[00:14] <Psi-Jack> Hmmm.. Now, evaluating KeePassXC and iOS options, if any exist yet that are viable.
[00:17] <pragmaticenigma> To my knowledge XC is uses the same file format as KeePassX (kdbx format) ... so I should think any are potential candidates... no?
[00:17] <Psi-Jack> They do use that format, newer versions of it. And there are /some/ options. Not always that good, specifically on the IOS side.
[00:18] <Psi-Jack> like, there's Strongbox now. I don't recall that being a thing a few years ago.
[00:18] <Psi-Jack> But, you have to pay for some features of it, heavily, if you want it. Like syncing.,
[00:18] <Psi-Jack> By heavily, I mean up to ~$38 for a lifetime license.
[00:19] <sixwheeledbeast> KeePassium?
[00:20] <Psi-Jack> Hadn't tried that one yet.
[00:27] <Psi-Jack> Heh, well, my old KeePassXC database still works. :)
[00:28] <Psi-Jack> keepassium also has a lifetime license, but $27.
[00:38] <Psi-Jack> Doesn't support TOTP though.
[00:38] <Psi-Jack> Strongbox does. :)
[00:46] <ducasse> i wish my phone supported nfc so i could use my yubikey to unlock my keepass db
[00:51] <Psi-Jack> heh
[00:53] <Psi-Jack> Wellll.. Hmmmm..
[00:57] <Psi-Jack> This approach is indeed, working so far... I think the main issue I had before was form fields not always being reliable.
[01:00] <pragmaticenigma> cause Cook and Jobs know better... you don't need form fields :-)
[01:00] <pragmaticenigma> you just need a single button and an imagination
[01:00] <Psi-Jack> Well I wouldn't mind SQRL being more popular. But I get it's so very new, still. :)
[01:01] <pragmaticenigma> Oh... the gibson and twit... if it never gets promoted anywhere... it's not really going to take off I fear
[01:03] <Psi-Jack> Well, they need apps, first. Cross platform all over./
[01:03] <Psi-Jack> And not just in beta. :)
[01:06] <Psi-Jack> Hmmm yeah..
[01:07] <Psi-Jack> KeePassXC still failing form fields. In this easy example, AWS IAM login.
[01:11] <Psi-Jack> Oh boo! And the Alt+Shift+T they suggest is to fill in the TOTP, just opens the Tools menu option of Firefox. :/
[01:11] <Psi-Jack> Looking more and more like Enpass, for me, is still the most viable solution.
[01:11] <Psi-Jack> Lacks a CLI tool, but with some work I could get that working again.
[04:15] <lotuspsychje> good morning
[04:38] <lotuspsychje> !info stacer
[05:03] <lotuspsychje> nt0: dell has mostly good ubuntu support
[05:03] <lotuspsychje> nt0: but every machine has hardware components inside, that possible 'could' trigger a bug on a specific kernel
[05:04] <nt0> aye--my worry awhile back was that my laptop didn't ship with an intel wifi chipset as in the developer editions (but i'm not really concerned about that)
[05:04] <lotuspsychje> iwlwifi suffers some bugs lately on 5.0 and 5.3 kernels
[05:04] <lotuspsychje> realtek can also be a pain, but that doesnt mean things cant be solved right
[05:05] <nt0> i'm not an expert but i'm pretty savvy with this sorta stuff--patched the source for a relatively new (at the time) usb-wifi card to enable monitor mode a year or so ago.  that's my proudest HW moment.
[05:06] <nt0> 5+ kernels have regressions in iwlwifi?
[05:07] <lotuspsychje> some intel chipsets do yeah
[05:07] <nt0> i use gentoo on my main box (this one) but use ubuntu in a VM on my 9560 (which i continue to allow evil MS to exist on out of my own laziness) and recently helped a friend out with a fresh 19.10 install
[05:08] <nt0> tbh 19.10 was snappy/clean enough to cause me to think about using ubuntu again (i had forsaken it for various reasons) and now that i've read that the amazon cruft will no longer be default i'm back on board
[05:09] <nt0> bummer re: regressions.  i'll figure out which one i actually have and look into it.  thank you
[05:10] <lotuspsychje> i have both intel & realtek working on 20.04 here
[05:10] <lotuspsychje> kernel 5.4 currently
[05:12] <Psi-Jack> Hmmm.
[05:13] <Psi-Jack> I'd noticed that Fedora 31 had no iwl issues. Just Intel HDA issues on 5.5.5
[05:18] <nt0> pure curiosity: what causes bugs in HDA?  isn't that pretty much set in stone as a standard?
[05:19] <nt0> i can imagine hardware that might be on one side or the other of some expectation set by software
[05:20] <nt0> iirc HDA has been around for a long, long time now
[05:23] <lotuspsychje> i havent seen alot of audio bugs lately
[05:59] <Psi-Jack> Yeah. with many varations and bugs. LO
[07:07] <lordievader> Good morning
[07:41] <ducasse> good morning
[11:04] <code1o6> Hi everyone, I followed this tutorial https://help.ubuntu.com/community/LiveCDCustomization . Unfortunately, my resulting iso doesn't have a partition table. I'm able to boot my image from KVM but not from a usb
[11:06] <code1o6> the only step I didn't follow was cdrecord as I used dd instead.
