[hibernate-dev] Strange issues with -snapshot dependencies

[Sanne Grinovero]

Hi Scott, thanks for your suggestions.
It turns out it wasn't a Maven or caching problem: everything made
more sense in the morning. My issue was Arquillian not running the
tests on the correct deployments, which explains why the log messages
weren't being triggered as the configuration wouldn't match what I had
been debugging.

All good now..
Thanks,
Sanne

On 10 July 2015 at 00:31, Scott Marlow <smarlow at redhat.com> wrote:
>
>> I'm starting to wonder if actually sometimes Maven runs my integration
>> tests using a previous SNAPSHOT build, which would be missing both the
>> logs and the fix, and in those cases it could either fail or not (as
>> it wasn't always reproduced).
>>
>> Anyone else ever seen such an issue?
>>
>
> I worry about this problem happening but haven't experienced it.  If your
> building the SNAPSHOT build locally that you want to use, would it make
> sense to give it a unique name that wouldn't be found in the remote maven
> repository?  Something like:
>
> hibernateTargetVersion = '5.0.1-SANNE-SNAPSHOT'
>
> That might be a pita but at least you would have more confidence that its
> right.
>
> Another approach would be to disassemble (jad or procyon-decompiler) one of
> the SNAPSHOT jar classes to see if your code changes are in it.