More follow-up from the afore mentioned Frostcast featuring Manjaro developer Philip Müller. Just past the 16 minute mark.
We learn and everyone makes mistakes. And the new server change every package is new synced from Arch Linux so there is no security issues. … We sync daily so if there is any problems with our system it’s ninety percent from Arch itself, so I don’t know why they bash us.
I am not going to claim Arch is the bastion of all things security – in fact I know Arch is far from perfect here – but Manjaro claiming that they are on par with Arch is wrong. Saying “we sync daily” is frankly deceptive. The daily syncs are to the Manjaro unstable branch, so packages can take a while to reach the stable branch where the vast majority of users get the package. As I have pointed out previously, Arch does not separate out security updates from plain upstream updates, so when Manjaro holds back updates on the unstable branch in the name of stability, they are also holding back security fixes. The updates need monitored for security fixes and either 1) pushed more quickly to the users, or 2) have the fixes backported to the “stable” packages.
But, lets use an example, because facts are good. Recently there was an privilege escalation issue found in polkit. This was made public on 2013-09-18. And over the next couple of days there were a lot of distribution updates to fix this issue. So, I have not picked an obscure bug given the number of distros dealing with the issue, and it is a privilege escalation one (potentially with proof-of-concept available, although I have not checked that out). Lets compare the Arch and Manjaro response to this issue by monitoring the location of the polkit-0.112 package:
I will admit that this is actually better than I thought it would be… I thought packages stayed longer in Manjaro’s testing repositories to catch bugs. Then again, I noticed that there are packages that were pulled into Manjaro from Arch and put into their stable repos within ten minutes, including packages in the [core] repo, so I’ll assume that the testing that occurs in the Unstable and Testing branches is rather limited. (Evidence: pool/ directory with timestamp file was synced from Arch, stable/extra/x86_64/ directory with repo database timestamp.)
In summary, the indiscriminate holding back of all updates in the name of testing(?) is why I “bash” Manjaro security. With this system, Manjaro is always running behind Arch, so claiming the Manjaro security issues are “ninety percent from Arch itself” is full of… optimism. And before the “leave Manjaro alone” comments, I will stop posting about it when I have no need to correct such false statements.