Hide Forgot
Created attachment 1192150 [details] Patch for postgresql-setup.in to handle redhat_seckets_hack correct. Description of problem: There is an error in /usr/bin/postgresql-setup when it decides whether "upgrading from redhat server". The value of the parameter is ignored, it can be "yes" or "no", but the script checks only if the parameter is not empty. I provide a patch that solves the problem. Version-Release number of selected component (if applicable): postgresql-9.5.4-1.fc24.x86_64 postgresql-server-9.5.4-1.fc24.x86_64 postgresql-upgrade-9.5.4-1.fc24.x86_64 How reproducible: Always
You are right again, thanks for the patch, I'll apply this. Misinterpreting the yes/no value should not cause big troubles as the exported shell variable is used only by patched PostgreSQL in RHSCL. I'm curious, have you been affected by this issue?
(In reply to Pavel Raiskup from comment #1) > I'm curious, have you been affected by this issue? No, I'm not affected. I was searching for the reason why postgresql-setup has failed, and one try was to correct this. But this did not solve the failed script. Later I found the reason described in bug #1368477. But I had already a patch for the wrong code, so I thought it is worth to report it so it can be corrected for other users.
Applied upstream: https://github.com/devexp-db/postgresql-setup/commit/82abbb0ec7d19
As Edgar is not affected and this is fixed in Rawhide (setup_version 5.0), I'm closing this bugreport. Thanks again!