Summary: | dev-ruby/http-parser-1.2.3: test failures (on sparc) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Sam James <sam> |
Component: | Current packages | Assignee: | Gentoo Ruby Team <ruby> |
Status: | CONFIRMED --- | ||
Severity: | normal | CC: | matoro_gentoo |
Priority: | Normal | Keywords: | TESTFAILURE |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
build.log
build.log (ppc, 1.2.3) |
Description
Sam James
![]() ![]() ![]() ![]() (In reply to Sam James from comment #0) > 1) HttpParser::Parser#initialize the error should be inspectable > Failure/Error: expect(@inst.error).to > be_kind_of(::HttpParser::Error::INVALID_METHOD) > expected #<HttpParser::Error::UNKNOWN: an unknown error occurred > (UNKNOWN)> to be a kind of HttpParser::Error::INVALID_METHOD > # ./spec/error_spec.rb:20:in `block (2 levels) in <top (required)>' I've just looked at this one assuming that they may all have the underlying cause. What happens here is that the test should trigger HPE_INVALID_METHOD in the http-parser.c code but it doesn't and then falls back to UNKNOWN. The http-parser code in the gem looks to be 2.8.1, so perhaps this is an issue already fixed in net-libs/http-parser. Unfortunately we can't just drop in the .so since the C code gets compiled with ffi-compiler which adds ruby-specific code to make it easy to handle on the ruby side. Upstream also seems to consider moving from http-parser to llparser from llhttp, e.g. https://github.com/httprb/http/pull/639 Not sure if it's worth the time to investigate this further or hope for a new release based on llhttp. Created attachment 787640 [details]
build.log (ppc, 1.2.3)
Same on ppc.
|