Date: Tue, 9 Jun 2020 20:09:21 -0700 From: Donald Wilde <dwilde1@gmail.com> To: freebsd-stable <freebsd-stable@freebsd.org> Subject: question on porting Message-ID: <CAEC73934akXpUU50Z8WDS0nakQ5Dy73FhTGUh8vpKv%2B8r0ArEg@mail.gmail.com>
next in thread | raw e-mail | index | archive | help
I am running into different problems with 'synth upgrade-system' now, after rebuilding my system yet again to 12-STABLE status. My development mule is an i3 with 4GB of RAM. Specifically, my synth operation ran swap into the ground several times while it was attempting to rebuild both llvm80 and gcc9 at the same time. This caused two failures and over 150 skipped ports (one more failure and 21 more skips happened for a known reason). I think in several previous synth ops this same thing caused a bug to occur that trashed my disk and I had to reinstall from scratch. The fact that synth continues suggests to me that the swap fault is expected behavior, but I suspect that whatever trashes the disk is an un-accounted-for bug. Both gcc9 and llvm80 are huge code-bases when you include in all the dependencies. The LLVM project seems to be less GNU-centric, but GCCx is suffering from more and more code bloat, IMHO. I realize that we are talking about a _lot_ of ports, but where do we (and FreeBSD's port maintainers) reach the point of diminishing returns by supporting GCC and other GNU-oriented Linux-isms like libsigsegv? It seems that CLANG supports all flavors of C++ so it is more a question of linkage than compiling? On the specific synth crash, If I re-run it, does synth have code that reorders failed ports such that it has a better chance of not having such swap-space faults/failures happen? -- Don Wilde **************************************************** * What is the Internet of Things but a system * * of systems including humans? * ****************************************************
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAEC73934akXpUU50Z8WDS0nakQ5Dy73FhTGUh8vpKv%2B8r0ArEg>