#jogamp @ irc.freenode.net - 20140512 02:59:13 (UTC)

20140512 02:59:14 -NickServ- This nickname is registered. Please choose a different nickname, or identify via /msg NickServ identify <password>.
20140512 02:59:19 * jogamp (~jogamp@anon) has joined #jogamp
20140512 02:59:19 * Topic is 'http://jogamp.org | Hacking 3D Graphics, Multimedia and Processing across Devices'
20140512 02:59:19 * Set by rmk0 on 20130116 23:58:04
20140512 02:59:19 -NickServ- You are now identified for jogamp.
20140512 03:14:04 <sgothel> http://jogamp.org/deployment/archive/master/gluegen_786-joal_532-jogl_1249-jocl_959-signed/
20140512 03:14:07 <sgothel> http://jogamp.org/deployment/archive/master/gluegen_786-joal_532-jogl_1249-jocl_959/
20140512 03:14:26 <sgothel> maven (jogamp) 2.2.0-rc-20140409
20140512 03:17:31 * guardianp (~gp@anon) Quit (Remote host closed the connection)
20140512 03:40:30 <sgothel> maven (jogamp) 2.2.0-rc-20140512
20140512 03:40:52 <sgothel> @Mark: There is a bug w/ the deployment scripts .. I have overwritten 2.1.5
20140512 03:41:42 <sgothel> (due to an non clean 'output' folder .. containing 2.1.5 .. and 2.2.0-rc-.... - it used pom files for 2.1.5 but jar files of 2.2.0-rc.. oops)
20140512 04:12:19 <sgothel> OT: Kudos to 'Novorossiya in progress' .. (despite the bloodshed happened to the civilians) - Ridiculous German gov. is already concerned about Bulgaria .. oh well, looks like the IMF threat to reduce the loan wasn't that effective. A glimpse of democracy ? I wish them peace and no more killing. Actually same to the 'western' Ukraine .. but I guess they have to deal with an even greater corruption, where 'lines' are quite fuzzy. Good night :)
20140512 04:56:53 <[Mike]> Looking at https://jogamp.org/bugzilla/show_bug.cgi?id=823 , In AWTTextureData there are multiple cases to handle different formats. In these cases some set rowLength and then call setupLazyCustomConversion(). The setupLazyCustomConversion method sets the rowLength to the width. Is this correct functionality? See AWTTextureData:211-214 for example.
20140512 05:01:08 <[Mike]> I'm thinking about if we check (BufferedImage.getRaster().getParent() != null) then we call setupLazyCustomConversion(). Or I refactor the class :\
20140512 05:03:43 * jvanek (jvanek@anon) has joined #jogamp
20140512 06:19:54 * monsieur_max (~maxime@anon) has joined #jogamp
20140512 07:35:11 * hija (~hija@anon) has joined #jogamp
20140512 07:42:56 * guardianp (~gp@anon) has joined #jogamp
20140512 07:44:58 * [Mike] (~Mike]@anon) Quit ()
20140512 07:55:56 * hija (~hija@anon) Quit (Quit: hija)
20140512 07:58:01 * guardianp (~gp@anon) Quit (Remote host closed the connection)
20140512 08:30:20 * guardianp (~gp@anon) has joined #jogamp
20140512 08:40:50 <rmk0_> sgothel: checking now
20140512 08:40:55 * rmk0_ is now known as rmk0
20140512 08:41:24 * rmk0 is now known as Guest52983
20140512 08:41:30 <Guest52983> damnit
20140512 08:41:51 * Guest52983 is now known as rmk0_
20140512 08:43:08 <rmk0_> sgothel: ah, do you mean you did: ./make.sh 2.2.0 && ./make.sh 2.1.0 ?
20140512 08:43:35 <rmk0_> and that the poms for the 2.1.0 versions turned out to be those from 2.2.0?
