Skip to main content
aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2015-03-06Remove content of the lttng folder and corresponding target dependenciesMarc-Andre Laperle51-9913/+0
The LTTng component (also known as TMF) of Linux Tools has moved to the Trace Compass project. A README file was added to explain that the code moved. Change-Id: I939275594beccc78155f37671d1765539764eabc Signed-off-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com> Reviewed-on: https://git.eclipse.org/r/43270 Tested-by: Hudson CI Reviewed-by: Alexander Kurtakov <akurtako@redhat.com> Tested-by: Alexander Kurtakov <akurtako@redhat.com>
2014-10-01lttng: Annotate 3.2 APIsAlexandre Montplaisir2-5/+5
Many APIs were not actually part of 3.1, mark them @since 3.2 instead. Change-Id: Idf17db685b4c0ed57c7e772f2718bdb0bfeffc03 Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/34242 Reviewed-by: Genevieve Bastien <gbastien+lttng@versatic.net> Tested-by: Hudson CI
2014-09-17ctf: make CTFTraceReader access streams in consistent wayMatthew Khouzam1-2/+2
make all accesses pass by getTopStream() accessor Change-Id: I41b21289d8d7bff7bcf56273545e2c1190af1eef Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/32566 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-09-15ctf: Reduce complexity of CTFTrace#openStreamInputMatthew Khouzam1-31/+34
Change-Id: Id9ffa4ddffbe5e44640c5703af9ca1581bbfb74b Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/32563 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-09-03ctf: Provide a utility for "safe" mapping of a ByteBufferMarc-Andre Laperle3-7/+9
This works around a bug on Windows which prevents deleting a file after it was mapped. See http://bugs.java.com/view_bug.do?bug_id=4715154 Running TraceSeekBenchmark, I noticed a 10% decreased in performance (3.8s vs 3.4s) on Windows. On Linux, the difference is negligible (2.41s vs 2.40). Change-Id: Ie590709df4149708a37257ee3bc3d3d017f31c21 Signed-off-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com> Reviewed-on: https://git.eclipse.org/r/32448 Tested-by: Hudson CI
2014-09-02ctf: make CtfStreamInput not hold onto any resourcesMatthew Khouzam5-85/+115
Change-Id: I61cb09c4c325479115084c54bba03d4746cd6610 Reviewed-on: https://git.eclipse.org/r/32553 Tested-by: Hudson CI Reviewed-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com>
2014-08-22ctf: improve unit testsMatthew Khouzam1-3/+3
There were several recent changes introduced to the ctf parser without fixing the test suite. This patch brings the tests up to date and coverage up to 78% Change-Id: I38b0f09d4005934d7a05d8cc3384ee7767bbd321 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/31936 Tested-by: Hudson CI Reviewed-by: Bernd Hufmann <bernd.hufmann@ericsson.com>
2014-08-21ctf: replace HashMaps with ArrayLists for EventDeclaration storageMatthew Khouzam4-35/+142
Parser no longer supports events with IDs larger than Integer.MAX_VALUE Change-Id: I088943c5a041f16638cb89b832ab79f8ef9c1d76 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/25924 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-08-18ctf: make id conversions in event headers saferMatthew Khouzam1-2/+3
Change-Id: I4fcfb986f5e1ff0010fcd4b6692075c3dbfd175e Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/31569 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-08-13ctf: add Event header data typesMatthew Khouzam10-83/+259
This is a composite datatype described in the examples of the CTF spec, section 6.1.1 . Right now, chances are that many people will re-use this pattern so why not enhance its performance. Change-Id: I2042be35fee38e42680691af879357ebb1aff95b Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/25937 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-08-04CTF: Do not recalculate the single clock each time in CTFTrace#getClock()Geneviève Bastien1-1/+4
Change-Id: Ia42daadc9606335073da59a547e93867f445ed91 Signed-off-by: Geneviève Bastien <gbastien+lttng@versatic.net> Reviewed-on: https://git.eclipse.org/r/30990 Tested-by: Hudson CI Reviewed-by: Matthew Khouzam <matthew.khouzam@ericsson.com>
2014-08-04ctf: remove duplicate code in array/sequence declarationsMatthew Khouzam3-62/+27
Change-Id: I6fcb6325ae24de3984412a151137f8a5dca83deb Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30837 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-08-02ctf: remove redundant code in utilMatthew Khouzam2-16/+7
Change-Id: I4db24325762b054a031925e6fb0bc861a6511429 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30835 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-08-02ctf: remove redundant code in structDeclarationMatthew Khouzam1-19/+11
