Skip to main content
aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2017-09-04Bug 521827 - Move artifact.repository bundle to Java 8 Alexander Kurtakov29-95/+227
* Bump BREE. * Bump minor version for the BREE change. * Enable more warnings. * Add missing Override annotations. * Remove redundant type declarations. * Try-with-resources. * Lambda conversion. Change-Id: I198af985183d1429fb18ed3198f238f42afd0c25 Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2017-08-31Bug 518031 - XML External Entity Vulnerability in Eclipse IDEI20170904-0230I20170903-2000I20170902-1500I20170901-2000Brian de Alwis1-4/+3
Ensure XML processors are configured to use XMLConstants.FEATURE_SECURE_PROCESSING=true to avoid accessing external DTDs and expanding external entities. Change-Id: Ic29e4a0aab1ea5f642ce49914bc6fcecd238efe8 Signed-off-by: Brian de Alwis <bsd@mt.ca>
2017-06-22Bug 517929 - Update POMs in the master for 4.8 releaseI20170622-0950Alexander Kurtakov1-1/+1
Point to 4.8 parent pom. Change-Id: Ic754058b23731334425f98aaca527bcf90991a7d Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2017-05-11Bug 516457 - Many of plug-ins need a version increaseAlexander Kurtakov2-2/+2
Update the following (only service bump): org.eclipse.equinox.p2.artifact.repository org.eclipse.equinox.p2.console org.eclipse.equinox.p2.metadata org.eclipse.equinox.p2.metadata.repository org.eclipse.equinox.p2.operations Change-Id: Ic0eff63a2836ab7068e2c6dc3bef1648c764843a Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2017-04-24Some mirroring tests use wrong MessagesI20170425-0910I20170425-0700I20170425-0240I20170425-0215I20170424-2000Mykola Nikishov2-3/+0
MD5-related test methods in ArtifactMirrorApplicationTest, NewMirrorApplicationArtifactTest and MirrorTaskTest should assert log message using the same Messages class as MD5ArtifactComparator's compare method uses when generating the log message in question. Change-Id: I6986d1d5748d55bed9589d71e9c4f62dfde91665 Signed-off-by: Mykola Nikishov <mn@mn.com.ua>
2017-04-06Bug 514807 - Avoid instantiation of empty array for processing stepsY20170413-1000I20170413-2000Karsten Thoms1-2/+3
Change-Id: Ie1a3b45efc2ce33b871db4476c4b86e7ab74f3a4 Signed-off-by: Karsten Thoms <karsten.thoms@itemis.de>
2017-03-30Bug 423715 - Rename checksum-related constants in SimpleArtifactRepositoryI20170402-2000I20170401-2000I20170331-2000I20170330-2000Mykola Nikishov2-7/+16
SimpleArtifactRepository provides two properties to control if p2 client should validate artifact's MD5 checksums. Rename constants that store names of these properties to better reflect their purpose. Change-Id: I2147e366fb5d20da21c123998cf9c73510e8de24 Signed-off-by: Mykola Nikishov <mn@mn.com.ua>
2016-06-28Bug 496465 - Prep for 4.7 (Oxygen) and 4.6.1 (Neon.1) maintenance buildsThomas Watson1-1/+1
Change-Id: I90a85375f7864204f482d76130da772798f0a28b Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2016-04-19Bug 489706 - Replace new Integer() with Integer.valueOf()I20160419-0800Alex Blewitt3-8/+8
There are a lot of new Integer() calls in the P2 codebase, which results in many duplicate values of Integer being stored, particularly when performing updates or checks. Integer.valueOf() performs this caching already, which means that it's unnecessary to call new Integer() and instead can call Integer.valueOf(). In the places where the int value is being used directly, the call can be further optimised to Integer.parseInt() instead. Replace calls to new Integer() with Integer.valueOf() or .parseInt() as appropriate. Change-Id: Ic760e66084c856fc90cb7c8a358007c975213638 Signed-off-by: Alex Blewitt <alex.blewitt@gmail.com>
2016-01-02update pomsPascal Rapicault1-1/+1
Change-Id: Icc9608edfcdd0a615c4edf42df4398a1f30a271c
