summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Add com.jogamp.common.util.PrimitiveStack (FloatStack and IntegerStack), a ↵Sven Gothel2012-12-305-5/+551
| | | | | | | | | | | | simple primitive stack implementation. Currently only FILO put/get operations are implemented using either primitive arrays as I/O itself or <Type>Buffer. Unit tests are included.. Note: Only FloatStack is implemented in a manual, where others (IntegerStack) is derived (generated) from it. Same goes w/ unit tests.
* Fix Bug625: StructAccessor missing setShortsAt() and getShortsAt() methods ↵Sven Gothel2012-12-301-0/+13
| | | | | | | | | | for short[] For short[] Java code gets emitted for an StructAccessor object that uses: void setShortsAt(int i, short[] shorts) short[] getShortsAt(int i, short[] shorts) Problem was that StructAccessor.java had no such methods - added.
* Refine AWTEDTExecutor.invoke(..): Allow control whether a non AWT-EDT may ↵Sven Gothel2012-12-281-15/+36
| | | | | | | execute the runnable. For some 'rare' AWT/GL lifecycle actions, it is required to only run the command on the AWT-EDT, hence adding an argument determining the restriction.
* Complete commit 43163af2618a0aaa3cf41de8027ef402d7e89cc3 - ↵Sven Gothel2012-12-271-6/+6
| | | | jogamp-androidtasks.xml was using ANDROID_SDK_HOME, -> ANDROID_HOME
* Android build scripts: Fix env. name ANDROID_SDK_HOME -> ANDROID_HOMESven Gothel2012-12-274-24/+24
| | | | | | | | Intuitively I assumed ANDROID_SDK_HOME to be pointing to the SDK root dir, however this is not true: Semantics by Android tools are: ANDROID_SDK_HOME - Users ~/.android folder ANDROID_HOME - SDK root folder
* Android Launcher: Fix ClassLoaderUtil's JNI lib-path determination, use ↵Sven Gothel2012-12-271-9/+19
| | | | | | | | | | | | ApplicationInfo's nativeLibraryDir (API level 9). On Android > 4.0.3 (maybe even earlier), w/ a split filesystem (internal and SDCARD) the JNI libs maybe stored at a different location than it's data path. ApplicationInfo's nativeLibraryDir properly determines the JNI storage location, hence use it. Prev. code also derived JNI lib path by the launcher's ApplicationInfo's nativeLibraryDir, which might be different than the user package's nativeLibraryDir. This is especially true, since the launcher may not hold any native libraries.
* IntIntHashMap: Add privileged for 'getDeclaredMethod(..)'; Add StringBuilder ↵v2.0-rc11Sven Gothel2012-10-311-42/+75
| | | | toString(..); Add more generics coding.
* IntBitfield: Add API doc; Enhance put() method, return previous value to be ↵Sven Gothel2012-10-271-7/+27
| | | | used more versatile and write only if value changed.
* Adding com.jogamp.common.util.IntBitfield: Simple bitfield holder class ↵Sven Gothel2012-10-271-0/+71
| | | | | | | using an int[] storage. IntBitfield comes in handy to store bit states of a wide value range w/o being a memory hog an O(1) access, e.g. keyCode -> isPressed maps etc.
* Fix Bug 608: Push down TempJarCache.findLibrary(..) from ↵Sven Gothel2012-10-252-16/+15
| | | | | | | JNILibLoaderBase.loadLibraryInternal() to NativeLibrary.findLibrary(..) This allows using TempJarCache (if used/initialized) for native 'tool' libraries as well. This is the case of JOAL's attempt to load the provided 'libopenal.so'
* VersionNumber: API doc fixSven Gothel2012-10-241-1/+1
|
* VersionNumber: Add API doc for string parsingSven Gothel2012-10-241-0/+7
|
* VersionNumber: Remove printing Exception (was added for debugging purposes)Sven Gothel2012-10-241-1/+1
|
* Fix VersionNumber: Non digits cut off pattern was '\D.*' and cut off digits ↵Sven Gothel2012-10-243-44/+145
| | | | | | | | | | | | | | | | | | | in case leading non digits appear. '\D.*' matches a leading non-digit and then any character. First the string is tokenized by delim: "OpenGL ES GLSL ES 1.0.16" -> "OpenGL ES GLSL ES 1", "0", "16" Enhance pattern as follows and access collected group if matching: // group1: \D* == leading non digits, optional // group2: \d* == digits // group3: .* == any pending chars, optional final java.util.regex.Pattern nonDigitsCutOff = java.util.regex.Pattern.compile("(\\D*)(\\d*)(.*)"); Reorganized storage from atomic values to int[3] array, allowing simple pattern matching loop. Added unit test.