Change-Id: I0945d838d214cea852f5929c5c4ffbf4de6d50b6 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30836 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-07-28ctf: Make LabelAndRange static in EnumDeclarationMatthew Khouzam1-2/+11
Change-Id: I9bab7c058fec93879f9ec5b6eef5ef55fb7405c6 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30155 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-07-25ctf: remove redundant null check in StructDeclarationMatthew Khouzam1-10/+5
Should speed things up too! Change-Id: I011c3ddcbb3750e15a8a19c6e328c22a637b0661 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30154 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-07-25ctf: fix call to equals of two different types (CTFStreamIn...)Matthew Khouzam1-2/+2
CTFStreamInputPacketReader's lookup definition compares the lookup path to lexical scopes, and not their strings. It should be comparing the strings. Change-Id: I35aafb3462ca4e03906d3d8d7636bf49581f459a Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30151 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-07-25ctf: add hashcode and equals to CTFCallsiteMatthew Khouzam1-0/+55
Change-Id: Ia63701468c591a78a37ae245964d80993e39fb5b Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30150 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-07-25ctf: remove redundant null check in lexical scopeMatthew Khouzam1-9/+9
Change-Id: Id00df1f7a43e69d770790c63e7d91fdb871129e3 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/30149 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-07-10ctf: introducing the ICompositeDefinitionMatthew Khouzam3-19/+49
This is a definition that contains subfields. Change-Id: I347de51cb12022262891b59c00d0342c52bc7439 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/29628 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-07-09ctf: introducing IDefinitionMatthew Khouzam6-22/+57
This is a step in adding a real api to CTF. Now an IDefinition can be used instead of a definition. Change-Id: I2f1cea03265afcfd1125ccecff4657324b55ad84 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/29573 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-07-09ctf: accelerate n-grams for ctf scopesMatthew Khouzam14-26/+434
CTF-based LTTng traces have 30-40% of their requests directed to the same fields, this makes looking up these field _names_ faster. Change-Id: I3cf32b10d22de04a1bd94ecadde37afa8e50b0a1 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/25861 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-07-07ctf: minor CTFTrace cleanupMatthew Khouzam1-9/+18
moved IDefinition scope stuff together Change-Id: I5d21ef90d44b1cc6f3824d6050e538888f8f2223 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/25860 Tested-by: Hudson CI
2014-06-30ctf: Flatten fixed-size structsMatthew Khouzam4-5/+28
This patch will unroll nested arrays and structs retaining only floats, enums and integers in a 1d array. This means we won't need to create "useless" definitions. Change-Id: Ia22730b6479df5b4a4f296e609bb98bdc4f1736e Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/25926 Tested-by: Hudson CI
2014-06-26ctf: Cleanup StructDeclarationAlexandre Montplaisir3-47/+75
No need to keep separate lists for the fields and values of structs. A LinkedHashMap preserves the insertion order, and Iterable defines a specific iteration order. No apparent performance impact, both versions hover around 975-980 ns/ev on my machine. Change-Id: I487b2c756d5827a7489a2ef65e550fb3d018cb72 Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/29009 Reviewed-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Tested-by: Hudson CI
2014-06-26ctf: make BitBuffer's Constructor only take non-null ByteBuffersMatthew Khouzam3-15/+16
Change-Id: I70490af35ea9d1ec559828d81fcb6b65b65d12ca Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/28685 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-06-25ctf: make BitBuffer's buffer @NonNullMatthew Khouzam1-17/+11
This accelerates canread and thus every read Change-Id: I53254b9d7760b9f7a4915751a9fedb4297a5ebb7 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/28684
2014-06-25ctf: make BitBuffer's ByteBuffer finalMatthew Khouzam5-38/+46
No more chances of someone swapping it while you're reading. Remove an unnecessary alignment issue and add a proper byteorder check. Change-Id: I344ed316489918dc0751c09c464fada3d4ac02ab Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/28681 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-06-20ctf: Introducing the ByteArrayDefinitionMatthew Khouzam11-17/+179
