Closed AndrewLevin closed 4 years ago
The PR seems good. But, it indeed fixes only the most urgent issue with the DAC scans which affects/ed P5 commissioning operations. A full review of all the file accesses will probably be required in the future.
Yes, I will follow up, although I think this is only place where we are actually copying or moving a file inside of a python script -- if we just change the file permissions, it shouldn't matter if the file is flushed, right?
Yes, I will follow up, although I think this is only place where we are actually copying or moving a file inside of a python script -- if we just change the file permissions, it shouldn't matter if the file is flushed, right?
Right, changing the file permissions shouldn't matter. And, indeed, your PR fixes the only known issue caused by files not being flushed after a write.
However, depending on how the files and the code are used in the future, every file write is a potential issue... For example, that code in testConnectivity.py
is OK since the file is closed. But, I haven't all file writes.
OK, yes, I see, that would be a problem if the file were not closed.
needs a rebase
rebased
This pull request fixes the issue of files being accessed prematurely by using a
with...as
block which defines the scope of the file, so it is automatically flushed and closed at the end of that block.Description
In addition, this pull request uses a
pandas
Series
to output the nominal DAC values instead of thewrite()
function.Types of changes
Motivation and Context
This pull request partially resolves https://github.com/cms-gem-daq-project/vfatqc-python-scripts/issues/291. I will make another pull request
vfatqc-python-scripts
to fix the unclosed files there.How Has This Been Tested?
Yes, I reran a DAC scan analysis and found that the output NominalValues files were the same as when the HEAD of develop was used.
Screenshots (if appropriate):
Checklist: