https://en.wikipedia.org/wiki/Mars_Clim … of_failure
Heads buried in sand. No one seemed to care after the discrepancy was noticed.
]]>Mostly went with the original pkgbuild.
A man went to the doctor complaining of horrible abdominal pain. The doctor, after some examination and tests, asked if the man had been swallowing any non-food items. The man replied 'Oh gosh no, I mostly just eat food.' The doctor diagnosed him as mostly healthy and sent him on his way.
]]>Are you using different build options?
No, they weren't any different. All I did was disable link-time optimization on both devices to see if that might have had an effect. Mostly went with the original pkgbuild.
I'm confused now, since I originally thought the files were created on one computer, then somehow got modified after copying them to another computer.
Originally, they were, then I bit the bullet and tried to compile the packages natively to see if that would resolve the situation.
Oh well. Guess it's just one of those things.
]]>I've come to believe the automatic compiler options - such as link-time optimization - and the processors at hand have a lot to do with what I'm seeing.
That certainly would result in differences if you are using different compile options on each machine (or anything like -march=native which expands to something different on each machine). Are you using different build options?
Still doesn't explain why some files in the packages awaiting compilation appear as identical when others don't, but at least it's half of an answer and that's good enough for me.
If it is good enough, you could mark this as SOLVED, but if you want to figure it out, you should specify some examples of files are or are not identical.
]]>All of the binaries in the util-linux-selinux package (which the scanner readouts first directed me to) are exactly 8 bytes less than they are on the source machine. That made me think of some kind of corruption/compression loss on package transfer - which I then ruled out as this is happening to files belonging to packages not copied via scp. I've also ruled out update schedules not matching up, and even underlying filesystem issues. Strangest part then being, I reinstalled util-linux-selinux from a cached copy, and the md5/sha256sums taken from command line became completely different than even that which was messed up to begin with.
I'm at a loss as to explain what's going on. If there's a good reason for file integrity to just change out of the blue like this would someone please let me know?
]]>