This patch introduces a special acceleration that will make sequences and arrays read faster when it's an array of bytes or a string, as the CTF spec specifies that they are aligned on 8 bits and always 8 bits wide per character. Sequences and arrays can either be ArrayDefinitions or ByteArrayDefintions. This patch gives a 20% acceleration in sequential reading. Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Change-Id: I7933268b7f9271f74f7d59339ea9b8343fdb021e Reviewed-on: https://git.eclipse.org/r/26084 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-06-18ctf: remove asserts from bitbufferMatthew Khouzam1-4/+12
Change-Id: I8a0dcff9f365a313c40032f4b5b4c2b95de91ea6 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/28680 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-06-18ctf: Fix simple resource leak warningsAlexandre Montplaisir2-33/+6
Many warnings left though. For instance, the Stream/StreamInput/ StreamInputReader etc. constructor should create their inner Closeable objects themselves (instead of having to pass say new StreamInput(new StreamInputReader(...)) ). That way they can keep a reference to it, and close() it in their own close() method. Right now nobody keeps a reference to them, and they never get closed. Change-Id: I708cfb7d2da12ea4d5c7c8f8a8bbcdf77263ddc1 Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/26485 Reviewed-by: Genevieve Bastien <gbastien+lttng@versatic.net> Tested-by: Hudson CI
2014-06-17ctf: add notion of ScopedDefinitionMatthew Khouzam3-194/+152
extract common code from variants and structs into a ScopedDefinition Change-Id: I702dc54227efccace892b5787bc93a8a333c3e1a Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/28492 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-06-17ctf: Add get(byte[]) to BitBufferMatthew Khouzam1-0/+28
This adds a bulk reader for data. Change-Id: I8a43b292b8240d070a94473861b7aca711474661 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/26083 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-06-05tmf-ctf: fix NPE in callsite handling (bug 436625)Bernd Hufmann1-0/+3
Change-Id: Iaaecf310226eeb84cfa30c69035eb8f34cda3c91 Signed-off-by: Bernd Hufmann <Bernd.Hufmann@ericsson.com> Reviewed-on: https://git.eclipse.org/r/27991 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI Reviewed-by: Bernd Hufmann <bernd.hufmann@ericsson.com> Tested-by: Bernd Hufmann <bernd.hufmann@ericsson.com> Reviewed-on: https://git.eclipse.org/r/28025 Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-06-04tmf: Bug 421475: Inaccurate seek for traces with timestamp transformPatrick Tasse1-0/+30
The CtfLocation now holds the original (untransformed) timestamp value. Change-Id: I8f1b435469fd3c9ee32720fc367524d734abb145 Signed-off-by: Patrick Tasse <patrick.tasse@gmail.com> Reviewed-on: https://git.eclipse.org/r/26750 Tested-by: Hudson CI
2014-05-22ctf: Rename Stream* classes to CTFStream*Alexandre Montplaisir8-102/+79
"Stream" will become very confusing with java.util.stream.Stream when Java 8 comes. Rename those classes to CTFStream, for better namespacing. Change-Id: Ifb239af0366553472463c50d8e6ed899e002f880 Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/27067 Tested-by: Hudson CI Reviewed-by: Matthew Khouzam <matthew.khouzam@ericsson.com>
2014-05-13ctf: add scope to packet headerMatthew Khouzam1-1/+1
add a trace defintion scope to a packet header, instead of it being null. Change-Id: I6e8d361101fb0a71f25665f921378131744d8d92 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/25862 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-05-09ctf: get iterable of stream input readersMatthew Khouzam2-0/+31
This patch adds a way to get all event declarations from a CTFTraceReader Change-Id: If12805ba6436ab5f5ce562686f5870c0fbf05261 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/23284
2014-05-02ctf: minor LexicalScope fixupsMatthew Khouzam1-2/+3
Two mislabelled scopes were fixes Change-Id: I2a44d3fda57551f84a320432493c1e1d1cb61e8f Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/25859 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-05-02lttng: Add Lttng relayd connectorMatthew Khouzam2-13/+58
This allows connections to the lttng session daemon so that live trace reading can work. (rebased on lttng: Rename lttng2 feature/plugins to lttng2.control) Conflicts: Everything (then rebased on ctf: immutable events) LttngRelaydConsumer updated Change-Id: Ida7f76f09d3f01a25b67cc552cb5869dc5e4c3a4 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Signed-off-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com> Reviewed-on: https://git.eclipse.org/r/23564 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-05-02ctf: Make events immutableMatthew Khouzam31-991/+1753
