Summary: | [MAINTAINER] security/libgcrypt: Fix crash on amd64 | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | Ports & Packages | Reporter: | Carlos J Puga Medina <cjpm> | ||||||||
Component: | Individual Port(s) | Assignee: | freebsd-ports-bugs (Nobody) <ports-bugs> | ||||||||
Status: | Closed FIXED | ||||||||||
Severity: | Affects Some People | CC: | koobs, marino, pi | ||||||||
Priority: | --- | Keywords: | patch, patch-ready | ||||||||
Version: | Latest | ||||||||||
Hardware: | amd64 | ||||||||||
OS: | Any | ||||||||||
Bug Depends on: | |||||||||||
Bug Blocks: | 196674 | ||||||||||
Attachments: |
|
Description
Carlos J Puga Medina
2015-01-28 01:21:26 UTC
Thanks Carlos! Please include QA output as attachments for portlint and poudriere please (In reply to Kubilay Kocak from comment #1) % portlint -ac WARN: Consider to set DEVELOPER=yes in /etc/make.conf 0 fatal errors and 1 warning found. Build logs via poudriere: https://fbsd.es/~cpm/freebsd/poudriere/freebsd:8:x86:32/libgcrypt-1.6.2_1.log https://fbsd.es/~cpm/freebsd/poudriere/freebsd:9:x86:32/libgcrypt-1.6.2_1.log https://fbsd.es/~cpm/freebsd/poudriere/freebsd:10:x86:32/libgcrypt-1.6.2_1.log Created attachment 152313 [details]
libgcrypt-1.6.2_1.zip
At last, all build logs via poudriere on 8.4/amd64, 9.3/amd64 and 11/amd64 https://fbsd.es/~cpm/freebsd/poudriere/freebsd:8:x86:64/2015-01-31_22h43m38s/build.html https://fbsd.es/~cpm/freebsd/poudriere/freebsd:9:x86:64/2015-01-31_22h45m23s/build.html https://fbsd.es/~cpm/freebsd/poudriere/freebsd:11:x86:64/2015-01-31_22h54m30s/build.html Just to clarify, "INSTALL_TARGET= install-strip" doesn't work, so that's why you replaced it with a post-install target that does? (In reply to John Marino from comment #5) Yes, I changed it because using ${STRIP_CMD} works fine. I'm not sure why "INSTALL_TARGET= install-strip" doesn't work here. I think you mean you changed it *because* INSTALL_TARGET doesn't work (because INSTALL_TARGET is preferred). The most likely reason is that install-strip isn't a valid target on that makefile. The rest looks okay to me, so I'll promote this PR. Regarding to ${COMPILER_TYPE} variable. Probably, it will be necessary to add this variable in /etc/make.conf instead declare it in libgcrypt/Makefile. Are you agree? (In reply to Carlos Jacobo Puga Medina from comment #8) I mean to add a new entry in /usr/ports/UPDATING. nope. (as documented in private email) Created attachment 152642 [details]
libgcrypt-1.6.2_1.diff
Patch modified according to Mk/Uses/compiler.mk
# Fix crash at cipher/salsa20.c module on amd64
.if ${ARCH} == "amd64" && exists(/usr/bin/clang)
CFLAGS:= ${CFLAGS:N-O*} -O2
.endif
A commit references this bug: Author: pi Date: Sat Feb 7 07:25:23 UTC 2015 New revision: 378572 URL: https://svnweb.freebsd.org/changeset/ports/378572 Log: security/libgcrypt: fix some issues - Fix crash at cipher/salsa20.c module on amd64. - Fix strip shared object. - Bump PORTREVISION. PR: 197146 Submitted by: cpm@fbsd.es (maintainer) Changes: head/security/libgcrypt/Makefile head/security/libgcrypt/files/patch-cipher-Makefile.in Committed, thanks. |