20140512 08:45:54 * bbbruce (~bx@anon) Quit (Quit: bbbruce)
20140512 08:56:45 * hija (~hija@anon) has joined #jogamp
20140512 09:04:59 * bbbruce (~bx@anon) has joined #jogamp
20140512 09:24:22 * guardianp is now known as budder
20140512 09:24:47 * budder is now known as budder69
20140512 09:58:27 <xranby> (12:51:46) sgothel: http://www.cafc.uscourts.gov/images/stories/opinions-orders/13-1021.Opinion.5-7-2014.1.PDF (Oracle's won the appeal re. API 'copyrightable') <--- gosh..!
20140512 09:59:42 <xranby> some of the classpath developers think about reviving the icedrobot project.. to run android on top of openjdk
20140512 13:50:26 * bbbruce (~bx@anon) Quit (Read error: Connection reset by peer)
20140512 13:51:55 * bbbruce (~bx@anon) has joined #jogamp
20140512 13:52:02 * bbbruce (~bx@anon) Quit (Client Quit)
20140512 13:52:10 * bbbruce (~bx@anon) has joined #jogamp
20140512 15:17:38 <zubzub> I don't get Oracle
20140512 15:18:06 <zubzub> if they really press forward with their Java "strategy", they'll kill it
20140512 15:44:32 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20140512 15:44:46 * budder69 (~gp@anon) has left #jogamp
20140512 16:29:44 * monsieur_max1 (~maxime@anon) has joined #jogamp
20140512 16:43:28 * hija (~hija@anon) Quit (Quit: hija)
20140512 16:44:56 * hija (~hija@anon) has joined #jogamp
20140512 17:25:42 <sgothel> @Mark: ./make.sh 2.2.0 && ./make.sh 2.2.0
20140512 17:26:09 <rmk0_> sgothel: from an empty output directory?
20140512 17:26:19 <sgothel> poms for the 2.2.0 versions turned out to be those from 2.1.0 + jars from 2.2.0
20140512 17:26:36 <sgothel> output holding 2.1.0, 2.2.0 .. and others
20140512 17:26:41 <sgothel> when I cleaned it up - it works
20140512 17:26:44 <rmk0_> right, can reproduce it in that case
20140512 17:26:55 <rmk0_> it's written under the assumption that the output directory will be cleaned each time
20140512 17:27:04 <sgothel> so either ./make.sh cleans output .. or a better fix :)
20140512 17:27:10 <sgothel> ah ..
20140512 17:27:10 <rmk0_> right
20140512 17:27:32 <sgothel> so 2.1.5 contains some 2.2.0-rc now .. (jogamp maven)
20140512 17:27:34 <rmk0_> i can make ./make.sh call ./clean first
20140512 17:27:55 <sgothel> ./clean also does a 'rm -rf release' what is that ?
20140512 17:28:14 <rmk0_> rm -rfv output
20140512 17:28:14 <rmk0_> rm -rfv repository
20140512 17:28:24 <sgothel> yes repo.. whats that ?
20140512 17:28:30 <rmk0_> hehe, didn't realize i'd left that in there...
20140512 17:28:38 <rmk0_> was deploying to a local directory to test
20140512 17:29:59 <rmk0_> ok, pushed a couple of patches
20140512 17:30:06 <sgothel> thx!
20140512 17:30:14 <rmk0_> 30e387d9c6ab34fed8132437321ea31ed0b6a62a: Always clean before building
20140512 17:30:26 <rmk0_> e92feede1441fba02a9b5a178a2d4e03d7a77752: Remove removal of local repository!
20140512 17:45:40 * [Mike] (~Mike]@anon) has joined #jogamp
20140512 17:47:04 <sgothel> @Mike .. The AWT TextureIO tools .. well, I have to admit that I have little interest in those, since I favor the AWT agnostic code path
20140512 17:48:12 <sgothel> I mentioned an alternative .. (forgot) to extract the subImage .. which would need some sort of refactoring
20140512 17:48:22 <sgothel> if you do have an interest - pls go ahead!
