-
nikolamwhat distrubted file system on illumos in general to use? (SmartOS or in VM)
-
gitomat[illumos-gate] 18172 ZFS tests spew false negatives after 17694 -- Gordon Ross <gordon.w.ross⊙gc>
-
gitomat[illumos-gate] 18176 nvme sctostr test fails -- Robert Mustacchi <rm⊙fo>
-
sommerfeldso my predraft IPD 63 has some formatting glitches around double-underscore when my doc is displayed on github. I don't see this in the html generated by the asciidoc CLI python script. anyone bump into this?
-
sommerfeldsource has: Illumos defines `LC_GLOBAL_LOCALE` as `(__global_locale())`
-
sommerfeldand github treats the `__` as "go italic", while the CLI doesn't treat it as special.
-
richlowewhen I was using asciidoc for arm64-gate documentation, I think someone suggested (maybe a random website? maybe someone here?) that I use asciidoc and asciidoctor
-
richloweI can't imagine how parsing _ within ` as special isn't a bug in something though.
-
sommerfeldfound another site on github where they used +++<code>__enter__</code>+++
-
sommerfeldand that appears to have done the trick..
-
richloweI wish I understood why
-
richlowegwr: one thing you could do with tests, though I don't endorse it, is this: 1) a rule `skip_pattern` that takes a list of regular expressions of ests that would be found but should not be run (see also, the ZFS noise, a useful tool for `autotest` too). 2) the general concept of arch-specific rules, I don't know how ConfigParser allows for it, but for eg you accept rules of the form `i86pc.tests =`, `i86pc.timeout = `, and crucially
-
richlowe`i86pc.skip_pattern`. You can then do `aarch64.skip_pattern = [r'\.32$']` for the 32-bit ignorance, and can also tune things like timeouts for platform variability. With extension (the 'arch' value alone is a bit shit), that would help oxide-y environments to othat are not i86pc but are intel.
-
richloweI say "I don't endorse it" in the sense I came up with it in the bath after wherever people last chatted about tests, and have not thought about it further :)
-
richloweit might be great, I'll accept praise :)
-
richloweit would also let me skip known failures easily, where a test circumstantially needs a facility not yet on illumos/arm
-
richlowelike the sleep test using dtrace in its implementation
-
gitomat[illumos-gate] 18163 Fix low-hanging 17694-discovered test bugs -- Dan McDonald <danmcd⊙ei>