* Revert "Minor revert of clone replacement w/ copt-ctor of commit ↵Sven Gothel2012-10-183-7/+4
| | | | | | | | | | | | | | | d7fb6a7bcfbd2d5ac452abdcdd31fb4d71441c70" This reverts commit 73e8828566404e864170688dfb4fd530a83f8add. Convinced after discussing semantics w/ Harvey Harrison: "the copy-constructor is way for predictable (as in branch-predictable) and has better cache behavior, it can issue almost all the writes in parallel, and has no exception catching. So, the copy-constructor actually ends up being more efficient, and you get typechecking to boot."
* Fix Buffers.copy<Type>Buffer[asByteBuffer](..): Reset position of the passed ↵Sven Gothel2012-10-181-14/+15
| | | | | | | | | buffer (was missing). This is an API regression as introduced in commit 25cc744f6bd5ca97e0ae58fa7e1c35f7b0f3046d where JOGL's BufferUtil and GlueGen's were merged and API doc elaborated. The latter states the desired fact that the source passed buffer's position shall remain unchanged, however no code was added to take care of this detail.
* Minor revert of clone replacement w/ copt-ctor of commit ↵Sven Gothel2012-10-173-4/+7
| | | | | | d7fb6a7bcfbd2d5ac452abdcdd31fb4d71441c70 clone of ArrayList is more shallow than it's clone operation (?)
* Merge remote-tracking branch 'hharrison/RFC'Sven Gothel2012-10-177-21/+7
|\
| * gluegen: replace Thread name with name of lock file in ↵Harvey Harrison2012-10-161-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | SingletonInstanceFileLock error message The getName() call here is currently calling Thread.getName() inherited by the anonymous inner class. This is ambiguous, and likely meant to call the getName() method of SingletonInstanceFileLock, which would output the lockfile name. In any event, infoPrefix() is already adding the Thread name for us. Signed-off-by: Harvey Harrison <[email protected]>
| * gluegen: remove two fast-path tests that can never triggerHarvey Harrison2012-10-161-14/+0
| | | | | | | | | | | | | | | | ignores and unimplemented are both Set<Pattern>, they cannot contain a String, immediately after these 'fast paths' we loop over the patterns calling .matches() which makes sense, so remove the misleading tests. Signed-off-by: Harvey Harrison <[email protected]>
| * gluegen: generics annotations for various ArrayList/List useHarvey Harrison2012-10-155-6/+6
| | | | | | | | | | | | | | - Use the copy constructor rather than ArrayList.clone() - constrain listsEqual so the compiler will warn about comparing lists of different types Signed-off-by: Harvey Harrison <[email protected]>
* | Buffers: Add normalized put methods, i.e. incl. value range conversion; New ↵Sven Gothel2012-10-174-3/+408
|/ | | | util. ValueConv for primitive type value conversion
* ProcAddressEmitter: Remove 'guessing' argument names from types, since they ↵Sven Gothel2012-10-162-18/+2
| | | | are ambiguous - better use default generic ones to avoid confusion.
* Reducing Type's equalsSven Gothel2012-10-161-32/+0
|
* Intermediate revised Type.java edit before reducingSven Gothel2012-10-161-8/+34
|
* gluegen: conditional cleanup in Type.equals()Harvey Harrison2012-10-141-12/+7
| | | | | | | | - instanceof includes null checking - simplify comparison of name, either they are the same object (possibly null) or they compare equal as strings Signed-off-by: Harvey Harrison <[email protected]>
* gluegen: simplify expression comparing against Boolean.TRUEHarvey Harrison2012-10-131-7/+2
| | | | | | | | | As a last resort we can just return the boolean value directly here, also no need to try/catch as valueOf does not throw any exceptions. Pointed out by Findbugs. Signed-off-by: Harvey Harrison <[email protected]>
* gluegen: remove trailing whitespace from psuedo-C parserHarvey Harrison2012-10-131-5/+5
| | | | Signed-off-by: Harvey Harrison <[email protected]>
* gluegen: break apart a complex conditional in Type.equals() for readabilityHarvey Harrison2012-10-131-2/+15
| | | | | | - move the cheap integer compares earlier before the String comparisons Signed-off-by: Harvey Harrison <[email protected]>
* gluegen: fix equals comparison in cgrammer Type classHarvey Harrison2012-10-131-1/+1
| | | | | | | name was being compared to itself, rather than the name of the type we are comparing, looks like a simple typo. Signed-off-by: Harvey Harrison <[email protected]>
* gluegen: remove trailing whitespace from cgrammer Type classHarvey Harrison2012-10-131-10/+10
| | | | Signed-off-by: Harvey Harrison <[email protected]>
* gluegen: small cleanup and type annotation in GlueGenTaskHarvey Harrison2012-10-131-37/+23
| | | | | | | | - remove an unused import - annotate includesList with type information - small code cleanup now that the type information is present Signed-off-by: Harvey Harrison <[email protected]>
* gluegen: trim trailing whitespace from GlueGenTaskHarvey Harrison2012-10-131-64/+64
| | | | Signed-off-by: Harvey Harrison <[email protected]>
* Adding gcc linker cfg: '-static-libgcc' for all def. build platforms:Sven Gothel2012-10-135-0/+16
| | | | | | | | | | | | | | | | | | | | | | | | | - linker.cfg.linux - linker.cfg.linux.x86 - linker.cfg.linux.amd64 - linker.cfg.linux.armv6 - linker.cfg.solaris - linker.cfg.solaris.spacv9 - linker.cfg.solaris.amd64 - linker.cfg.macosx - linker.cfg.linux64.mingw64 - linker.cfg.linux64.mingw32 - linker.cfg.win32.mingw32 - linker.cfg.win32.mingw64 - android.armv6 - android.armv7 - linux.armv6 - linux.armv6hf These flags shall now go through autobuild and results will be validated, i.e.: - working - memory footprint
* Buffers.toString(): Add optional format string for single elementSven Gothel2012-10-101-8/+39
|
* Buffers: Add convenient "public static StringBuilder toString(StringBuilder ↵Sven Gothel2012-10-101-1/+63
| | | | | | sb, Buffer buffer)" Sometimes we need to see details of a Buffer w/ it's content, e.g. for debugging purposes.
* Fix NPE in Buffers.slice2Float(), regression of commit ↵Sven Gothel2012-10-091-3/+12
| | | | 86e8c3a8d9f430700e07c485127130da68618e9d
* Buffers.slice2Float(): Reset position and limit of host bufferSven Gothel2012-10-091-22/+25
|
* SingletonInstanceServerSocket: serverSocket.setReuseAddress(true); // reuse ↵Sven Gothel2012-10-042-2/+2
| | | | same port w/ subsequent instance, i.e. overcome TO state when JVM crashed
* TestSingletonServerSocket01: unlock singleton lockSven Gothel2012-10-041-0/+2
|
* Fix regression of 55b4552aef7882c358d545d020d6f12c958ed8ed, which impacts ↵Sven Gothel2012-10-041-1/+1
| | | | non native unix.
* Fix Bug 624 - Compile and generate android JAR files ↵Sven Gothel2012-10-042-6/+25
| | | | (gluegen-rt-android.jar) if android SDK JAR files are available.
* Refine API doc of Buffers, fix API doc Platform.NEWLINE.Sven Gothel2012-10-032-1/+9
|
* SingletonInstanceServerSocket: Add another (manual) unit tests for timeouts.Sven Gothel2012-09-212-5/+77
|
* SingletonInstanceServerSocket: Add unit tests; Create new server Thread @ ↵Sven Gothel2012-09-215-23/+186
| | | | start, otherwise we may collide w/ a failed start. Misc: Cleanup / reuse strings.
* SingletonInstanceServerSocket: Add kill @ JVM Shutdown _and_ if normal ↵Sven Gothel2012-09-213-9/+45
| | | | unlock fails ; Added unit tests.
* JNLP: Cache gluegen-rt-natives for linux-armv6 and linux-armv6hf.Xerxes Rånby2012-09-161-0/+8
| | | | | Due to the lack of OS Arch and ABI detection in JNLP launchers force us download both armv6 armel and armv6hf armhf ABI jars on ARM.
* AWTEDTExecutor: Add convenient "invoke(Object treeLock, boolean wait, ↵Sven Gothel2012-09-101-1/+25
| | | | Runnable r)" to be used directly
* windows scripts: bump to 6u35Sven Gothel2012-09-074-8/+8
|
* Platform: Slight reorder of initialization reducing one PrivilegedAction block.Sven Gothel2012-09-071-42/+41
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-02 14:34:30 +0000