[00:04] <mwhudson> ExecStart=/bin/echo $HOME would probably be enough here?
[00:05] <CarlFK> just that one line?
[00:08] <xnox> mwhudson, well, systemd will substitute that in the unit. I'd rather see something like
[00:08] <xnox> ExecStart=/bin/sh -c 'echo $HOME'
[00:08] <mwhudson> oh right
[00:09] <xnox> or maybe it needs like \$HOME
[00:09] <xnox> ah
[00:09] <xnox> ExecStart=/bin/sh -c 'env | grep HOME'
[00:09] <xnox> muahahahhaha
[00:09] <sarnold> lol
[00:09] <CarlFK> yeah - ans sh is linked to dash, which is different from bash  and $HOME and ~ are surprising me a little too
[00:19] <xnox> CarlFK, that's why i want specifics of how/what works for you in one release, but not the other, and wheather it is sensible or not.
[00:20] <xnox> sarnold, bah
[13:45] <ahasenack> doko: hi, do you remember what happened to your mongodb "No-change rebuild for boost soname change" in cosmic, 1:3.6.3-0ubuntu2? I don't see it in rmadison, nor publishe in https://launchpad.net/ubuntu/+source/mongodb
[13:50] <ahasenack> ah, it failed to build
[13:57] <teward> ah, small question, I should probably know this but, if a package I'm working on ends in a +nmuX version string, would the Ubuntu one be +nmuXubuntuY for an Ubuntu modified version of the package?
[13:58] <teward> there's no documentation of such examples on the security team's documentation for the 'version string' guide so sorry for the stupid question
[14:01] <rbasak> teward: I'm not aware of any convention. I think +nmuXubuntuY would be fine.
[14:02] <teward> rbasak: that's what I thought, and what sarnold thought, interesting though there's no examples in the documentation :p
[14:04] <xnox> teward, i recall hainv a ubuntuX+nmuYubuntuZ
[14:04] <rbasak> My long term plan is to have "git ubuntu lint" expect one (and only one) answer for every input case.
[14:04] <xnox> teward, because somebody uploaded ubuntuX string to debian, then it got nmued, and then we had to patch that again in ubuntu....
[14:04] <rbasak> The code is already there; it's just not integrated yet.
[14:04] <rbasak> https://git.launchpad.net/usd-importer/tree/gitubuntu/versioning.py#n342
[14:04] <rbasak> Please feel free to add additional cases :)
[14:05] <teward> xnox: indeed, that makes sense, though I"m surprised it was never added into the documentation as an example, guess I"ll ask the sec team to update their documentation :p
[14:05] <teward> rbasak: not every package is in git ubuntu last i checked though?
[14:05] <rbasak> And then eventually I'd like to split that code out into some kind of library, and for Ubuntu's official answer to be "what the library does; file a bug against the library where it's wrong"
[14:05] <teward> or has that changed?
[14:05] <rbasak> teward: you're correct
[14:06] <teward> yeah the package in question isn't ;)
[14:06] <rbasak> But the code that works out the next version is fairly independent - just not split out as an API, but it could be.
[14:06] <teward> (iptables-persistent is not there, nor is it a quilt package which made me headscratch for a moment xD)
[14:23] <ahasenack> doko: did you investigate more that mongodb ftbfs in cosmic?
[14:24] <doko> ahasenack: sorry, no
[14:24] <ahasenack> it looks like gcc issues
[14:24] <ahasenack> new warnings, that kind of stuff
[14:24] <doko> cosmic?
[14:25] <ahasenack> yeah
[14:25] <ahasenack> cc1plus: all warnings being treated as errors
[14:25] <ahasenack> https://launchpadlibrarian.net/416133571/buildlog_ubuntu-cosmic-amd64.mongodb_1%3A3.6.3-0ubuntu1.18.10.1_BUILDING.txt.gz
[14:25] <ahasenack> Werror=class-memaccess pops up a lot
[14:27] <doko> ahasenack: and works in disco?
[14:27] <doko> die, -Werror, die
[14:27] <ahasenack> doko: haven't checked, but it's another version in disco
[14:27] <ahasenack> seems to build fine in bionic, same version as cosmic
[14:27] <ahasenack> (ppa still buildinf)
[14:27] <doko> bionic is GCC 7
[14:28] <ahasenack> I remember having to deal with this with squid
[14:28] <ahasenack> which is c++ too