| opendevreview | Ivan Anfimov proposed openstack/openstackdocstheme master: wip https://review.opendev.org/c/openstack/openstackdocstheme/+/980865 | 09:46 |
|---|---|---|
| opendevreview | Ivan Anfimov proposed openstack/openstackdocstheme master: wip https://review.opendev.org/c/openstack/openstackdocstheme/+/980865 | 09:49 |
| opendevreview | Ivan Anfimov proposed openstack/openstackdocstheme master: Highlight the current page in the left menu https://review.opendev.org/c/openstack/openstackdocstheme/+/980865 | 09:49 |
| opendevreview | Ivan Anfimov proposed openstack/openstackdocstheme master: Highlight the current page in the left menu https://review.opendev.org/c/openstack/openstackdocstheme/+/980865 | 09:54 |
| opendevreview | Ivan Anfimov proposed openstack/openstackdocstheme master: Highlight the current page in the left menu https://review.opendev.org/c/openstack/openstackdocstheme/+/980865 | 10:14 |
| opendevreview | LeeChunghwan proposed openstack/osprofiler master: Create root span in the otlp driver https://review.opendev.org/c/openstack/osprofiler/+/968428 | 14:27 |
| dansmith | tkajinam: not trying to pester, but just wanted to remind you about this specific question I directed to you :) https://review.opendev.org/c/openstack/oslo.utils/+/978097/comments/68a98181_0cba9b32 | 16:04 |
| tkajinam | dansmith, let a comment but I generally agree that we should consider splitting that format_inspector.py | 16:22 |
| dansmith | tkajinam: okay, so general split regardless of the LUKS one? | 16:23 |
| dansmith | also, you said "not feasible" to split out the luks stuff, but I think we could have some of those decryption functions in a luks_utils module | 16:23 |
| dansmith | but maybe if we split all the inspectors out it's not necessary | 16:24 |
| dansmith | or maybe you weren't thinking splitting them all out? | 16:24 |
| tkajinam | IMO it's a bit overkilling at this moment to split each inspector to single file, but we can push (1) too large inspector (luks) or (2) a few we can categorize (like vhd + vhdx) to separate files. | 16:28 |
| dansmith | okay | 16:30 |
| tkajinam | regarding luks_util, we can create a separate helpers, maybe we can create one, but I'm not very clear about how we balance 'generic' implementation and alignment with requirements by format inspector | 16:34 |
| tkajinam | and I assume that inspector can't be very much shrunk even if we prepare helpers outside of it | 16:35 |
| dansmith | I think it would be, but okay.. splitting just luks out (for now) will also address it, so that's fine | 16:54 |
| dansmith | I will push up one more copy of what I have now with some feedback addressed and then will iterate on a reshuffle from there | 16:55 |
| dansmith | tkajinam: I don't really understand the concern(s) over the ContainerFileInspector class... to me, there seems to be a fundamental split between things that contain other things and things that are .. uh, just things :) | 16:56 |
| dansmith | if I move that interface to the base class, then I can't add the obligatory inner_safety check with a standardized name, and will have to rely on anything that is going to implement that to do it | 16:57 |
| dansmith | maybe I'm misunderstanding your latest comment.. did you mean keep the intermediate class and raise instead of return None? I don't think that makes sense, | 17:05 |
| dansmith | because we don't know our inner_format until we have parsed enough of our own container header to know | 17:05 |
| dansmith | in the LUKS case, we have to parse the luks header, then grab the first data block, decrypt it, and then inspect.. that's kinda the whole point of doing that.. to be able to see what is inside | 17:05 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!