=== guiverc2 is now known as guiverc | ||
adrien | \o | 08:18 |
---|---|---|
adrien | is there a way to get a serialized output from apparmor profiles? I'd like to see which paths are covered (I don't need exact data, only approximate) | 08:32 |
adrien | I've see -p to only preprocess profiles which might be enough to grep with the paths that I am interested in but I'm wondering if there's a cleaner way | 08:33 |
adrien | I finally had another look at erlang/rabbitmq-server and the TLS configuration appears to be simply forwarded to erlang/tls without going through code which means that any code change would have to be done in erlang instead and I'm not proficient in erlang | 11:41 |
tsimonq2 | adrien: jjohansen or sarnold may know the answer to your AppArmor question. That's a good one, let me know what you find. | 18:32 |
jjohansen | adrien: so currently apparmor_parser -p is probably your best bet, there is wip to be able to expose more controls around the profile dump (different levels of processing) but nothing that is available atm | 18:35 |
jjohansen | at least not cleanly, there are the -D flags, but their dumps can change (its not a fixed format) and they are fairly messy | 18:36 |
jjohansen | apparmor_parser --help=D | 18:37 |
jjohansen | will give the full list supported by a give parser | 18:37 |
jjohansen | apparmor_parser -D rule-exprs would be the next closest thing to -p, it late stage dumping what the rules get converted to before the statemachine build | 18:41 |
jjohansen | it comes after includes, variable expansion, rule dedup, etc. It has a certain order to it, but exactly what depends on the dedup phase and that has changed over time, and not all rule types have always supported dedup, etc. | 18:41 |
jjohansen | and a single rule might expand into multiple rule-exprs | 18:41 |
jjohansen | so it probably not what you are looking for | 18:41 |
georgiag | --debug might be helpful as well | 18:45 |
jjohansen | but as with dump, its messy and not a fixed format | 18:46 |
jjohansen | debug also can vary based on how the parser is compiled, as some of the output will only happen when DEBUG is set at build time | 18:47 |
jjohansen | but most it doesn't | 18:47 |
adrien | thanks, these all look interesting; I'm not sure which one I'd settle on right now and I guess I'll know when I start using that at scale | 19:02 |
adrien | for context, I am interested in finding out which packages ship server programs which communicated over TCP and do TLS, and maybe other protocols | 19:03 |
adrien | I have some heuristics but more heuristics is better (and I'll also expand the scope but as you can imagine, there are already quite a few package that match this) | 19:03 |
plugwash | hi, the autopkgtest documentation mentions a file called big_packages, but it doesn't say anything about how to view said file or request changes. Can anyone elaborate? | 20:15 |
plugwash | ok, I just reread the page and saw the git repo, still leaves the question of how to request changes. | 20:18 |
vorlon | plugwash: MP | 20:21 |
plugwash | do you mean MR? if so where? (the repo i've found doesn't appear to be on a gitlab instance) | 20:23 |
vorlon | merge proposal; the repository is in launchpad, are you familiar with launchpad code hosting? | 20:23 |
vorlon | plugwash https://code.launchpad.net/~ubuntu-release/autopkgtest-cloud has some info at the top about how this works | 20:26 |
plugwash | hmm, I've cloned the repo, but I can't find big_packages, looking at the cgit url it looks like it's on it's own branch, but I don't seem to have that branch in my clone | 20:38 |
vorlon | oh sorry one sec | 20:38 |
vorlon | that's simply the wrong repo I mangled urls when trying to translate | 20:39 |
vorlon | https://code.launchpad.net/~ubuntu-release/autopkgtest-cloud/+git/autopkgtest-package-configs | 20:39 |
plugwash | ok I've made and pushed a commit, how do I turn that into a merge proposal? | 20:55 |
bdmurray | when you did the push you should have seen a url for creating an MP | 20:57 |
bdmurray | https://code.launchpad.net/~plugwash/autopkgtest-cloud/+git/autopkgtest-cloud/+ref/main or there is a propose for merge link there | 20:58 |
bdmurray | In the MP it would be helpful to point a failure log | 20:58 |
bdmurray | plugwash: Do you suspect that there are more packages that might need moving to big_packages? | 21:47 |
plugwash | I would not be at all surprised but i'm not going to read every log to find out. | 21:51 |
bdmurray | plugwash: there is a tool called retry-autopkgtest-regressions that aids in searching log files but I'll do that | 21:52 |
rbasak | plugwash: hello! Long time! Thank you for your help! | 22:17 |
bdmurray | liushuyu: what did you do to test the regression potential of gdb e.g. "may break ARM64 execution"? | 22:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!