John Paul Adrian Glaubitz

Results 98 comments of John Paul Adrian Glaubitz

Testsuite still fails: ``` PASS: test/test-hex_string_to_array.sh FAIL: test/test-aircrack-ng-0001.sh FAIL: test/test-aircrack-ng-0002.sh FAIL: test/test-aircrack-ng-0003.sh FAIL: test/test-aircrack-ng-0004.sh PASS: test/test-aircrack-ng-0005.sh FAIL: test/test-aircrack-ng-0006.sh PASS: test/test-aircrack-ng-0007.sh FAIL: test/test-aircrack-ng-0008.sh FAIL: test/test-aircrack-ng-0009.sh PASS: test/test-aircrack-ng-0010.sh PASS: test/test-aircrack-ng-0011.sh PASS: test/test-aircrack-ng-0012.sh...

Here's what I got for the tests in the static build: ``` (sid-sparc64-sbuild2)root@stadler:/build/perl-7Cs41Q/perl-5.36.0/build-static# cd t; TEST_JOBS=1 ../perl harness -v re/reg_mesg.t re/regex_sets.t; cd - re/reg_mesg.t .... ok 1 - /[[=foo=]]/ died...

> Could you build your perl with -g and -DDEBUGGING and then run the tests that throwing exceptions in some kind of debugger so we can see the C backtrace...

Interesting. The testsuite passes on git master. Let me bisect this.

> @glaubitz I suspect this is fixed; if I don't hear from you otherwise by May 19, 2024, we will close this ticket It's fixed with 5.38.x, but still present...

Debian is using 5.38.2 in unstable which is the only distribution available for sparc64. Hence, Debian isn't affected by the bug anymore either.

Any updates on this? Missing LoongArch support is currently blocking packages in Debian on loong64: See: https://buildd.debian.org/status/fetch.php?pkg=firebird3.0&arch=loong64&ver=3.0.11.33637.ds4-2&stamp=1693212495&raw=0

_Update as of 2017-10-27: New location for installation images: https://cdimage.debian.org/cdimage/ports/_

I have begun adding the initial bits for linux/sparc64: > https://github.com/glaubitz/go/commit/0eca2c04e3a3707ca9e85f524c7ea1e13a8873eb It's still missing the stuff from src/runtime. I copied the syscall stuff from Debian's golang-golang-x-sys package here: > https://anonscm.debian.org/cgit/pkg-go/packages/golang-golang-x-sys.git/tree/unix...

I'm waiting for James Clarke (@jrtc27) to comment on this. He's done a fair bit of the stuff in x/sys.