2016-01-02Update bundles versionsPascal Rapicault1-1/+1
Change-Id: I479ee5a8b507090ce496e7b334e0055e027896fe
2016-01-02Bug 466249 - Propagate more exceptions to ease problem analysisEike Stepper2-3/+7
We often get problem reports with stack traces that "end" somewhere in p2 even though it's obvious from looking at these places that there was a causing exception. These causing exceptions are often not propagated with the new exception that p2 creates and throws. I'm not sure if that's on purpose or just an oversight. I've prepared a proposal to enhance these places and make it easier to analyze the problems that users report. Change-Id: Iaa53448c53c18301113b42dbe80558eccec49e8e Signed-off-by: Eike Stepper <stepper@esc-net.de>
2015-10-12List childs' URI in composite repositories' toString()I20151103-0800I20151029-1100I20151028-2000I20151028-0800I20151027-2000I20151027-0800I20151026-2000I20151026-0800I20151025-2000I20151020-0800I20151014-1100I20151013-0800Mykola Nikishov1-0/+5
Change-Id: I1916c643da3158a0f9142b8eb2edae7746bc56a8 Signed-off-by: Mykola Nikishov <mn@mn.com.ua>
2015-09-07Bug 476724 - Remove new Boolean and Boolean.valueOf().booleanValue()I20150929-0800I20150922-0800I20150916-2000I20150916-0800I20150915-2000I20150915-0800I20150914-2000I20150914-1545I20150914-1130I20150914-0800I20150913-2000I20150908-0800Alex Blewitt2-6/+6
The new Boolean constructor creates a new instance of a Boolean object, but it can easily be replaced with Boolean.valueOf which returns the reference to the global Boolean.TRUE or Boolean.FALSE. Replace calls to new Boolean() with Boolean.valueOf() for identical semantics except without object collection. Additionally Boolean.valueOf().booleanValue() is identical to Boolean.parseBoolean() and will result in no garbage. In addition, methods will be (slightly) smaller and parseBoolean will often be in-lined by the JIT, which can often prove that the value is non-null for faster checking. Replace Boolean.valueOf().booleanValue() chains with Boolean.parseBoolean(). Some other tests can use Wrapper.valueOf() to take advantage of the built-in caches that these objects maintain (for values in the range -128..127). Signed-off-by: Alex Blewitt <alex.blewitt@gmail.com> Change-Id: I5da4216a26ffbb6b8fd3365515ee800dd82b36ae
2015-07-02Bug 469765 - Prep for 4.6 (Neon) and 4.5 (Mars) maintenance buildsI20150818-0800I20150811-1400I20150805-2000I20150805-0800I20150804-2000I20150804-0800I20150803-2000I20150803-0800I20150802-2000I20150728-0800I20150721-0800I20150714-0800I20150707-0800Thomas Watson1-1/+1
- update parent poms for 4.6
2015-05-13Bug 464614 Use XZ as compression formats of metadata filesI20150514-2000I20150514-1000I20150513-2000Pascal Rapicault3-2/+133
This part takes care of the artifacts.xml file. Change-Id: If18777d4057ec273d219c4aa2b959fa3093a9a7f
2015-04-23Update versions for MarsThomas Watson2-2/+2
2015-04-01bug 457619: MirrorRequest fails for remote after failing locally withChristian Schneider1-0/+15
"is a folder but the repository is an archive or remote location" MirrorRequest discards descriptor when changing sourceRepository Change-Id: I513abf4adaab4b91a9b54ea473b0b67e29950b7a Signed-off-by: Christian Schneider <schneider@yatta.de>
2015-02-26Bug 460967 - Use type safe service retrievingI20150303-0800Alexander Kurtakov2-4/+6
Instead of retrieving services based on their String class name there is newer implementation that takes the class directly and returns the correct class preventing casts. Change-Id: I8b26f4d294b927f4c28a6ba4e68e35bd401df4ec Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2015-02-24Bug 460627 - Fix compiler problems from generified IAdaptable#getAdapterAlexander Kurtakov2-7/+6
