Currently xlockmore ebuild uses pam_stack in its pamd file to mimic system-auth auth class. This is non-portable outside of sys-libs/pam, so won't work on G/FBSD (for example). The attached patch applied over current ebuild, instead of using a pam_stack-based pamd file, creates the pamd file "on the fly" with the pamd_mimic_system function inside pam eclass (still not in tree, see bug #93118), which uses include syntax (which requires sys-libs/pam 0.78 or openpam, so virtual/pam). I've also cleaned the ebuild up a bit, moving the patching into src_unpack, and using econf instead of ./configure. HTH, Diego
Created attachment 59268 [details, diff] Ebuild patch
fixed, and added a version bump while i was at it
this change appears to have broken regular pam authentication on linux as I can no longer unlock my screensaver (using the combo of xautolock and xlockmore).
reopened due to comment #3
Here works fine, you're using sys-libs/pam-0.78 right?
Just to provide more info, I experienced this on an ~x86 system running all of the latest and greatest versions of everything and all of my config files are up to date. However, the odd part here is the logs show no authentication attempts. Let me know if you want me to provide "emerge info" or anything else you'd consider to be useful. Yes I'm using sys-libs/pam-0.78-r2. I just re-emerged it and it appears that some of the configuration files have changed since it was originally installed. However, this caused no change in the behavior of xlockmore for me.
Ok found the problem, and belongs to another bug. The same setup with version 5.12 works (and was the one I tested here because it was the one I used to make the patch) but doesn't on 5.18, so the problem seems to lie there.
What other bug does this problem belong to?
Well a new one to open :P
Well, you basically said "I know what the problem is" but you provide 0 details on what it is. So how can I open a new bug to address the problem if you're not providing me any details on what the real problem is here?
It's not like "i know what the problem is".. the only thing i know is that it's not in this patch :) The problem must lay on something in >=xlockmore-5.13 which changed.. so it should be reported and investigated as another bug (of which I don't know anything, I just dealt with it for the virtual/pam openpam compatibility).
