CCing the list since others might like to see where we are at on Debian Alpha.
On Wed, Feb 15, 2017 at 03:43:02AM -0600, Bob Tracy wrote: > The list of held packages was once again starting to get uncomfortably > long, so I took a dive into the swamp to see what was gumming up the > works. There's a new (?) dependency in the various "plasma-*" packages > on "nodejs", and attempting to build "nodejs-4.7.2~dfsg" uncovered a few > issues. OK, I hadn't seen that because been working on kernel and binutils/glibc bugs. If you build your own kernel make sure you revert commit 6cd9dc3e75078ef646076fa63adfb9b85ced0b66, as it leads to random segfaults in user space. [1] > First one is a circular dependency on "node-marked" and "node-yamlish". So there is a dependency loop between nodejs and node-yamlish, thus nodejs will need manual building and uploading. > Next issue is the "-m32" argument getting passed to the compiler. Not > appropriate for Alpha. That's a bug that should be reported to the package maintainer. I've got bigger fish to fry at the moment. In particular a binutils/glibc bug that is causing segfaults in the dynamic symbol resolver. Try this: write a simple "Hello world" program in C. Compile with "-Wl,-z,now" linker option which causes the dynamic loader to resolve all symbols at program invocation, rather than resolving symbols when first used. If compiled with a recent toolchain it segfaults [2]. That is holding up re-building quite a few packages because (recent?) configure scripts use the -Wl,-z,now option when testing to see if the C compiler can generate working executables! A segfaulting executable is not (unsurprisingly) considered a working executable by configure! Cheers Michael. [1] And if you run the Debian kernel you should consider building your own kernel with that commit reverted. [2] Toolchain gcc 4:6.1.1-1, binutils 2.27-8 produces a working executable, but toolchains later than gcc 4:6.2.1-1, binutils 2.27.90.20170124-2 are known to be bad.