20140512 17:51:06 <sgothel> to deal w/ generic 'pixels', we have GLPixelBuffer and the AWTGLPixelBuffer
20140512 17:51:50 <sgothel> the latter can be used to extract a BufferedImage .. and is used e.g. in GLJPanel and in our AWT printing interface
20140512 17:52:18 <sgothel> maybe .. a refactoring of AWT rel. texture utils can use those, since those classes are 'more to the point'
20140512 19:19:44 * hija (~hija@anon) Quit (Quit: hija)
20140512 20:25:47 * hija (~hija@anon) has joined #jogamp
20140512 20:44:55 <sgothel> dinner .. will give our API versioning validation a shot afterwards .. before cont. hacking
20140512 20:56:38 * monsieur_max1 (~maxime@anon) Quit (Quit: Leaving.)
20140512 21:19:23 * hija (~hija@anon) Quit (Quit: hija)
20140512 21:23:57 * hija (~hija@anon) has joined #jogamp
20140512 21:40:38 * hija (~hija@anon) Quit (Quit: hija)
20140512 21:42:29 * hija (~hija@anon) has joined #jogamp
20140512 21:55:16 * hija (~hija@anon) Quit (Quit: hija)
20140512 21:56:33 * hija (~hija@anon) has joined #jogamp
20140512 23:41:19 * hija (~hija@anon) Quit (Quit: hija)
20140512 23:57:27 * hija (~hija@anon) has joined #jogamp
20140513 01:51:04 * guardianp (~gp@anon) has joined #jogamp
20140513 05:18:14 <sgothel> https://jogamp.org/wiki/index.php/Jogamp_Versioning_and_Releases#Semantic_Version_Numbers
20140513 05:20:52 <sgothel> @Mark: Seems to work - excellent ..
20140513 05:22:00 <sgothel> I will add them to Gluegen .. and add unit tests right into all modules validating compatibility with a good known version.
20140513 05:47:04 * monsieur_max1 (~maxime@anon) has joined #jogamp
20140513 05:50:19 * monsieur_max1 (~maxime@anon) Quit (Client Quit)
20140513 05:59:43 * [Mike] (~Mike]@anon) Quit ()
20140513 07:18:28 <sgothel> @Mark and all: http://jogamp.org/git/?p=gluegen.git;a=commit;h=c06288d2a12586ab8df3715cf130549fdd7499fb
20140513 07:18:37 <sgothel> jogl .. is coming up ..
20140513 07:51:03 <sgothel> java.lang.AssertionError: Current version 2.1.5 (2.1.5) is not BACKWARD_COMPATIBLE_USER to previous version 2.1.4 (2.1.4), but NON_BACKWARD_COMPATIBLE
20140513 07:51:31 <sgothel> ^^ would have detected the JOGL versioning error 2.1.4 -> 2.1.5 ! good !
20140513 08:37:26 * hija (~hija@anon) Quit (Quit: hija)
20140513 09:17:21 * hija (~hija@anon) has joined #jogamp
20140513 10:34:58 <rmk0_> sgothel: nice!
20140513 10:36:22 <rmk0_> how does that library cope with the, er, incompatible notion of version numbers in jogl?
20140513 10:36:32 <rmk0_> the minor == incompatible change part
20140513 10:36:35 <rmk0_> is it patched?
20140513 10:55:33 <zubzub> the problem with deviating from the real semver is that it leaves it open for interpretation
20140513 10:55:47 <zubzub> what is a minor incompatible change? what is major one?
20140513 10:56:24 <rmk0_> zubzub: i've made this argument already
20140513 10:57:06 <zubzub> then it must be a valid argument! ;)
20140513 11:14:53 <sgothel> ^^ see latest wiki entry linked above
20140513 11:15:21 <sgothel> so unit test expect a 'compat' type ..
20140513 11:16:05 <sgothel> hence the version NUMBER semantics is implicit by our test itself .. well
20140513 11:16:52 <sgothel> while writing the wiki .. IMHO the most important semantic diff is the major/minor incompat. API change .. like easy/safe .. to risky (JOGL1 -> JOGL2)
20140513 11:17:28 <sgothel> nothing to really loose here .. since pure bugfix patches are rare right now
20140513 11:17:43 * zzuegg (~zzuegg@anon) Quit (Read error: Connection reset by peer)
20140513 11:17:57 <sgothel> since the test is performed all the time now .. we have the detailed data available
20140513 11:19:51 * zzuegg (~zzuegg@anon) has joined #jogamp
20140513 11:37:43 <sgothel> hmm .. not so sure whether the result doesn't produce false positives (incompatible)
20140513 11:38:29 <sgothel> example: GLCanvas override Component setVisible() method
20140513 11:38:32 <sgothel> v1
20140513 11:38:38 <sgothel> then remove it -> v2
20140513 11:39:15 <sgothel> semver: incompat
20140513 11:39:46 <sgothel> in this regard: compat has been broken a few times before ..
