[07:15:01] *** Quits: gila_ (~gila@ec2-54-91-114-223.compute-1.amazonaws.com) (Ping timeout: 246 seconds) [07:16:39] *** Joins: gila (~gila@5ED4FE92.cm-7-5d.dynamic.ziggo.nl) [08:33:19] sethhowe, drv: build pool appears down [08:41:45] looking into it now [08:46:08] jimharris, wrt your comment on the need for the _P macro in the mock thing, I still don't see an easy way to have one macro that, for the *, also sets up storage for a mock return value like I do in the _P variant [08:46:30] jimharris, but I'm open to suggestions of course... [08:55:00] Build Pool is back online. The pool script was still running, but it appears that something was blocking it from finding new reviews. I will let you know when I find the root cause. [08:55:13] cool - thanks [09:06:23] *** Quits: gila (~gila@5ED4FE92.cm-7-5d.dynamic.ziggo.nl) (Quit: My Mac Pro has gone to sleep. ZZZzzz…) [09:38:08] peluse: pulled your patch down and played around with it and now I understand why you need it :) [09:38:20] I'll finish up a review on the full patch here shortly [09:40:24] jimharris, gracias [09:53:23] peluse: posted [09:55:37] thanks [10:04:02] jimharris, only 2 comments that I'd rather not implement. Take a look, let me know if you agree after reading the explanation and then I'll go make the changes [10:36:01] peluse: responses posted [10:38:20] jimharris, thanks. OK, I'll do both and find a happy medium on the verbosity in the comments :) [10:44:23] the GET_STUB macro I think will simplify the blob and blofs tests immensely [10:45:32] here's the patch I was referring to where I added a ton of comments in the unit tests - I definitely do not seeing this be the normal case but hopefully explains my responses to your review :) https://review.gerrithub.io/#/c/367611/6 [10:46:00] sorry - that's a pointer to an old review [10:46:04] https://review.gerrithub.io/#/c/367611/ is the current [11:15:55] jimharris, cool, will check it out in a bit [11:29:33] jimharris, one more reply back on the mock thing when you get a sec... [11:38:52] otherwise its ready to go [11:48:13] oops, forgot to hit post earlier. Just did [11:55:45] yeah - not sure what I was thinking [11:55:51] ignore that comment [11:56:08] jimharris, LOL but it did lead to me removing that stack variable :) [11:58:57] also started looking at detecting c++ style comments, doesn't look like astyle does it but we could add something to check_format script - there are a bunch of files with them in it and we'd have to ignore some of the stuff in the JSON parrser [11:59:03] that isn't a comment... [11:59:29] maybe after a beer on Jul 4 :) [12:08:19] *** Joins: gila (~gila@5ED4FE92.cm-7-5d.dynamic.ziggo.nl) [14:43:54] this rdma/nvmf threading stuff is pretty tricky [14:44:25] I'm probably going to have to write a novel explaining how it works [14:44:53] but I can't find a way to make it simpler than what I'm doing and still support 1) multiple transports, 2) multiple NICs within a transport [14:45:17] future NYT #1 bestseller: "SPDK NVMe-oF threading model improvements" [14:46:20] I'll probably have to settle for a conference talk [14:53:41] drv, thanks for the comments on the mock patch, its been updated [14:54:00] cool [14:54:02] hell I'd buy a copy! [15:38:55] jimharris: odd multi-process failure here, seems to have ignored our --base-virtaddr option and mapped the shared memory region somewhere else [15:38:58] http://spdk.intel.com/public/spdk/builds/review/27d6944dddd2a63f41a94d1e37415976bc293262.1498861819/vm-fedora-01/build.log [15:45:19] *** Quits: gila (~gila@5ED4FE92.cm-7-5d.dynamic.ziggo.nl) (Quit: My Mac Pro has gone to sleep. ZZZzzz…) [15:52:07] drv, haven't seen that one before... guess that why they call that an "option" :) [15:52:40] I'm not a big fan of the R in ASLR :) [15:54:31] c/mon, security is here to benefit all of us! [15:55:24] I mean, the security part is good, just not the "tests aren't reproducible" part