Bug #14873: Compile failure - MinGW build as of 63768

• Status: Closed
• Assignee: -
• Target version: -
• ruby -v: ruby 2.6.0dev (2018-06-27 trunk 63758) [x64-mingw32]
• Backport: 2.3: UNKNOWN, 2.4: UNKNOWN, 2.5: UNKNOWN

[MSP-Greg (Greg L)]

normalperson (Eric Wong),

The MinGW build (ruby-loco) on 63768 had the following error:

building .ext/include/x64-mingw32/rb_mjit_min_header-2.6.0.h
./miniruby.exe -I../ruby/lib -I. -I.ext/common  ../ruby/tool/transform_mjit_header.rb "x86_64-w64-mingw32-gcc " rb_mjit_header.h .ext/include/x64-mingw32/rb_mjit_min_header-2.6.0.h
../ruby/tool/ifchange "--timestamp=.rbconfig.time" rbconfig.rb rbconfig.tmp
rbconfig.rb updated
generating enc.mk
creating verconf.h
verconf.h updated
../ruby/tool/transform_mjit_header.rb:118:in `block in check_code!': undefined method `success?' for nil:NilClass (NoMethodError)
	from ../ruby/tool/transform_mjit_header.rb:181:in `block in with_code'
	from C:/projects/ruby-loco/src/ruby/lib/tempfile.rb:295:in `open'
	from ../ruby/tool/transform_mjit_header.rb:178:in `with_code'
	from ../ruby/tool/transform_mjit_header.rb:115:in `check_code!'
	from ../ruby/tool/transform_mjit_header.rb:210:in `<main>'
make: *** [uncommon.mk:224: .ext/include/x64-mingw32/rb_mjit_min_header-2.6.0.h] Error 1
make: *** Waiting for unfinished jobs....
==> ERROR: A failure occurred in build().
    Aborting...
Command exited with code 1

Thanks, Greg

[normalperson (Eric Wong)]

[e-mail] wrote:

verconf.h updated
../ruby/tool/transform_mjit_header.rb:118:in block in check_code!': undefined method success?' for nil:NilClass (NoMethodError)

Sorry for the delay. Do you SIGCHLD on that platform?

"pp Signal.list" should show it

I'm not seeing where it's failing to call rb_last_status_set
from rb_waitpid based on recent changes...
(but I had a long and tiring day dealing with stuff offline
and I'm running on fumes at this point)

[MSP-Greg (Greg L)]

@normalperson (Eric Wong)

Do you SIGCHLD on that platform?

Really, not sure. But, I don't believe so. For instance, Puma describes using signaling for server control, and I think all the tests dealing with it are bypassed for windows. I'm really *nix challenged. One of these days...

PS C:\Greg\GitHub> ruby -e "pp Signal.list"
{"EXIT"=>0,
 "INT"=>2,
 "ILL"=>4,
 "ABRT"=>22,
 "FPE"=>8,
 "KILL"=>9,
 "SEGV"=>11,
 "TERM"=>15}

Thanks, Greg

[normalperson (Eric Wong)]

• Status changed from Open to Closed

Applied in changeset trunk|r63785.

---

process.c: fix typo in non-SIGCHLD waitpid :x

I expect this to fix [Bug #14873]

[usa (Usaku NAKAMURA)]

Eric, your recent changes about process.c actually broke Windows build.
r63785 fixed a little, but test-all still reports 14904E.
I recommend to revert all commits from r63754, and restart.

[normalperson (Eric Wong)]

[e-mail] wrote:

Eric, your recent changes about process.c actually broke Windows build.
r63785 fixed a little, but test-all still reports 14904E.
I recommend to revert all commits from r63754, and restart.

Can you try r63790? (Cc: you directly since ruby-core seems very slow)

[usa (Usaku NAKAMURA)]

Hi, Eric

r63790 seems almost OK.
Thx!!

--
[e-mail]