fds
fds copied to clipboard
Nightly validation statistics has undefined date
I pushed up alot of validation data sets yesterday, but the Utilities/Scripts/validation_git_stats.sh
script is having trouble identifying the date. I tried to fix, but could not figure out what was wrong.
Ill take a look
the script was failing because the fds version string format changed (something like when we changed from FDS_6.x.y to FDS-6.x.y ). we needed to specify an extra '-' character in the gitdate/awk portion of the validation_git_stats.sh script to get to the git revision portion of the FDS version string. let me know if the table looks ok after firebot runs tonight
I figured it was something like that, but awk
is something that still eludes me. Things are working now except for the last entry:
should the FDS revision string for the last entry be the same as the previous ones? .
On Fri, Dec 29, 2023 at 11:12 AM Kevin McGrattan @.***> wrote:
I figured it was something like that, but awk is something that still eludes me. Things are working now except for the last entry: image.png (view on web) https://github.com/firemodels/fds/assets/11333911/226287e5-d896-4e80-a02f-5caae2326d7c
— Reply to this email directly, view it on GitHub https://github.com/firemodels/fds/issues/12314#issuecomment-1872195109, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC6UCRWFJMUTNSYK7TZVNL3YL3TX3AVCNFSM6AAAAABBF5KIRKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZSGE4TKMJQHE . You are receiving this because you were assigned.Message ID: @.***>
-- Glenn Forney
No, the git string is right. I just don't know where the date came from
gforney wrote the following in an email to mcgratta, but such arcane minutiae deserves to be preserved in perpetuity:
Let me know if there is a problem with my latest fix. my first fix was really a hack. I recomputed things if i found a variable was blank. this hack failed for that one particular case. this time I found an awk feature that should make the script more robust. the problem was that at some point the number of '-' characters in our FDS revision string increased from 3 to 4. In both cases if NF was the number of - signs, the git revision we wanted would be after the NF-1'st - sign. In awk I can get that by just saying awk -F - '{print $(NF-1)}' . google found this for me
this script lives in both the fds and bot repo. I've had problems in the past with scripts used to make releases living in the fds or smv repos if I needed to make changes to the scripts but needed to build a release for a particular fds/smv revision. this is probably why I put a copy of the script in the bot repo (where firebot uses it). we can rethink this but I would prefer to keep things the way they are at least for now. I'm going to be out next week.
I noticed some extra code in this script not being used. I'm going to remove it but will wait until I get back.
This fix worked. The index of validation cases is correct. I'll leave this issue open should you want to clean up the scripts a bit more.