search

siemens / kas

Setup tool for bitbake based projects
MIT License
371 stars 148 forks source link

Kas is showing errors when updating repos #78

Closed jclsn closed 2 years ago

jclsn commented 2 years ago

I always get errors when repos are updated. Everything is fine though

2022-02-25 11:48:19 - ERROR    - From https://git.openembedded.org/meta-openembedded
2022-02-25 11:48:19 - ERROR    - 23598caea..24f4e6e8d  hardknott-next -> origin/hardknott-next
2022-02-25 11:48:19 - ERROR    - c05ae80ba..0fb490a08  honister       -> origin/honister
2022-02-25 11:48:19 - ERROR    - c05ae80ba..0fb490a08  honister-next  -> origin/honister-next
2022-02-25 11:48:19 - ERROR    - ea223bd3d..cf0ed4239  master         -> origin/master
2022-02-25 11:48:19 - ERROR    - + 67a38f8ed...293e679e7 master-next    -> origin/master-next  (forced update)
2022-02-25 11:48:19 - INFO     - Repository meta-openembedded updated

Version is 2.3.3 on Ubuntu Impish. Maybe the version is outdated?

jclsn commented 2 years ago

Just tried 2.6.3 and same issue

jan-kiszka commented 2 years ago

That ERROR is an annoying side effect of git dumping those messages to stderr. kas just picks up the output streams and classifies them according to stdout vs. stderr. Now, we could say git's stderr is actually stdout, but that may also incorrectly classify real errors. Maybe someone should try to discuss this with git upstream?

jclsn commented 2 years ago

I wouldn't know how to even describe the issue. I still don't understand why there are errors thrown in the first place, when everything works.

jan-kiszka commented 2 years ago

You can observe that by doing git fetch 2>git.stderr >git.stdout against a remote repository that has pending updates since you last updated. Then look at what ends up in the different files. The question for git upstream would be why those update information lines need to go to stderr, rather than stdout. In contrast, when you try that while the remote is offline, the real error will also go to stderr - as it should.

henning-schild commented 2 years ago

What about git fetch -q per default? I guess progress will not be reported on stderr but errors hopefully still.

henning-schild commented 2 years ago

I will give that a try, stay tuned for a patch.

henning-schild commented 2 years ago

@jclsn thanks for the report it looks like the problem will be solved soonish

I took the liberty to dig out your email and take things to the list, hope my guess was right.

Please respect our wish for not using issues and place your next issue on the list. When i see your name again here, no matter how urgent/valid the report will be, i will likely not be in the mood to fix it.

jclsn commented 2 years ago

Okay, thank you for fixing it. I don’t understand your choice regarding mailing lists for issues, but I will use them next time.

Am 25.02.2022 um 22:22 schrieb Henning Schild @.***>:

 @jclsn thanks for the report it looks like the problem will be solved soonish

I took the liberty to dig out your email and take things to the list, hope my guess was right.

Please respect our wish for not using issues and place your next issue on the list. When i see your name again here, no matter how urgent/valid the report will be, i will likely not be in the mood to fix it.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.