20140513 11:40:10 <sgothel> Q: Shall we enforce public-only or also protected (class, method, field) ?
20140513 11:40:33 <sgothel> the latter would be req. if user overrides .. so I guess the latter -> SimpleDiffCriteria
20140513 11:43:32 <rmk0_> are the protected parts part of the "public" API? as in, is a user required to use them in order to use the API?
20140513 11:44:17 <zubzub> if the user *can* use them then it's user api
20140513 11:44:21 <zubzub> aka public api
20140513 11:46:28 <rmk0_> well, actually, the semver described by http://semver.org requires the public parts to be documented. as in, there has to be some unambiguous statement that says "this part of the API is public, you are allowed to depend on it not changing"
20140513 11:47:02 <rmk0_> java's access control is weak... there are some things that have to be public in the java sense that aren't actually supposed to be part of an API that a user can depend on
20140513 11:47:37 <rmk0_> was considering putting together a package of annotations for documenting "really public" parts
20140513 11:48:00 <rmk0_> think people often handle this by having "api" and "implementation" jars
20140513 11:48:32 <rmk0_> so you explicitly depend on the "api" jar, and pull in the "implementation" when a program is packaged
20140513 11:49:20 <sgothel> we have our jogamp.* package: all private impl. .. otherwise we depend on the access qualif.
20140513 11:49:35 <rmk0_> right
20140513 11:50:23 * rmk0_ is now known as rmk0
20140513 11:50:28 * rmk0 (~rmk0@anon) Quit (Changing host)
20140513 11:50:28 * rmk0 (~rmk0@anon) has joined #jogamp
20140513 14:23:40 <sgothel> https://github.com/sgothel/semantic-versioning/commits/jogamp .. oh dear :)
20140513 14:25:45 <rmk0> wonder if he applies semantic versioning to his own library
20140513 14:25:55 <rmk0> madness lies in that direction
20140513 14:26:36 <sgothel> hours of playing w/ the regexp 'creator' .. then I dropped it .. passing regexp directly .. duh
20140513 14:26:51 <sgothel> and the '.' '/' in class names .. interesting
20140513 14:29:05 <sgothel> however - it's a great start .. good tool
20140513 14:29:21 <sgothel> just need to understand and come clear regarding the .. err .. semantics :)
20140513 14:33:21 * guardianp (~gp@anon) Quit (Read error: Connection reset by peer)
20140513 14:47:38 <sgothel> http://jogamp.org/git/?p=jogl.git;a=commit;h=8acae6df2ca6884f65ad4b32ed2d35d6a1c46f87
20140513 14:47:41 <sgothel> jogl ..
20140513 14:48:14 <sgothel> TestVersionSemanticsNOUI: Added (disabled) test to see whats happening w/ diff versions .. (2.1.2 - 2.1.3, .. etc)
20140513 14:48:23 <sgothel> so far .. they are all incompatible !
20140513 14:48:29 <rmk0> hehe
20140513 14:49:07 <sgothel> the case above ..
20140513 14:49:11 <sgothel> A ext B
20140513 14:49:13 <sgothel> B.a()
20140513 14:49:16 <sgothel> A.a()
20140513 14:49:20 <sgothel> ^^ v1
20140513 14:49:27 <sgothel> remove A.a()
20140513 14:49:31 <sgothel> ^^ v2
20140513 14:49:34 <sgothel> incompatible ?
20140513 14:50:09 <sgothel> maybe .. if a() is referenced via type A directly ? . have to read the java compat thingy
20140513 14:50:12 <sgothel> (again)
20140513 14:50:50 <sgothel> the semver tool makes a very simply diagnostic .. i.e. if something has been removed -> incompatible
20140513 14:50:58 <rmk0> don't think it's possible to say in general
20140513 14:51:01 <sgothel> guess there is some work to do ..
