Bug#730216: please consider allowing pypy to build on machines with less ram.
Hi Stefano! On 02/09/2015 08:07 PM, Stefano Rivera wrote: Hi John (2015.02.09_00:36:31_+0200) Please let the decision whether a certain package should be built on a certain architecture up to the porters. OK, you make persuasive arguments. I'll remove it. Thank you very much! I appreciate that. At the time this check was implemented, many of Debian's architectures had a wide range of RAM in buildds. There are far fewer "small" buildds in the supported architectures these days. I think by now, most architecture that do have some small buildds have blacklisted pypy on them, so I am not too worried about blocking up the build queue. Well, it isn't really an issue either way. The m68k buildds have 768 MiB RAM at most, for example. However, we can use very fast swap areas (RAM disk boards etc) such that the amount of RAM isn't much of an issue. I was more concerned about wasting time architectures with varying RAM in buildds than archs that only have small buildds, where it'd be unlikely to build anyway. But I can see why this would be an issue on those archs. Yeah, I understand. But normally we will just blacklist the package per arch or build it manually so that isn't really an issue. Thanks a lot for your help! Adrian -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#730216: please consider allowing pypy to build on machines with less ram.
Hi John (2015.02.09_00:36:31_+0200) > Please let the decision whether a certain package should be built on > a certain architecture up to the porters. OK, you make persuasive arguments. I'll remove it. At the time this check was implemented, many of Debian's architectures had a wide range of RAM in buildds. There are far fewer "small" buildds in the supported architectures these days. I think by now, most architecture that do have some small buildds have blacklisted pypy on them, so I am not too worried about blocking up the build queue. > They (we) are in a much better decision to decide that and if we > actually don't want a package to be built at all on a certain > architecture, we just set it to "Not-For-Us" in the wanna-build > database on the buildd master. There is no need for a package > maintainer to influence that. I was more concerned about wasting time architectures with varying RAM in buildds than archs that only have small buildds, where it'd be unlikely to build anyway. But I can see why this would be an issue on those archs. SR -- Stefano Rivera http://tumbleweed.org.za/ +1 415 683 3272 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#730216: please consider allowing pypy to build on machines with less ram.
Hello Stefano! > Is that not a decision for those porters to make? What's the > disadvantage of letting it continue to try to build there? Exactly, it should be up to the porters whether we want a particular package to build on a certain target architecture or not and the package maintainer should not interfere with that unless there is an actual reason (e.g. the built package would be in any way harmful or problematic). > It seems fairly harmless, to me. After installing the build-deps the > package realises there isn't enough RAM, and aborts, wasting only a > few minutes of buildd time. Depending on the target architecture, installing the build dependencies of a package together with the preceding run of "apt-get update" can take a considerable amount of the time share of the total build time, so it does actually harm in a sense by wasting build time, yes. Please let the decision whether a certain package should be built on a certain architecture up to the porters. They (we) are in a much better decision to decide that and if we actually don't want a package to be built at all on a certain architecture, we just set it to "Not-For-Us" in the wanna-build database on the buildd master. There is no need for a package maintainer to influence that. Thanks, Adrian Porter for m68k and sh4 -- .''`. John Paul Adrian Glaubitz : :' : Debian Developer - glaub...@debian.org `. `' Freie Universitaet Berlin - glaub...@physik.fu-berlin.de `-GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#730216: please consider allowing pypy to build on machines with less ram.
Hi peter (2013.11.22_20:36:12_+0200) I assume this was bug was triggered by the thread on pkg-openstack-devel. My reply there hasn't appeared in the archives yet - so I'll repeat the relevant bits here. > I appreciate that grinding swap on buildds is less than ideal but I > still think it is preferable to not having the package on those > architectures at all. Especially as slow architectures are where > having optimised implementations of stuff is most important. Is that not a decision for those porters to make? What's the disadvantage of letting it continue to try to build there? It seems fairly harmless, to me. After installing the build-deps the package realises there isn't enough RAM, and aborts, wasting only a few minutes of buildd time. Ubuntu has armhf buildds that are capable of building PyPy. I personally have a machine that's capable, and if the Debian armhf porters would like me to, I'd be quite happy to do binary uploads for it, when necessary. There is a JIT implementation for ARM, which makes PyPy builds on that platform a useful thing to have (I think). There also seems to be interest in porting the JIT to MIPS. Someone occasionally posts on the list and IRC channel about it - I doubt we'll see anything any time soon, though. SR -- Stefano Rivera http://tumbleweed.org.za/ H: +27 21 461 1230 C: +27 72 419 8559 -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#730216: please consider allowing pypy to build on machines with less ram.
Package: pypy Currently pypy refuses to build on the armhf and mipsel buildds because they have less than 1400 MiB of ram (the most recent build log for mipsel reports 967 MiB, the most recent one for armhf reports 1011 MiB). I appreciate that grinding swap on buildds is less than ideal but I still think it is preferable to not having the package on those architectures at all. Especially as slow architectures are where having optimised implementations of stuff is most important. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org