Generify where needed and bump the min version requirement for o.e.core.runtime and o.e.equinox.common to ensure that generified dependency will be used. Change-Id: I3e177a3c8384b5aaabe4c77c949e9802e53923b1 Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2015-01-23Bug 458123 - don't mask Error or RuntimeException from getArtifact inI20150126-0800I20150125-2000Christian Schneider1-3/+20
MirrorRequest Change-Id: Iac5f2ae2666d539805f899f64dece3361b882675 Signed-off-by: Christian Schneider <schneider@yatta.de>
2014-06-19Bug 436736 - Prep for 4.5 (Mars) and 4.4 maintenance buildsPaul Webster1-1/+1
Update parent versions from 4.4.0 to 4.5.0 Signed-off-by: Paul Webster <pwebster@ca.ibm.com>
2013-12-11bug 423827 - Disable "Treat above errors like fatal compile errors"I20131217-0800I20131212-1600I20131211-2000Krzysztof Daniel1-1/+27
Change-Id: If2b8652ff7deed04dea308880d444f1b4e7c6ffe Signed-off-by: Krzysztof Daniel <kdaniel@redhat.com>
2013-12-10bug 423628 - False positives warnings in the official build.I20131211-0800I20131210-2000I20131210-0800Krzysztof Daniel1-1/+2
False positives are caused by projects settings being different than build settings. Typical example is that project is configured to warn about local variable hiding a field, and suppression is added, but the regular releng build is not configured to warn, and it finds unnecessary suppression. I have did following things: * fixed warnings that I was able to fix (renamed variables, added generics, comments, etc, etc). * reduced the number of warnings in tests * commented warnings that must not be suppresed in the workspace unless we want "Unnecessary @SuppressWarnings" at build time. Change-Id: Iab2c962e5c331eab170b5f6e08d9bb29d05f0cc4
2013-12-09bug 418662: Get rid of compile warnings in official build -Krzysztof Daniel1-1/+2
org.eclipse.equinox.p2.publisher.eclipse Change-Id: I02f9f6367913e01988f5e65452d2c52c7898a5a7 Signed-off-by: Krzysztof Daniel <kdaniel@redhat.com>
2013-11-19Bug 422026 - Get rid of ColletionUtils.empty[Set|List|Map]Alexander Kurtakov1-2/+2
Now that p2 has Java 1.5 as a minimum there is no reason to not use the JVM methods. Change-Id: I89444e8d1e174c316e2b17fb4f53bc7b1d097c0a Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2013-11-15Bug 421798 - Fix various warnings.Alexander Kurtakov1-1/+1
* Remove unneeded casts. * Remove unneeded throws. * Remove unneeded semicolons. Change-Id: I7454ea216cfc194a0d394550feecc29042d23363 Signed-off-by: Alexander Kurtakov <akurtako@redhat.com>
2013-11-13Bug 421339 - Accept java 1.5Krzysztof Daniel1-2/+1
Make 1.5 setting default in Eclipse (P2 is compiled as 1.5 with tycho anyway). Drop jsr14. Change-Id: I848d7865211126ad32a226914a046f9553a48f04 Signed-off-by: Krzysztof Daniel <kdaniel@redhat.com>
2013-07-11Increment bundle version and tidyI20130723-0800I20130717-1600I20130716-0800John Arthorne3-4/+3
2013-07-11Bug 411035 - Some tests depend on ↵Thomas Watson1-1/+9
SignedContentFactory.getSignedContent(File) to return an empty unsigned SignedContent object when the input is an invalid zip file. Signed-off-by: Thomas Watson <tjwatson@us.ibm.com>
2013-06-26Bug 411170 -- Updated the parent poms for Luna (Eclipse 4.4)Ian Bull1-1/+1
2013-05-15Bug 407707: A number of .class files changed, after changing to M7 JDT compilerIan Bull1-0/+1
Updates two bundles because of a new JDT compiler.
2013-03-27Bug 403971 - Update Equinox p2 parent versionsI20130402-0800Pascal Rapicault2-3/+4
2013-02-25Bug 397850 - [CBI] use the correct group IDs for all artifactsv20130225-185631Ian Bull1-1/+1
Upated all the pom.xml files (in the bundles) to include the proper groupID. The groupID is the first three segments of the artifactID.
2013-01-20Revert "Remove non existant dependency"v20130120-161028Pascal Rapicault1-1/+2
This reverts commit 77ba1dec8f5448479ca50611e3ab8af509939304.
