On Fri, Aug 4, 2017 at 1:32 PM, Matthias Klose <d...@debian.org> wrote: > On 03.08.2017 12:05, Jeff Law wrote: >> On 08/02/2017 08:06 PM, Ximin Luo wrote: >>> Jeff Law: >>>> On 07/21/2017 10:15 AM, Ximin Luo wrote: >>>>> (Please keep me on CC, I am not subscribed) >>>>> >>>>> >>>>> Proposal >>>>> ======== >>>>> >>>>> This patch series adds a new environment variable BUILD_PATH_PREFIX_MAP. >>>>> When >>>>> this is set, GCC will treat this as extra implicit >>>>> "-fdebug-prefix-map=$value" >>>>> command-line arguments that precede any explicit ones. This makes the >>>>> final >>>>> binary output reproducible, and also hides the unreproducible value (the >>>>> source >>>>> path prefixes) from CFLAGS et. al. which many build tools (understandably) >>>>> embed as-is into their build output. >>>> I'd *really* avoid doing this with magic environment variables. Make it >>>> a first class option to the compiler. Yes, it means projects that want >>>> this behavior have to arrange to pass that flag to their compiler, but >>>> IMHO that's much preferred over environment variables. >>>> >>>> Jeff >>>> >>> >>> Hi Jeff, >>> >>> If by "first class option" you meant a command-line flag, GCC *already has* >>> that (-fdebug-prefix-map) and it wasn't enough to achieve reproducibility >>> in many cases we tested. dpkg-buildflags actually already adds these flags >>> to CFLAGS CXXFLAGS etc on Debian. However, with this patch using the >>> environment variable, we are able to reproduce 1800 more packages out of >>> 26000.Then take what you've done with the environment variable and instead >> implement it on top of a switch. An environment variable is absolutely >> the wrong thing to do here. >> >>> >>> GCC already supports a similar environment variable SOURCE_DATE_EPOCH, >>> which was accepted about 2 years ago in a patch written by one of our GSoC >>> students. We are not planning any more environment variables like this, and >>> are committed to fixing other sources of non-determinism by patching the >>> relevant build scripts. >> I would have rejected that as well :-) One of the few times I would >> have disagreed with Bernd. > > You can argue that gcc has command line options to set these, but most build > systems can be influenced by well documented environment variables like > CFLAGS, > CXXFLAGS, LDFLAGS, so adding one more variable like SOURCE_DATE_EPOCH makes > sense, considering that many tools dealing with timestamps don't even have > command line options to do these (and there it's not just about compilers). > > BUILD_PATH_PREFIX_MAP might be passed as well in CFLAGS / CXXFLAGS, but there > are build systems that even ignore these environment variables, and it's > common > to require the use of environment variables for distributions, like > > http://pkgs.fedoraproject.org/cgit/rpms/gcc.git/tree/gcc7-foffload-default.patch > > If changes like that are not allowed to go upstream, where do you want to keep > these instead, and keeping these in sync across distributions?
In general any distro provides some way to set default parameters for every package. Adding hacks to compiler to work around limitations of particular build system does not sound right. Here's a relevant discussion from binutils ML: https://sourceware.org/ml/binutils/2009-05/msg00086.html -Y
