Bug 886 - jogl-all-natives-linux-amd64.jar content has not changed in new release
Summary: jogl-all-natives-linux-amd64.jar content has not changed in new release
Status: RESOLVED INVALID
Alias: None
Product: Jogl
Classification: JogAmp
Component: core (show other bugs)
Version: 2
Hardware: pc_x86_64 linux
: --- critical
Assignee: Sven Gothel
URL:
Depends on:
Blocks:
 
Reported: 2013-11-02 16:31 CET by Marco Möller
Modified: 2013-11-05 13:03 CET (History)
2 users (show)

See Also:
Type: ---
SCM Refs:
Workaround: ---


Attachments
diff 2.0.2 2.1.0 2.1.1 2.1.2 of jogl-all-natives-linux-amd64.jar (4.77 KB, text/plain)
2013-11-02 22:31 CET, Sven Gothel
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Marco Möller 2013-11-02 16:31:58 CET
the content of the file 	jogl-all-natives-linux-amd64.jar
has not changed in 2.1.2 compared to 2.1.0 on binary diff level, for other OS the files have changed... maybe a bug in your build?

Best regards

-marco
Comment 1 Sven Gothel 2013-11-02 22:31:41 CET
Created attachment 534 [details]
diff 2.0.2 2.1.0 2.1.1 2.1.2 of jogl-all-natives-linux-amd64.jar
Comment 2 Sven Gothel 2013-11-02 22:34:26 CET
The native JAR manifest file states the proper build version.
Hence I have to assume that no binary diff of the native libs occurred between:
  2.1.0 -> 2.1.1 -> 2.1.2

however, there is a binary diff of the native libs:
  2.0.2 -> 2.1.0

where we actually changes things.

So either you have more indications that something is missing (reopen in this case)
or we have to come to the conclusion that the compiler created 1:1 binaries
from the same source code, incl. GlueGen generated native code.
(This makes sense ofc.)
Comment 3 Marco Möller 2013-11-05 11:50:26 CET
Hi,

I agree that the manifest is different, but are their NO changes in linux code for jogl between 2.1.0 and 2.1.2 ? On win 32/64 and mac I have seen changes in the native libs....

maybe something is wrong with the build process?

THX
-marco
Comment 4 Sven Gothel 2013-11-05 12:26:21 CET
(In reply to comment #3)
> Hi,
> 
> I agree that the manifest is different, but are their NO changes in linux
> code for jogl between 2.1.0 and 2.1.2 ? 
Right!

> On win 32/64 and mac I have seen
> changes in the native libs....

We build macos binaries w/ new xcode clang now.

Win compiler has changed as well.

> 
> maybe something is wrong with the build process?

Please prove it (see my prev. comment), I couldn't find anything wrong.

In general, as long as the compiler nor linker add a timestamp
or other time related / build number related artifacts to the binaries
it _should_ be assumed that the resulting binaries are equal.

Further more, we clean the workspace before build!

If you can show us that indeed an old binary is being used .. great,
then pls reopen.

~Sven

> 
> THX
> -marco
Comment 5 Marco Möller 2013-11-05 12:33:43 CET
perfect!

I was only suspicious that this was the only binary platform which was different...

anyway: its hard for me to prove to you where your (for me blackbox) build process is wrong. I only can give you some hints what I have found. If you dont like to get such feedback.. just tell me.

Anyway: Thanks for your great tool!
-Marco
Comment 6 Sven Gothel 2013-11-05 13:03:23 CET
(In reply to comment #5)
> perfect!
> 
> I was only suspicious that this was the only binary platform which was
> different...
> 
> anyway: its hard for me to prove to you where your (for me blackbox) build
> process is wrong. I only can give you some hints what I have found. 

No problem - the build process is actually sort of public, i.e. jogamp-scripting
contains the aggregation scripts .. i.e. how the 7z files are build
based on the test nodes results.
Ofc .. not all jenkins settings are public, but most of our setup is.

> If you dont like to get such feedback.. just tell me.

No - sorry for sounding harsh, this was not personal at all.
Input and indications of suspicion is _always_ _welcome_ !
I was just missing the 'proof' thats all.
So please don't hold back next time .. I will hear you 'crying wolf' :)

> 
> Anyway: Thanks for your great tool!
> -Marco

Thank you, you are very welcome.

~Sven