2013-01-19Remove non existant dependencyPascal Rapicault1-2/+1
2012-11-21Bug 390385 - Stop the use of jsr14v20121121-194827Pascal Rapicault (Ericsson)1-2/+0
2012-11-14Updated manifest for bundles that use Java 5 class library methodsv20121114-150134John Arthorne1-3/+1
2012-10-11356561 Revert default of composite loading back to non-atomicv20121011-091400I20121030-0800I20121029-2000I20121029-0800I20121028-2000I20121024-1200I20121024-1130I20121023-0800I20121016-2240I20121016-0800Tobias Oberlies1-2/+1
- There has been a suggestion of how to fix non-atomic loading (i.e. implement it in a way that network problems are not silently ignored), so we may keep non-atomic loading as the default. Bug: 356561 Composite repositories ignores errors when loading children (errors are treated like missing children)
2012-10-09391400 Include repository URL in status if getArtifacts failsTobias Oberlies4-6/+15
Bug: 391400 Status returned by getArtifacts does not include repository URL of problematic repository in case of an error
2012-09-28356561 Require that all children of a composite repository can be loadedv20120928-180332Tobias Oberlies1-1/+3
- Loading a composite repository now fails if any of its children cannot be loaded. So in case of transient network problems, a p2 operation now fails fast, instead of generating obscure (and possibly severe, yet unnoticed) follow-up problems. - p2 repository providers who rely on the old, lenient behaviour may re-enable it by adding the property 'p2.atomic.composite.loading' with value 'false' in their compositeContent.xml/ compositeArtifact.xml. - p2 clients who need to communicate with broken composite repositories which are out of maintenance can revert to the old behaviour by setting the system property eclipse.p2.atomic.composite.loading.default=false. Bug: 356561 Composite Repositories have non-deterministic content by default
2012-08-29Bug 381673: Fixes a few compile warningsv20120829-182520Ian Bull2-6/+4
This fixes 2 types of compile warnings. 1. Stream not closed warnings. 2. Unused object warnings
2012-04-30Bug 378095 - Install new Software writes to stdout v20120430-1959Ian Bull1-3/+0
I accidently committed some debug output.
2012-04-27Bug 377976 : Don't continually try to download bad artifacts.v20120427-2140Ian Bull3-3/+29
If there is a processing error (non-transport error), we can now set the status code ARTIFACT_PROCESSING_ERROR. This will skip the retry attempts. https://bugs.eclipse.org/bugs/show_bug.cgi?id=377976
2012-02-09Bug 369834.v20120209-1054Mengxin Zhu1-4/+2
Rework the download progress event reports in the scenario of multiple agents. Signed-off-by: Mengxin Zhu <kane.zhu@windriver.com>
2012-01-17364929 support customizable download statisticsv20120117-1140v20120117-0330Mengxin Zhu3-4/+29
Signed-off-by: Mengxin Zhu <kane.zhu@windriver.com>
2012-01-12316328 engine should be more verbose while performing an installationv20120112-1825Mengxin Zhu4-7/+76
1. the event of phase starts and finishes 2. download events, including how many artifacts to be downloaded, how many artifacts to be downloaded from a specific repository, the mirror request(download) result, download progress event(a wrapper of ProgressStatistics) 3. the events before/after configuring/unconfiguring an IU Signed-off-by: Mengxin Zhu <kane.zhu@windriver.com>
2011-12-19Bug 366781 - isReadOnly(File) doesn't work on Windows.v20111219-0328Mengxin Zhu1-6/+12
It wants to determine whether new files can be created under the given file that actually denotes a folder. There's no more elegant way besides creating a temporary file for testing. Signed-off-by: Mengxin Zhu <kane.zhu@windriver.com>
2011-12-07Fix 351944 to avoid unnecessary loading artifact repository when doing ↵v20111207-0646Mengxin Zhu1-4/+12
query, getting artifact descriptors via artifact key. Without fixing though simple artifact repository always checks the last modify of artifacts.jar file, however it still cost some time, it will waste a lot of time especially the repository locates on a read-only NFS server. This version of fix introduces the dependency on o.e.core.filesystem to use EFS API to change the file permission crossing platforms. Signed-off-by: Mengxin Zhu <kane.zhu@windriver.com>
2011-12-05Revert "Fix 351944 to avoid unnecessary loading artifact repository when ↵v20111205-2311Ian Bull1-12/+4
doing query, getting artifact descriptors via artifact key." This reverts commit 82aeb8dee11f4114494fe3dd0ca1097312d2dfd5.

Back to the top