Nobuyoshi Nakada wrote: > If `dbx` can't handle `sigaltstack`, try `configure` with `--enable-debug-env` option and run with environment variable `RUBY_DEBUG=core`. I've tried to reproduce the issue before I've got your note: - fi...TRex58 (Tony Reix)
I've got a different crash in the same place. The .spec file does have the `export LDR_CNTRL=MAXDATA=0x80000000` . ~~~ gmake test .... Generating RDoc documentation ./miniruby -I./lib -I. -I.ext/common ./tool/runruby.rb --extout...TRex58 (Tony Reix)
Nobuyoshi Nakada wrote: > `configure` with `cflags=-qsuppress=1506-434` may work. Yes. That works. However, it is not enough. I know have: ~~~ CC = xlc CFLAGS = -qsuppress=1506-434 ... Run options: "--ruby=./m...TRex58 (Tony Reix)
The issue appears only when compiling with XLC. ~~~ XLC: 1 tests, ** 7** assertions,** 1 failures**, 0 errors, 0 skips GCC: 1 tests, **12** assertions, **0 failures, 0 errors**, 0 skips ~~~ Anyway, how can I get information abou...TRex58 (Tony Reix)
Tony Reix wrote: > Now, I will clean and rerun the test that makes the core appear, **WITH** export LDR_CNTRL=MAXDATA=0x80000000, so that we are sure if it is - or not - related to the memory allocation. I've rerun 3 times the .ext/r...TRex58 (Tony Reix)
Rei Odaira wrote: > I discussed this with Tony using emails, and it turned out that it was due to the default size limit in memory allocation on AIX. Users must set the following environment variable to allocate more than 256MB on AIX. ...TRex58 (Tony Reix)
I had attached all the details in a file. However, I have an "internal error" when trying to look at the file. I haven't put all details in defect because, in some other projects, we are required to not put all details at first, only a ...TRex58 (Tony Reix)
Testing Ruby 2.2.3 on AIX6.1 built with xlc (same with gcc) leads to a reproducible crash. See complete trace in attached file. Ruby version is: RUBY_PROGRAM_VERSION=2.2.3 . ~~~ # gmake V=1 test-all ... objcopy -w -L 'Init_*' -L ...TRex58 (Tony Reix)