Healthchecker cannot properly catch the error for different hash #62
Labels
No Label
bug
component
executor
component
framework
component
parser
component
UI
duplicate
enhancement
help wanted
invalid
priority
p0
priority
p1
priority
p2
priority
p3
question
wontfix
No Milestone
No project
No Assignees
5 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: JOJ/JOJ3#62
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
When testing healthchecker, we encounter following pbs:
ERROR
message on the debug-log on the server.ERROR
message on gitea action for non-ascii stuff, but not for hash detectBy running locally, we use:
Here is the output locally:
and currently the debug-logs on the server looks like:
relevant issue:
engr151-24fa/hteam-00#158 This is for the failed hash detect
engr151-24fa/hteam-00#157 This is for the success non-ascii detect success
For a healtch checker problem, why do you paste the teapot log here?
The hash of which file is wrong in the first issue?
looks like
/home/hydraallen/Desktop/Github/Others/hteam-00/.gitea/workflows/push.yaml
(sopush.yaml
on server)we probably don't need to output the hash values (this is long and students will feel it's complicated so they won't really read...) so drop that sentence.
So the content of
push.yaml
?@allen_wr testing file change on
push.yaml
might not be the easiest test as this file is involved itself in the process. maybe first start with a readonlyReadme
, then a.gitignore
, after thatpush.yaml
and finally a combinations of them. this should cover all cases.test with no other problem (don't mix with non-aascii chars). only combine test types after confirming they all work "individually".
this should hopefully help tracing the source of the bug
Tried
.gitignnore
, same issue. Healthcheck outputs difference in hash for.gitignore
locally, but no output on gitea issue.where is the issue?
.gitignore
is also kind of involved in the process. what about a readonlyreadme
or any other "neutral" file?Tried
README.md
, same issue... The issue is that when files like.gitignore
,push.yaml
orREADME.md
as just tested are set to be unchangeable, we do sha256 check in healthcheck. When running locallyhealthcheck
orrepo-health-check
binary file, it can notice that file has been changed and output the msg as shown below:However, when we tried on gitea, it outputs nothing and healthcheck says everything is fine.
Where is the gitea issue that proves everything happens?
engr151-24fa/hteam-00#169
readme is changed in this commit, healthcheck should fail
add more debug logs before calling the hash check, during, and after. then send the link of the gitea action
first, set and check the debug log file for joj3, not only for teapot.
set and done @bomingzh
so the content of the log?
new issue occurs that:
but I already has
and it seems that healthcheck didn't shut down by this
full content:
need to add more debug logs inside healthcheck, and also log the version of healthcheck
how to achieve?
add slog.Debug in pkg/healthcheck/xxx.go
which server are you using @jon-lee I didn't log to the server for a long time
@zzjc123
tt@111.186.58.48
it looks good now @allen_wr engr151-24fa/hteam-00#170
noticed that, thank!
Checksum for .gitea/workflows/push.yaml failed. Expected 8d1229900c6fc6711b5cc141d1ab5ea7f5b7b7a4b921d9cfa3957408b43ae723, but got 4f4091b5e83442d46936959fefde7b657740cd76f4f3ade217726d90091d386e. Please revert your changes or contact the teaching team if you have a valid reason for adjusting them.
adjust for a more simple message that freshmen can understand, eg.
Some files were altered .gitea/workflows/push.yaml. Please revert your changes or contact the teaching team if you have a valid reason for adjusting them.
(or any other message)basically remove the word
checksum
and the "values" of the checksums.