Skip to main content
aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2016-03-25Bug 489669 - Some bundles decrease versions compared to Mars.2I20160329-0800Thomas Watson2-2/+2
Change-Id: I127134ef79bfee32c5370e91e46f4a4d15a04f05 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-03-24Bug 488935 - Avoid "workspace dirt" in the launcher bundlesMat Booth20-26/+23
Change the ant task to copy the artifacts in the maven "target" directory and include them in the bundle via tycho-packaging-plugin instead of build.properties. This is the idiomatic way to include dynamic or generated artifacts. Change-Id: Ib60544a804b2e1d4b93627a9360e7ee6ffdfe4f1 Signed-off-by: Mat Booth <mat.booth@redhat.com>
2016-03-24Bug 490374 - Enable framework tests for gerrit build jobThomas Watson6-52/+59
Change-Id: I4f575a3fbcffc919487f3fe77e9f8ccb29da8427 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-03-22Bug 490008 - Avoid storing complete jar Manifest objectY20160324-1000Alex Blewitt4-38/+222
Storing a whole Manifest for what is essentially minimal data is very space intensive. Not only does this result in an additional 1/2Mb of string data, the other data structures (particularly HashMaps) store dby Manifest objects ends up being able to reduce the start-up of a generic Eclipse SDK instance from a minimum of around 44Mb to 39Mb after performing repeated GC operations. The custom data structures are not necessarily as optimal as they could be (for example, repeated -Vendor strings aren't interned or uniquifed) but if nothing is present then a singleton NONE instance is used and no HashMap is instantiated. For an initial implementation it does well. Change-Id: I7e0a049a3cfeb42f49695576bb2feb71be8700ed Signed-off-by: Alex Blewitt <alex.blewitt@gmail.com> Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-03-22Bug 490008 - Add to test override of main attributes for package infoThomas Watson5-0/+75
Change-Id: I0d5dc773bd768525082045afaabfb6d631606238 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-03-21Bug 490008 - Add a test for package info in manifest main attributesI20160322-0800Thomas Watson2-10/+19
Change-Id: I5229b936ef6139b6150c1a55d8687302b4f40b81 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-03-21Bug 489675 - ContextFinder.getResources() searches resources inThomas Watson4-10/+30
different order with getResource Change-Id: I39d2d32b03ce92bc2879ec88ad60d59591291867 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-03-18Bug 489958: Java9 9-ea+108: Eclipse doesn't start because javax.annotation.* ↵Markus Keller1-6/+9
types are not on boot classpath java.version is now "9-ea". This commit adapts the workaround for bug 466683 to JEP 223. Change-Id: I96165e5935aa96cbc8ec6f894a476c58befbcf3f Signed-off-by: Markus Keller <markus_keller@ch.ibm.com>
2016-03-16Binaries v20160316-1045Y20160317-0800I20160317-0200I20160316-2000I20160316-1200Arun Thondapu19-19/+19
2016-03-16Bug 489642 - 32 bit launcher on linux is "ELF 64-bit LSB"Arun Thondapu1-0/+3
Change-Id: Id997f8c2da29da4bcb97419161f03022bf51e9d8
2016-03-12Binaries v20160312-1144I20160316-0800I20160315-2000I20160315-0800I20160314-2000I20160314-1400I20160314-0800I20160313-2000Arun Thondapu21-21/+21
2016-03-12Bug 399786 - launcher change for highDPI support - GTK onlySravan Kumar Lakkimsetti1-0/+3
Change-Id: I0ecfdfc3d3097308c39e7c5f006e1f5857752e06 Signed-off-by: Sravan Kumar Lakkimsetti <sravankumarl@in.ibm.com>
2016-03-12Bug 469766 - Support embedded JREs on MacOSChristian Georgi1-1/+1
Instead of assuming Oracle's JDK layout (Contents/Home/jre/bin), support pure JREs (Contents/Home/bin) as well. This is in line with the other platforms that assume the bin/ dir only, i.e. work with both JDK and JRE layouts. Bug 469766 Change-Id: I2be6bd9a28f06297ec2b89ff03dac75bd81fbae2 Signed-off-by: Christian Georgi <christian.georgi@sap.com>
2016-03-11Bug 469766 - Launcher-relative VM path to honor changed file layoutChristian Georgi1-2/+3
In the new Eclipse file layout on MacOS the launcher is now 2 levels below the app root dir (Eclipse.app/Contents/MacOS), while it was 3 levels before (eclipse/Eclipse.app/Contents/MacOS). This fix adjusts the relative path the launcher uses to find an embedded JVM in case no -vm launcher argument is given. Bug: 469766 Change-Id: I88f56fb03eaf99233504002b26d999d2dd5b1a15 Signed-off-by: Christian Georgi <christian.georgi@sap.com>
2016-03-09Bug 489312 - Prefer usage of Integer.valueof instead of new() for ↵Lars Vogel12-19/+19
org.eclipse.osgi Change-Id: I8b1062796d0d1f49146d1d21876de2aab303ae2a Signed-off-by: Lars Vogel <Lars.Vogel@vogella.com>
