09:44:33 Looking closer at https://www.illumos.org/issues/16523 09:44:34 → BUG 16523: int8_t should explicitly be signed char (New) 09:45:13 Now, I haven't hit any problems with applications by making int8_t be explicitly 'signed char', and it does solve a bunch of issues 09:45:31 It does, however, rather spectacularly break the gate build 09:47:10 yeh I've been running with a variant of that for over a year 09:47:59 Haven't had time to look at them all yet, but I get a bunch of signedness errors that seem off 09:48:05 error: pointer targets in passing argument 2 of 'strncpy' differ in signedness [-Werror=pointer-sign] 09:48:38 Where it's storing a string in an int8_t[] and pssing it to something that expects a char 09:50:22 Was using int8_t to store characters a pattern at some point? My guess would be that using char where char is actually meant would fix most of these. 10:06:29 richlowe will have had a load of experience in the arm port too, from the other side where char is unsigned. 10:06:38 So it's likely some of the fixes are already in that branch 11:55:58 Yeah, on quick inspection it looks like they're covered. Upstreaming would be nice, although I'm not entirely sure of the precise attribution for any given hunk in the diffs. 11:57:39 I think what I may do is pick what's necessary to build the next Tribblix release, which will have the modified header, and use that as a testbed 19:09:49 as andy said, I fixed a bunch of int8-as-char stuff 19:09:54 an awful lot in wifi drivers as I recall 19:10:06 it smelled like someone got told "used fixed size integers" and not why, and overreacted. 19:12:49 ptribble: attribution is `git blame`, if it blames to the first commit that credits everyone, credit everyone 19:13:07 and as the readme says, talk to me before upstreaming anything. 19:13:16 (I should update that to say "me, or anyone equivalent to me") 21:09:55 So illumos-discuss is now a single person's stream of consciousness? 21:12:20 you're the 2nd person to have mentioned that today 21:12:53 danmcd: I dunno if this is in your ballpark, since I don't think it's _inappropriate_ 21:13:01 or at least, not with the connotations that word has 21:14:12 I don't actually know how to classify it at all. 21:14:50 the bits about Windows/Crowdstrike/etc seem off-topic, not offensive 21:15:17 It feels like spamming a list. I see no apparent direction or purpose to these series of messages. 21:17:47 It might feel different if anyone else was responding to him, but it seems to be him just responding to himself over and over and over 21:25:06 if it's the same person that was spamming a month or so ago, that was the reason I unsubbed 21:28:47 From https://illumos.topicbox.com/groups/discuss he looks like he's been most of the list traffic since May or so 21:31:36 then yeah, danmcd's the person you want I think 21:32:17 I'm not sure he's around just now 21:32:45 yeah, I'm not either, that's just the extent to which I remember things 22:09:58 it's not yet at a level where it's urgent something be done, just getting noticable 22:36:28 I'm actually dealing with it 22:36:46 I apologise for the noise. 22:37:09 It's definitely off-topic. 22:37:36 jclulow: sorry, was hoping to save you 22:37:45 It's all good 22:39:30 anyone here using illumos on the new(ish) intel cpus with big.LITTLE?