This patch does some major changes to the ctf parser The event definitions are now immutable and can no longer be modified. The declarations will read the bitbuffer and then create event definitions. The event definitions are going to have fields. Initial tests show a slight speedup but the parser is 66% slower without eclipse. Change-Id: I52b8c0de9776fa7cd2b333628c2bb6d3dd2c86ac Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/23740 Tested-by: Hudson CI Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im>
2014-05-02ctf: potential memory optimizationMatthew Khouzam3-58/+1
This patch leads to a 300x memory improvement in lttng kernel traces but a performance regression. Change-Id: I655fa7b8ae94d742ee79d879bbbb5b2ddeec1374 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/23221 Reviewed-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Tested-by: Hudson CI
2014-04-26Change tabs to spaces; elim trailing whitespace.Andrew Ferrazzutti1-1/+1
To comply with Sonar style requirements, replace all tab characters with four whitespace characters/spaces, and remove all trailing whitespace. Perform this change to all Java and XML files. Also replace tabs with "\t" in test strings that require tabs, and edit STPIndenterTest so that it doesn't rely on indented comments. Change-Id: I48c3c5449a58fe5310967d998a05df1a28fbcbb0 Signed-off-by: Andrew Ferrazzutti <aferrazz@redhat.com> Reviewed-on: https://git.eclipse.org/r/25561 Reviewed-by: Alexander Kurtakov <akurtako@redhat.com> Tested-by: Alexander Kurtakov <akurtako@redhat.com>
2014-04-26ctf: java 8 compliance of javadocMatthew Khouzam8-10/+10
Change-Id: I9ad7d732f606365bcdef1608b7c98c6577558e25 Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/25579 Tested-by: Hudson CI Reviewed-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com> Tested-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com>
2014-04-25ctf: make MetadataPacketHeader more inline with coding styleMatthew Khouzam1-40/+59
This fixes some sonar warnings Change-Id: I67b401823d28cedf84d164af8ffe8a6fa01b6afe Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com> Reviewed-on: https://git.eclipse.org/r/25531 Reviewed-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com> Tested-by: Hudson CI
2014-04-25Revert "ctf: make MetadataPacketHeader more inline with coding style"Matthew Khouzam1-59/+39
This reverts commit 2ce5d7294581df7698ccd84f86ebd4d5ab11b392. Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com>
2014-04-25ctf: make MetadataPacketHeader more inline with coding styleMatthew Khouzam1-39/+59
Change-Id: Icafec1945eeb7edbe579a2fc05b20f4f0da7176e Signed-off-by: Matthew Khouzam <matthew.khouzam@ericsson.com>
2014-04-17ctf: Do not call System.gc() in CTFTrace.close()Alexandre Montplaisir1-2/+0
This call is used to workaround a bug/limitation of MappedByteBuffer (see JDK-4724038). But CTFTrace.close() is called quite often (for example, once per trace when closing an experiment), so it is not the best place to call System.gc(). Instead, only call it when we are about to delete/rename a trace. Change-Id: Ib9b346dd04df4f64289ac9db0b976e6b70ff2551 Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/25174 Tested-by: Hudson CI Reviewed-by: Patrick Tasse <patrick.tasse@gmail.com> Tested-by: Patrick Tasse <patrick.tasse@gmail.com>
2014-04-16ctf: Make CTFTrace and its reader classes AutoCloseableAlexandre Montplaisir4-14/+18
Basically just replace their dipose() method with close(). This small change allows us to benefit from the compiler warnings for when a resource is allocated locally in a method but never disposed. Which then allows to find leaks. Many leaks ! Many tests were leaking CTFTraceReaders and the like, but the biggest offender is probably the validation methods in LttngKernelTrace and LttngUstTrace: trying to validate a non-CTF trace could potentially leak a number of file descriptors equal to the number of files in that trace(!) Change-Id: I7e54416b608992a14a7bfd3cd29ffb8dc3cee898 Signed-off-by: Alexandre Montplaisir <alexmonthy@voxpopuli.im> Reviewed-on: https://git.eclipse.org/r/24986 Tested-by: Hudson CI Reviewed-by: Matthew Khouzam <matthew.khouzam@ericsson.com>
2014-04-14ctf: Use Antlr 3.5.2 for parser generation onlyMarc-Andre Laperle1-1/+1
This change keeps the Antlr rutime at the same version (3.2) but upgrades to 3.5.2 at the parser generation time (maven) so that it builds with Java 8. Change-Id: If9969034777534671d9ec23b19495ce2cf341fdb Signed-off-by: Marc-Andre Laperle <marc-andre.laperle@ericsson.com> Reviewed-on: https://git.eclipse.org/r/24821 Tested-by: Hudson CI Reviewed-by: Matthew Khouzam <matthew.khouzam@ericsson.com>

Back to the top