Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Unstable result of lz4 compression #12438

Closed
pcanal opened this issue Mar 6, 2023 · 2 comments · Fixed by #12439
Closed

Unstable result of lz4 compression #12438

pcanal opened this issue Mar 6, 2023 · 2 comments · Fixed by #12439
Assignees
@pcanal
Labels
bug
Milestone
6.30/00

Comments

@pcanal
Copy link
Member

pcanal commented Mar 6, 2023

As seen in https://lcgapp-services.cern.ch/root-jenkins/job/root-nightly-master/3440/LABEL=ROOT-centos9,SPEC=noimt,V=master/testReport/junit/projectroot.roottest.root.io/filemerger/roottest_root_io_filemerger_simple_lz4_compr_level1 :

hadd Target file: hsimple401.root
hadd compression setting for all output: 401
hadd Source file 1: hsimple.root
hadd Target path: hsimple401.root:/
-- TEST COMMAND -- 
cd /home/sftnight/build/night/LABEL/ROOT-centos9/SPEC/noimt/V/master/build/roottest/root/io/filemerger
/usr/bin/timeout -s USR2 270s /home/sftnight/build/night/LABEL/ROOT-centos9/SPEC/noimt/V/master/build/bin/root.exe -q -l -b /home/sftnight/build/night/LABEL/ROOT-centos9/SPEC/noimt/V/master/roottest/root/io/filemerger/testSimpleFile.C("hsimple401.root",25000,401,418980,7)
-- BEGIN TEST OUTPUT --

Processing /home/sftnight/build/night/LABEL/ROOT-centos9/SPEC/noimt/V/master/roottest/root/io/filemerger/testSimpleFile.C("hsimple401.root",25000,401,418980,7)...
Checking hsimple401.root
(int) 1000

-- END TEST OUTPUT --
-- BEGIN TEST ERROR --
Error in <testSimpleFile>: Disk size of hsimple401.root should have been 418980 but is 418988 (tolerance 7 bytes)


-- END TEST ERROR --
CMake Error at /home/sftnight/build/night/LABEL/ROOT-centos9/SPEC/noimt/V/master/build/RootTestDriver.cmake:186 (message):
  got exit code 255 but expected 0
@pcanal pcanal added the bug label Mar 6, 2023
@pcanal
Copy link
Member Author

pcanal commented Mar 6, 2023

The problem is actually due to an instability of the detection of the kIsOnHeap bit where sometimes the stack has the bit pattern that is used to detect that the memory was allocated by TObject::operator new.

pcanal added a commit to pcanal/root that referenced this issue Mar 6, 2023
@pcanal
Do not record kIsOnHeap nor kNotDeleted
0f9b635 
When reading back the "heap status" is determine by how the
read-into object was allocated, so there is no need to record
that information.

When reading back, the kNotDeleted is always on, (and should
always be on when writing :) ), so there is no need to record it.

Recording '0' for both allow for (slightly) better compression.

Fix root-project#12438
@pcanal pcanal mentioned this issue Mar 6, 2023
Merged
pcanal added a commit that referenced this issue Mar 8, 2023
@pcanal
Do not record kIsOnHeap nor kNotDeleted
3038a52 
When reading back the "heap status" is determine by how the
read-into object was allocated, so there is no need to record
that information.

When reading back, the kNotDeleted is always on, (and should
always be on when writing :) ), so there is no need to record it.

Recording '0' for both allow for (slightly) better compression.

Fix #12438
@pcanal pcanal added this to the 6.30/00 milestone Mar 8, 2023
@github-actions
Copy link

github-actions bot commented Mar 9, 2023

Hi @pcanal,

It appears this issue is closed, but wasn't yet added to a project. Please add upcoming versions that will include the fix, or 'not applicable' otherwise.

Sincerely,
🤖

omazapa pushed a commit to omazapa/root that referenced this issue Apr 13, 2023
@pcanal @omazapa
Do not record kIsOnHeap nor kNotDeleted
f42f2d7 
When reading back the "heap status" is determine by how the
read-into object was allocated, so there is no need to record
that information.

When reading back, the kNotDeleted is always on, (and should
always be on when writing :) ), so there is no need to record it.

Recording '0' for both allow for (slightly) better compression.

Fix root-project#12438
@pcanal pcanal mentioned this issue Jul 2, 2024
Open
1 task
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@pcanal pcanal

Labels
bug
Projects
None yet
Milestone
6.30/00
Development

Successfully merging a pull request may close this issue.

Do not record kIsOnHeap nor kNotDeleted pcanal/root
1 participant
@pcanal