20140513 14:51:14 <zubzub> if you remove something it's incompatible
20140513 14:51:21 <zubzub> according to the strict semver rules
20140513 14:51:32 <zubzub> if you add something it's compatible
20140513 14:51:46 <zubzub> logical no?
20140513 14:51:54 <sgothel> zubzub: you are the semver author ? or you use it alot ?
20140513 14:52:03 <zubzub> I have to use it all the time in osgi
20140513 14:52:12 <zubzub> and since my job involves a lot of osgi :)
20140513 14:52:29 <rmk0> was thinking in terms of API compatibility in the java sense... if a class overrides a method, and you then remove the override
20140513 14:52:38 <sgothel> for overriden stuff .. w/ virt-func table .. not so clear
20140513 14:52:51 <zubzub> ah ic
20140513 14:53:04 <sgothel> also signatures must be checked .. 'change' case
20140513 14:53:15 <rmk0> a strict interpretation of the semver rules says it's incompatible, but the result may still be type and semantically compatible
20140513 14:53:25 <zubzub> I guess the semver lib doesn't check the accessiblity of B
20140513 14:53:33 <zubzub> idd
20140513 14:54:08 <sgothel> I made a pullreq for my changes .. will see
20140513 14:54:12 <zubzub> yeah I guess the lib is in the wrong there
20140513 14:54:48 <sgothel> maybe we can refine the compat decision ..
20140513 14:55:05 <zubzub> I'm very interested how jogamp can make that lib work as we interested in using it here as well :)
20140513 14:55:11 <sgothel> I like the bytecode approach .. so I guess the basics are great
20140513 14:55:21 <sgothel> it's in now
20140513 14:55:49 <sgothel> (using my patched version)
20140513 14:59:37 <sgothel> will refine the stats as produced in VersionSemanticsUtil later on (separate add/remove/change) .. but at least we have a report now
20140513 15:00:09 <sgothel> laters
20140513 15:01:28 <zubzub> bye!
20140513 15:19:31 * guardianp (~gp@anon) has joined #jogamp
20140513 15:33:41 * guardianp (~gp@anon) Quit (Quit: liteIRC for Android)
20140513 15:58:46 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140513 16:37:57 * monsieur_max (~maxime@anon) has joined #jogamp
20140513 17:35:21 * [Mike] (~Mike]@anon) has joined #jogamp
20140513 17:46:13 * hija (~hija@anon) Quit (Quit: hija)
20140513 19:05:02 * hija (~hija@anon) has joined #jogamp
20140513 19:24:48 * hija (~hija@anon) Quit (Quit: hija)
20140513 20:06:43 * hija (~hija@anon) has joined #jogamp
20140513 21:29:36 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140513 21:50:12 * hija (~hija@anon) Quit (Quit: hija)
20140513 21:51:43 * kermyt (~kermyt@anon) Quit (Ping timeout: 276 seconds)
20140513 21:56:17 * kermyt (~kermyt@anon) has joined #jogamp
20140513 22:07:50 * hija (~hija@anon) has joined #jogamp
20140513 22:40:18 <sgothel> looks like all semver changes are welcome .. goody ..
20140513 23:35:58 <sgothel> https://jogamp.org/wiki/index.php/Jogamp_Versioning_and_Releases#Semantic_Version_Numbers
20140513 23:51:52 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=1008
20140513 23:52:39 <sgothel> @Mark: THANK you for pushing this ! I do agree - it is important and a solution will give us better good night sleep :)
20140514 00:20:20 * guardianp (~gp@anon) has joined #jogamp
20140514 00:52:53 * guardianp (~gp@anon) Quit (Ping timeout: 245 seconds)
20140514 01:52:52 * guardianp (~gp@anon) has joined #jogamp
20140514 02:04:35 * guardianp (~gp@anon) Quit (Ping timeout: 252 seconds)
20140514 03:48:49 * guardianp (~gp@anon) has joined #jogamp
20140514 05:05:14 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20140514050514.html