Migrate ansible 11 -> 12 : bug corrected

[pulse-mind]

No description provided.

[pulse-mind]

Sorry for these bad changes, I tested this one on my deployement

[ricardclau]

This does not change anything I believe

[pulse-mind]

This does not change anything I believe

It changes the behavior, without that it was failing in some cases when I used it on my deploy

[ricardclau]

Can you please provide examples? I cannot see any TBH El dv., 7 de nov. 2025, 12:09, Pulse-Mind ***@***.***> va escriure:

…

*pulse-mind* left a comment (ansistrano/deploy#405) <#405 (comment)> This does not change anything I believe It changes the behavior, without that it was failing in some cases when I used it on my deploy — Reply to this email directly, view it on GitHub <#405 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAEV2XIAB7U3RP5PT2SLUN333R4WZAVCNFSM6AAAAACLNU6LPGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTKMBRHEZDCNJWGQ> . You are receiving this because you commented.Message ID: ***@***.***>

[pulse-mind]

It is not easy for me, I spent 2 days to work on ANXS.postgresql and ansistrano and I shutdown one of my server to test the installation from scratch.
I tried many times from scratch, with the current version and with the version I proposed and there was a difference when retrieving the git repository, the repo was empty.
But it is up to you...

[ricardclau]

Then, unless you (or someone else) can provide an example where this is useful, sorry but we don't want to add any more changes to an already convoluted workflow

[pulse-mind]

you know, I provided the latest modification on this file so I was the one who did the mistake :D

[ricardclau]

Thanks for the PR. I compared it against the current master and also against the state before the original regression introduced in 12806de.

That regression was already addressed by:

• 4a24b39
• 632ef93

At this point, this PR does not seem to fix an additional failing case on current master. The changes in tasks/update-code/git.yml are mostly rewrites of the existing conditions without a functional improvement over the current implementation.

I’m going to close this to avoid extra churn. If there is still a concrete case failing on current master, please open a new PR with that failing scenario or a regression test.