Product: Fedora https://bugzilla.redhat.com/show_bug.cgi?id=904843 --- Comment #3 from Michael Schwendt <mschwendt@xxxxxxxxx> --- > http://fedorapeople.org/~ahs3/acpica-tools-20130123-1.f18.src.rpm 404 not found. f18 => fc18! ;) > obsolete tag is needed It's there (below the BuildRequires in the spec file), but it's too low for the latest Fedora package: > Obsoletes: iasl <= 20120913-6 Compare with: yum list iasl Also, since it includes %{_bindir}/iasl and shall replace package "iasl", a versioned "Provides" for iasl ought to be added. > further, the pmtools package -- which provides acpidump -- also provides > a /usr/sbin/acpixtract that we don't really want to collide with You do collide currently, however, because both builds of acpixtract are in $PATH. For normal users: $ rpm -qf $(which acpixtract) file /usr/bin/acpixtract is not owned by any package $ file $(which acpixtract) /usr/bin/acpixtract: symbolic link to `/etc/alternatives/acpixtract' For root: # rpm -qf $(which acpixtract) pmtools-20100513-3.fc18.x86_64 /sbin is before /usr/bin in $PATH for root, and vice versa for ordinary users. > NB: this package does not use the %{optflags} macro > Rationale: upstream claims that using -O will lead to miscompilation > and the resulting tools will be incorrect. Since ACPI is > a reasonably critical part of the environment, we are erring > on the side of caution. Furthermore, it is not important that > these tools operate more quickly than they do. Their present > performance level is sufficient. A few thoughts here: 1) Since I haven't checked whether the source code uses plain C only or also machine/assembler language, does the claim of miscompilation refer to the C source files? Does the test-suite discover the miscompilation? Given the fact that many thousands of packages are built with Fedora's optflags, miscompilation for this particular software could be due to a questionable programming style (such as dubious/unsafe assumptions about memory layout and e.g. casts). It could be enlightening to track down where it breaks, especially if this code is supposed to be a reference implementation. 2) Has this ever been reported to the GNU compiler developers? Or even Red Hat's compiler maintainers? 3) %optflags are not just for performance. It's also security related and helps locating bugs, too: $ rpm --eval %optflags -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic Could they be used with -O2 filtered out? https://fedoraproject.org/wiki/Packaging:Guidelines#Compiler_flags > %{_mandir}/man1/iasl.1.gz %{_mandir}/man1/iasl.1* is the more flexible wildcard for including manual pages, as it allows for the compression method to be disabled/reconfigured. > - NB: ACPICA documentation is not clearly redistributable so not included Apparently, this %changelog comment doesn't refer to files included within the src.rpm, does it? I find the comment confusing and ask about it, because if the src.rpm included the docs, they would need to be deleted from it, too. -- You are receiving this mail because: You are on the CC list for the bug. Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=JTRy1wnbV9&a=cc_unsubscribe _______________________________________________ package-review mailing list package-review@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/package-review