123237 2006-02-18 00:36 0000 cpio is broken on some non-glibc systems due to an unexistent stpcpy function 2006-04-12 10:50:27 0000 1 1 1 Unclassified Gentoo/Alt Other unspecified All Other RESOLVED FIXED P2 normal --- 1 bbj@gentoo.org base-system@gentoo.org bsd@gentoo.org bbj@gentoo.org 2006-02-18 00:36:37 0000 NetBSD has no such stpcpy function. I created a patch to fix this on systems that have no stpcpy. The patch must be applied after all the patches (especially after cpio-2.6-writeOutHeaderBufferOverflow.patch from #112140). bbj@gentoo.org 2006-02-18 00:37:00 0000 Created an attachment (id=80068) cpio-2.6-bsd.patch There goes the patch. arachnist@gmail.com 2006-02-18 00:46:46 0000 Compiled on DragonFly with this patch. Failed without it. reb@gentoo.org 2006-02-18 02:09:16 0000 On OpenBSD too, base-system, please apply this patch. flameeyes@gentoo.org 2006-02-18 03:56:48 0000 Going to commit in overlay in the mean time. Oh it needs an AT_M4DIR="m4" eautoreconf as it changes configure.ac, but it works otherwise. Benigno, did you send it to upstream already? bbj@gentoo.org 2006-02-18 04:32:00 0000 Yes, I just sent him a notification of this patch and the one from #112140, because this patch depends on cpio-2.6-writeOutHeaderBufferOverflow.patch. Thanks for adding it. vapier@gentoo.org 2006-02-18 11:31:56 0000 i dont think that's the right fix at all cpio already has a lib subdir for functions that the host doesnt support, you should be putting stpcpy.c in there like the other stuff and then updating AC_REPLACE_FUNCS flameeyes@gentoo.org 2006-02-18 11:41:32 0000 Uh that's gnulib stuff... it's quite difficult to patch it out, this solution works with the least changes. And it's already fixed upstream in next version. vapier@gentoo.org 2006-02-18 11:56:10 0000 i wasnt talking about the source tarball but about the upstream repo ... the patch posted is not correct for upstream but they've committed a fix upstream that involves pulling in stpcpy from gnulib bbj@gentoo.org 2006-02-18 12:53:01 0000 This replacing is not wrong. If this stpcpy implementation used by gnulib is something like this, I would say that this patch "cpio-2.6-bsd.patch" is better than it: http://cvs.savannah.nongnu.org/viewcvs/bison/bison/lib/Attic/stpcpy.c?view=markup Btw, the problem seems to be finished, and that's very good. :) vapier@gentoo.org 2006-02-18 13:31:49 0000 > This replacing is not wrong. it is wrong for upstream ... but it's a non issue since it's been fixed correctly in upstream cpio cvs vapier@gentoo.org 2006-02-18 13:47:49 0000 added a revised version of this patch to cvs 80068 2006-02-18 00:37 0000 cpio-2.6-bsd.patch cpio-2.6-bsd.patch text/plain LS0tIGNwaW8tMi42Lm9yaWcvY29uZmlndXJlLmFjCTIwMDQtMTItMjAgMDg6NDY6MzAuMDAwMDAw MDAwIC0wMjAwCisrKyBjcGlvLTIuNi9jb25maWd1cmUuYWMJMjAwNi0wMi0xOCAwNDowMToyNC4w MDAwMDAwMDAgLTAyMDAKQEAgLTkyLDcgKzkyLDcgQEAKICAgQUNfQ09ORklHX0xJTktTKHNyYy9m bm1hdGNoLmg6aGVhZGVycy9mbm1hdGNoLmgpCiBmaQogCi1BQ19DSEVDS19GVU5DUyhsY2hvd24g ZW5kcHdlbnQgZW5kZ3JlbnQpCitBQ19DSEVDS19GVU5DUyhsY2hvd24gZW5kcHdlbnQgZW5kZ3Jl bnQgc3RwY3B5KQogQUNfRlVOQ19WUFJJTlRGCiBBQ19GVU5DX0FMTE9DQQogQUNfQ0hFQ0tfTElC KG5zbCwgZ2V0aG9zdG5hbWUsIFtMSUJTPSIkTElCUyAtbG5zbCJdKQotLS0gY3Bpby0yLjYub3Jp Zy9zcmMvY29weW91dC5jCTIwMDYtMDItMTggMDQ6MDg6MTMuMDAwMDAwMDAwIC0wMjAwCisrKyBj cGlvLTIuNi9zcmMvY29weW91dC5jCTIwMDYtMDItMTggMDQ6MDE6MzMuMDAwMDAwMDAwIC0wMjAw CkBAIC0zNTYsNyArMzU2LDE2IEBACiAgIGNoYXIgYXNjaWlfaGVhZGVyWzExMF07CiAgIGNoYXIg KnA7CiAKKyNpZmRlZiBIQVZFX1NUUENQWQogICBwID0gc3RwY3B5IChhc2NpaV9oZWFkZXIsIG1h Z2ljX3N0cmluZyk7CisjZWxzZQorICB7CisgICAgdW5zaWduZWQgbG9uZyBfX2xlbmd0aCA9IHN0 cmxlbiAobWFnaWNfc3RyaW5nKTsKKyAgICBtZW1jcHkgKGFzY2lpX2hlYWRlciwgbWFnaWNfc3Ry aW5nLCBfX2xlbmd0aCk7CisgICAgcCArPSBfX2xlbmd0aDsKKyAgfQorI2VuZGlmIAorICAKICAg dG9fYXNjaWlfb3Jfd2FybiAocCwgZmlsZV9oZHItPmNfaW5vLCA4LCBMR18xNiwKIAkJICAgIGZp bGVfaGRyLT5jX25hbWUsIF8oImlub2RlIG51bWJlciIpKTsKICAgcCArPSA4Owo=