2016-03-08Bug 489233 - Avoid queue start-level operation if a bundle is notY20160310-0830Y20160310-0800Thomas Watson2-10/+43
persistently marked for activation Change-Id: I90372ece802642b0994ee091b73155e425e2bc3e Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-03-03[Bug 477787] Embedded Equinox in an entreprise application ear - ↵I20160308-0800John Ross3-1/+71
NullPointerException at start up In BundleInfo.Generation, check for null content and, if so, set lastModified to zero. In ModuleClassLoader, check for a null base file from BundleFile and, if so, create the ProtectionDomain with a null CodeSource. In both cases, the referenced variable will be null when the osgi.framework property was specified with an invalid value. Plus test. Signed-off-by: John Ross <jwross@us.ibm.com>
2016-03-03Bug 479180 - Delete launcher for unsupported Windows wpf platformY20160303-0800Mat Booth21-1233/+1
Change-Id: I91fc817f4954f7fa06b89ac466c170bbd44b7810 Signed-off-by: Mat Booth <mat.booth@redhat.com>
2016-03-02Bug 452486 - Remove Motif leftoversAlexander Kurtakov4-471/+0
* Cleanup launcher.releng. * Remove motif_root Change-Id: I9326d38a8dbcbf84ac9dbdd0369611d87a678fcc Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2016-03-02Bug 488846 - The profile.list entry for Java 9 is incorrectMat Booth1-2/+2
Change-Id: Ia9e20eebbc17797c8797e117d64b9f9ab717a4f4 Signed-off-by: Mat Booth <mat.booth@redhat.com>
2016-02-26Bug 488570 - Need to be more robust in the container persistence whenI20160301-1230I20160301-0800Thomas Watson3-21/+114
faced with invalid attribute types Change-Id: Ia8f3b7d6139d04c13e11917525517d55ffe4f0bb Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-25Bug 488501 - optimize storage and reading of strings and versionsThomas Watson3-150/+337
Change-Id: I4dcbe3fe97a85597812d6c507a38a535a754e77d Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-23Bug 483986 Footprint of module representation can be reducedEd Merks8-85/+184
Change-Id: I4c40ce395e3c2a3a00d79b3b18a03925839466f7 Signed-off-by: Ed Merks <ed.merks@gmail.com> Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-23Bug 487165 - Framework should read the parent configuration config.iniThomas Watson2-5/+49
file Change-Id: I4d0b8c2ea4e954b895c29d7a29fd9c3ad89cd3f7 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-19Bug 488127 - update xml-maven-plugin I20160223-0800David Williams1-1/+0
I have updated version in parent pom, and defined in pluginManagement, so please remove specific version number from equinox.executable. Change-Id: Ifd3136ee99cee1e3f0f60bb011a5a75f8b9e3aff Signed-off-by: David Williams <david_williams@us.ibm.com>
2016-02-19Bug 487842 - Unable to acquire state change lock for the module whenThomas Watson7-30/+78
starting a bundle Must increment inStart BEFORE acquiring the STARTED state the first time. There is still an unavoidable timing issue. The window is now as small as possible. A new flag is added to disable auto-starting newly resolved bundles that are marked for persistent start. Also fixed bug in how modules are restarted after a refresh. They must use their activation policy if it is set. Change-Id: I8175af9ed29e611659a8b8bc0b64b0a3874e68bf Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-18Bug 487842 - Unable to acquire state change lock for the module whenThomas Watson1-14/+11
starting a bundle - set the inStartResolve flag for the entire start operation while holding the STARTED state change lock. Otherwise there is a very small window where the check for inStartResolve can happen too early by the resolve operation resulting in another thread winning to start the module. Change-Id: I1e5a7f3216c8425a6dc7987a8d915a11886504db Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-18Bug 487954 - SignedBundleHook.getSignedContent() should report path David Williams1-3/+6
My guess is the "root" problem is in the JRE, but since they are deep in native methods, they may know the "file name" at the point the error occurs. But, we do, so I propose to catch the ZipException, and supplement it with the path information, and then re-throw it. I think that minimizes risk of effecting other code paths. Change-Id: Ieaf0072bb06f8be541a52cc24e35c38a82985e27 Signed-off-by: David Williams <david_williams@us.ibm.com> Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-17Bug 487842 - Unable to acquire state change lock for the module whenThomas Watson3-3/+19
starting a bundle - Give cause for BundleException to improve debugging. Change-Id: I31c91e258c841b01c3abdd1ff325201f74cc6c6e Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-15Bug 487842 - Unable to acquire state change lock for the module whenI20160216-1400I20160216-0800Thomas Watson4-35/+156
