On 11/9/22 4:33 AM, Toshiki Sonoda wrote: > Fix the virt-install progress bar file size > > These patches improve about the virt-install progress bar. > > Since version 4.0, implementation of the progress bar has changed. > If the file copy about vmlinuz and initrd.img completes quickly > (less than self.update_period), > the process ends without updating the progress bar. > In short, the result is printed as 0B. > > In addition, 0B is printed at "Creating domain" printing. > 0B is meaningless and can be misled as a failed transfer. > > Toshiki Sonoda (3): > virtinstall: Improve the progress bar output by meter.end() > virtinstall: Fix the allocating disk size printed by the progress bar > virtinstall: Hide total_size in the progress bar if it doesn't need > > tests/data/meter/meter7.txt | 2 ++ > tests/data/meter/meter8.txt | 6 ++++ > tests/data/meter/meter9.txt | 4 +++ > tests/test_misc.py | 49 ++++++++++++++++++++++++++++---- > virtManager/object/domain.py | 2 +- > virtinst/_progresspriv.py | 12 +++++--- > virtinst/diskbackend.py | 3 +- > virtinst/install/urlfetcher.py | 6 ++-- > virtinst/install/volumeupload.py | 3 +- > virtinst/progress.py | 4 +-- > virtinst/storage.py | 3 +- > 11 files changed, 76 insertions(+), 18 deletions(-) > create mode 100644 tests/data/meter/meter7.txt > create mode 100644 tests/data/meter/meter8.txt > create mode 100644 tests/data/meter/meter9.txt > Thanks for the report, and the patches! I see all the reported issues with this reproducer: ./virt-install --connect test:///default --location tests/data/fakemedia/fake-f26-netinst.iso Patch #1 was a bit tough to understand, since it's adding test cases, internal API change, and behavior change all at once. In the process of trying to understand it, I found a simpler fix that doesn't require any internal API change. I pushed that as patch #1, and pushed your patches #2 and #3 on top of that, with some adjustments due to patch #1 being different. Please test git and confirm the changes still fix your issues Thanks, Cole