[11:10] <code1o6> is this the right place to ask?
[11:18] <daftykins> not really as, given the topic, it's not a support channel - #ubuntu or #ubuntu-server might be some use
[17:15] <pragmaticenigma> giaco: dragonriver is already here too
[17:15] <dragonriver> giaco, a lot of modern hardware has 4k physical sectors, which it presents as 512 byte logical sectors. By default, old luks format will set the sector to 512 bytes, so do 4 calls when only 1 is needed. You get much better performance.
[17:16] <pragmaticenigma> dragonriver: Do you mean the block would have 4k sectors? or is that per track on the drive?
[17:16] <pragmaticenigma> how does the topology work?
[17:18] <giaco> dragonriver: thanks for the info, but fdisk print shows 512 for all my disks. I'll keep the info for the future
[17:20] <dragonriver> pragmaticenigma, just the sector size on the disk, it doesn't have to do with the block size. I'm trying to find links, but can't right now.
[17:20] <daftykins> 'Advanced Format'
[17:23] <pragmaticenigma> so a sector is still a sector, now I see how you get the multiple trips then
[17:24] <dragonriver> Yes. Most disks have hybrid advanced format 4k. It allows the disk to operate as having real sector size of 512 bytes, but in reality the sectors are 4k. In reality, the speedup isn't 4x, because the hdd controller is smarter than that, but it does have 4 times writing or checking the invisible crc that is on the 4k sector (invisible because that's handled internally by the hdd, not available to the OS.
[17:25] <dragonriver> The only real difference between 512 and 4096 byte sector size is lower overhead on the error correction on the sector, which allows hdds to have slightly higher density.
[17:26] <pragmaticenigma> do most OSes detect this and leverage it? or is that something the user needs to configure on format to properly leverage?
[17:26] <daftykins> besides partition sector alignment, it's the controller's job
[17:28] <giaco> so it could be that the disk has sector 4k but it shows as 512 because the controller says that?
[17:28] <giaco> for controller, I mean the usb disk envelope
[17:28] <pragmaticenigma> If Wikipedia is to be believed: For hard disk drives working in the 4K native mode, there is no emulation layer in place, and the disk media directly exposes its 4096, 4112, 4160, or 4224-byte physical sector size to the system firmware and operating system.
[17:28] <dragonriver> pragmaticenigma, virtually no OS detects this, and most time it's impossible to detect. Some HP hardware does expose this information (not consumer, I'm talking serious server hardware).
[17:30] <pragmaticenigma> It sounds like I'd need to know the drive has this, and configure it when I go to format the disk?
[17:32] <dragonriver> Oh. You can see the logical and physical sizes no problem. For example "Sector size (logical/physical): 512 bytes / 4096 bytes". What's hidden is the crc overhead that the sector has internally.
[17:39] <pragmaticenigma> How can "see" it then?
[17:42] <dragonriver> for example, fdisk -l
[17:44] <pragmaticenigma> giaco: Cool that you're taking steps on encrypting data at rest... do you have any concerns about the performance of the system though? with those being encrypted
[17:44] <pragmaticenigma> ah, okay... simple as that then
[17:44] <pragmaticenigma> thanks dragonriver
[17:51] <dragonriver> In my experience encrypting the entire drive has very very low performance penalty, if your cpu has the aes instruction set. You can always run a benchmark with "cryptsetup benchmark" to see what type of penalty you can expect.
[18:05] <pragmaticenigma> one of mine has a performance hit... but it has more to do with not enough RAM and the choice in drive encryption
[18:16] <dragonriver> You will definitely feel a performance hit, especially for things like SQL workloads. But whole disk luks is less bad than fscrypt. See for example https://www.phoronix.com/scan.php?page=article&item=ext4-crypto-418&num=3
[18:17] <dragonriver> Notice that for workloads like compilation and nginx page serving the penalty is around 0%.
[18:19] <pragmaticenigma> yeah... this setup is a windows laptop, i7 gen 2 or 3, with 8 GB ram, using veracrypt... it works well enough, but it's going to get wiped soon, and I'll move to having an encrypted partition, and leave the OS unencrypted... first issue is veracrypt isn't steller, but that's all that I have presently found for the Windows side of the world
[18:21] <dragonriver> veracrypt is very, very strong. You can setup deniable containers in files, the whole lot.
[18:22] <pragmaticenigma> yes, but FDE just doesn't seem like it is fully there
[18:23] <pragmaticenigma> As soon as anything starts swapping memory, I get a system crash... might be able to mitigate it by creating a partition for the windows swap file
[18:23] <pragmaticenigma> might be another option
[18:23] <dragonriver> I remember you can. You could with the old one.
[19:58] <pragmaticenigma> donofrio: please provide complete answers in the future. that was most unhelpful and confusing for the person that asked for help