starting a bundle Change-Id: Ibc7c58aaf610cbd7b0c2aaef154559e2ed7134c4 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-11Bug 487696 - ModuleContainer should avoid dynamically resolving packagesThomas Watson2-4/+42
that are exported Change-Id: I451d572e659815ed3266261bb643d0b1e6f48e6f Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-09Bug 487537 - Resolving multiple framework extensions should add contentThomas Watson3-35/+39
of all before starting any extension activators Change-Id: Ie44f992bcbc055f09aeaa14616bba37127d9c9d7 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-09Bug 487524 - extension bundle content is not added on framework creationThomas Watson2-2/+7
if the system.bundle is unresolved Change-Id: I494a119ec37f0519086bf8d7db6d2ede073b2c99 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-08Bug 487466 - BundleFiles are not closed for removal pending wirings fromI20160209-0800Thomas Watson1-0/+6
uninstalled bundles on shutdown Change-Id: I184f35602aa82e28694f408f492a11b6d6d11ca9 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-02-08Bug 486969 Optionally imported packages prevent a package from beingSamuel Bratton4-20/+161
dynamically imported. Change-Id: I66d300775edc19077e67bf8e480285e3d3a70dae Signed-off-by: Samuel Bratton <sbratton@us.ibm.com>
2016-02-04Bug 478175 - Adapt EE definition to java 9s new version scheme (JEP 223Thomas Watson3-6/+8
ak verona) - rename JavaSE-9.0 execution environment to JavaSE-9 Change-Id: Iaedb1185949ae5d759ca43780d40c6e86b6a16b1 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-01-28Bug 486662 - Neon features and bundles that might need a versionI20160202-1000I20160202-0800I20160128-2000Thomas Watson2-2/+2
increase Change-Id: Idafc952fff4a28f0aafb583014cc6846cb05656c Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-01-21Bug 483799 - Fragments with the 'system.bundle' host are not attached onI20160127-2000I20160127-0800I20160126-2000I20160126-0800I20160125-2000I20160125-0800I20160125-0400I20160124-2000Thomas Watson1-16/+45
incremental refresh - Fix getOndemandResources to properly lookup fragments with alias host names Change-Id: I7a64ce41706525b5ed21aafc4e07e9e026738d51 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-01-21Bug 483799 - Fragments with the 'system.bundle' host are not attached onThomas Watson2-1/+37
incremental refresh - add a testcase Change-Id: Ib97698223726a5950140db09d9fd7d749c0a83b3 Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-01-21Bug 486108: Latest felix resolver code has compile warnings in buildMarkus Keller1-1/+1
Change-Id: Ife03b2bee9bad7939479a25967ae8dadb9ad42fc Signed-off-by: Markus Keller <markus_keller@ch.ibm.com>
2016-01-18Bug 478175 - Adapt EE definition to java 9s new version scheme (JEP 223I20160119-0800Thomas Watson3-9/+62
ak verona) Change-Id: I395c83ef3863a3d73598687819506c4d3e0c50de Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-01-14Bug 485692 - Add test to get class loader from system bundle wiringThomas Watson1-0/+7
2016-01-12Bug 485695 - WrappedResource getCapabilities and getRequirements are incorrectI20160112-1800Thomas Watson1-52/+62
2016-01-11Bug 485217 - Update resolver implementation and default to using batchI20160112-0800Thomas Watson2-19/+36
resolves with a timeout Add an option to set the number of resolver threads. If set to 1 then only do single threaded resolve with a simple executor
2016-01-11Bug 485217 - Update resolver implementation and default to using batch ↵Thomas Watson1-22/+68
resolves with a timeout Fix felix resolver to implement add methods for List impl
2016-01-11Bug 485217 - Update resolver implementation and default to using batch ↵Thomas Watson1-3/+5
resolves with a timeout Update the Resolver service registration to be single threaded. Update the Resolver service registration to use Min service ranking. Some ResolveContext impls are not ready to be multi-threaded.
2016-01-07Bug 485300 - Bundle manifest with double byte UTF chars are parsedThomas Watson3-37/+142
incorrectly with line continuations Change-Id: I5f3c9a13db35683aa98820f990e8614acbc8b2ca Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-01-05Bug 485217 - Update resolver implementation and default to using batchThomas Watson29-1694/+3978
resolves with a timeout
2016-01-05Binaries v20160105-0445I20160105-1000I20160105-0800Arun Thondapu21-21/+21
2016-01-05Bug 383545 - Rename carbon folder into cocoaArun Thondapu2-1/+2
Export ARCHS="-arch x86_64" as it needs to be passed to the compiler as well as the linker Change-Id: Ibf595c272dfef30ff5510a92c5fcc7fc469ebec9 Signed-off-by: Arun Thondapu <arunkumar.thondapu@in.ibm.com>

Back to the top