REST version : v2.2.12_dev
REST commit : 0f0a7eae
When I launch restRoot File.root with a fresh generated ROOT originated from restG4 01.NLDBD example, I get several errors and seg.faults.
No problem when opening with the latest v2.2.12 release.
If I print out the pointer, run0, generated inside restRoot. I get the following message.
(TRestRun *) 0xdb57fc10 <invalid memory address>
Some updates were introduced in commit
If I revert this commit everything works back as usual.
@lobis I hope we can discuss at some point about test units, that allow us to check different parts of the code are still working after updates are pushed to the repository. For example, check write/read all TRest objects, generate a dummy simulation and read the different keys, read a dummy data file, etc.
1 Like
The previous one runs fine for me. Is this another operation system dependent problem?
We must find a way to avoid initializing two TRestRun objects. This is ugly.
I was running that in Linux, after I reverted the commit the problem disappeared, perhaps it is connected with the ROOT version? I am on 6.14/04
I am also on linux(centos 7.5, 64 bit). My root version is 6.12/06.
Somehow I am surprised that it works for you, then the pointer generated outside (in the compiled code) is recognised in the interpreted root interface?
Yes. If you add a cout for the address of the compiled TRestRun object, you can directly use this address in root’s command line
I still cannot reproduce that bug. Can you give me more details about it?
However @BofangJiang experienced similar problems, as reported in this topic
Are you in commit 543e2470?
Yes, and everything is fine when opening the file.
I have tested on
- SJTU bl2 server (CentOS 6.6, gcc 5.2.1)
- pandax3 daq server (CentOS 7.5, gcc 4.8.5)
- my own computer (WSL with Ubuntu 18.04.1, gcc 7.3)
- my virtual machine with ubuntu, 32-bit
None of them yield this problem.
The root versions include:
6.12/06, 6.14/06