:56, "Taher Alkhateeb" <slidingfilame...@gmail.com
<javascript:;>> wrote:
Okay cool. I am trying to eliminate all causes. The crash is
happening
exactly when trying to construct the classpath for the constructed
jar.
This line of code in build.gradle is what's causing the crash:
osClassPath = configurations.runtime.files.collect { "$it"
}.join('
')
And it is complaining about UTF-8 encoding. Not sure exactly why
but
it
very much smells like a bug around encoding and string parsing.
Still
investigating.
On Aug 6, 2016 4:43 PM, "Jacques Le Roux" <
jacques.le.r...@les7arts.com
<javascript:;>>
wrote:
We never used the Gradle daemon on BuildBot as it's not
recommended
on
CI
servers. There is no cache on BuildBot, it always starts anew. We
always
use Gradle and not the wrapper there.
Also the project name was fixed at r1754623 so 35 commits before
https://ci.apache.org/builders/ofbiz-trunk?numbuilds=35 I don't
see
how
it could be related, but this is weird indeed
Unrelated, just FYI: we also no longer use the Gradle daemon with
trunk
demo as it seems it was sometimes an issue. Though actually the
real
issue
(DB related) we got there was using "shutdown" instead of now
"terminateOfbiz"which seems to have fixed the DB issue
Jacques
Le 06/08/2016 à 17:06, Taher Alkhateeb a écrit :
Hi Jacques,
Can you try to disable the gradle Daemon? I suspect this might be
a
cause
as per this discussion ->
http://stackoverflow.com/questions/21267234/show-utf-8-text-
properly-in-gradle
Taher Alkhateeb
On Sat, Aug 6, 2016 at 5:50 PM, Taher Alkhateeb <
slidingfilame...@gmail.com <javascript:;>>
wrote:
Haa, I just noticed something interesting in
https://ci.apache.org/
builders/ofbiz-trunk/builds/1204/steps/shell/logs/stdio
If you look at the logs, ignore shiro for a second, ant note
the
following:
Could not resolve all dependencies for configuration
':runtime'.
> Could not resolve org.apache.shiro:shiro-core:1.2.5.
Required by:
:ofbiz:unspecified
There is no project called :ofbiz:unspecified. Something is not
parsing
the project tree correctly and I'm not sure of the reason yet.
It
could be
that when we fixed the project name as "ofbiz" in settings.gradle
something
got affected in the cache.
I don't think this is a transient bug, I suspect we introduced
something
that is causing this weird behavior.
On Sat, Aug 6, 2016 at 5:23 PM, Jacques Le Roux <
jacques.le.r...@les7arts.com <javascript:;>> wrote:
OK, here we go again https://ci.apache.org/builders
/ofbiz-trunk/builds/1204/steps/shell/logs/stdio
Anyway "wait and see" will come to an end
Jacques
Le 06/08/2016 à 16:08, Jacques Le Roux a écrit :
OK the dependency on Shiro is fixed in jcenter :)
There is no issues that "wait and see" can't solve :D
Jacques
Le 06/08/2016 à 14:19, build...@apache.org <javascript:;> a
écrit
:
The Buildbot has detected a restored build on builder
ofbiz-trunk
while
building . Full details are available at:
https://ci.apache.org/builders/ofbiz-trunk/builds/1203
Buildbot URL: https://ci.apache.org/
Buildslave for this Build: silvanus_ubuntu
Build Reason: The AnyBranchScheduler scheduler named
'on-ofbiz-commit'
triggered this build
Build Source Stamp: [branch ofbiz/trunk] 1755393
Blamelist: arunpatidar
Build succeeded!
Sincerely,
-The Buildbot