Re: [gentoo-dev] Last rites: */*: More Py2 stuff
On Tue, Jul 28, 2020 at 7:32 PM Aaron Bauman wrote: > On July 28, 2020 9:57:44 PM EDT, Gordon Pettey wrote: > >That dependency is only if USE="-gnuefi". sys-boot/gnu-efi has no > >Python > >dependency. Instead of masking/removing refind, remove the USE flag and > >force the gnu-efi dependency, or reverse the condition, add > >IUSE="tianocore", and mask that USE flag. > > > >On Tue, Jul 28, 2020 at 7:06 PM Aaron Bauman wrote: > > > >> On Tue, Jul 28, 2020 at 04:55:57PM -0700, Matt Turner wrote: > >> > On Tue, Jul 28, 2020 at 4:17 PM Aaron Bauman > >wrote: > >> > > sys-boot/refind > >> > > >> > How did you conclude that this package depends on Python at all? > >> > > >> > >> Hi, Matt. It has a dependency on sys-boot/udk which was masked due to > >> using py2.7 only. Hope that helps. > >> > >> -- > >> Cheers, > >> Aaron > >> > > That is for the maintainer to decide. Hence, all the previous discussions > surrounding this topic. It is a massive undertaking to remove py2.7 from the > tree. You've made a very strong case for filing bugs and asking maintainers to figure out what to do before masking packages for removal.
Re: [gentoo-dev] Last rites: */*: More Py2 stuff
On Tue, 2020-07-28 at 22:32 -0400, Aaron Bauman wrote: > > On July 28, 2020 9:57:44 PM EDT, Gordon Pettey wrote: > > That dependency is only if USE="-gnuefi". sys-boot/gnu-efi has no > > Python > > dependency. Instead of masking/removing refind, remove the USE flag and > > force the gnu-efi dependency, or reverse the condition, add > > IUSE="tianocore", and mask that USE flag. > > > > On Tue, Jul 28, 2020 at 7:06 PM Aaron Bauman wrote: > > > > > On Tue, Jul 28, 2020 at 04:55:57PM -0700, Matt Turner wrote: > > > > On Tue, Jul 28, 2020 at 4:17 PM Aaron Bauman > > wrote: > > > > > sys-boot/refind > > > > > > > > How did you conclude that this package depends on Python at all? > > > > > > > > > > Hi, Matt. It has a dependency on sys-boot/udk which was masked due to > > > using py2.7 only. Hope that helps. > > > > > > -- > > > Cheers, > > > Aaron > > > > > That is for the maintainer to decide. Hence, all the previous discussions > surrounding this topic. It is a massive undertaking to remove py2.7 from the > tree. > It is a massive undertaking to figure out which of the deep dependencies made you conclude to list perfectly maintained Python 3-only packages. -- Best regards, Michał Górny signature.asc Description: This is a digitally signed message part
[gentoo-dev] BugDay - August 1st - Everyone is welcome to join!
# Gentoo BugDay Come join us over at #gentoo-bugday on freenode IRC on the first Saturday of every month to squash bugs and make Gentoo a bit more awesome. You don't need to be a Gentoo developer or even a coder to help us on BugDay. Our next BugDay is on 1st Aug 2020 and we have started making preparations for selecting and prioritizing bug categories for that day. ## Bug categories The bug categories should be broad enough that there will be a lot of bugs being targeted. We keep a option poll open to everybody to help us narrow down the categories of bugs to focus. The opinion poll is there to get an input from everyone about how to best tackle the current bug situation and get an understanding of the community and developer priorities. The poll is open at https://dudle.inf.tu-dresden.de/Bugday_2020-08-01/ Be sure to vote in the poll to get your opinion heard. ## For developers Even if you have never coded for Gentoo you can help us with your knowledge. It's always valuable to have your experience to guide us. Things to help with - Find a related bug that piques your interest. - Look at upstream if this has been reported to them. - If not, make a bug report to the upstream developers. - If they have already seen it, check if they have managed to patch it. - If not, try to gather as much information as you can about the bug so that it may help the developer tackling it. - Alert us at #gentoo-bugday and interact with us to see if this can be squashed. ## For users Users are one of the most important part of Gentoo and this is the occasion for them to talk the developers and make your bugs looked at. Take a look at the categories for BugDay at the poll link and the final BugDay wiki page - Find a related bug that you have experienced and has not been fixed yet - Try to see how it can be reproduced. - The related bug reports have been ignored for months you say? Come poke us about these bugs at #gentoo-bugday on the freenode IRC and we will begin squashing any of those that are pending. ## Whats in it for me? Bragging rights, permanently being listed on the charts of BugDay, sense of entitlement. Any person who helps us solve valid problems will be given the honor of being listed on the page. Even users who help related bugs and find links which make our problem solving easier will be put on a pedestal. ## Contributors Thanks a lot to jstein@ for being the gracious organizer and making sure everything goes smoothly. And special thanks to contributors who have worked on our previous BugDays. Past contributors: - https://wiki.gentoo.org/wiki/Bugday_2020-06-06 - https://wiki.gentoo.org/wiki/Bugday_2020-07-04
Re: [gentoo-dev] systems-246 changes tmpfs default size from 50% to 10% of RAM
On Tue, Jul 28, 2020 at 03:19:26PM -0400, Mike Gilbert wrote: > On Tue, Jul 28, 2020 at 3:13 PM Mike Gilbert wrote: > > > > On Tue, Jul 28, 2020 at 2:50 PM Zoltan Puskas wrote: > > > > > > Hi, > > > > > > I've upgraded to and running systemd-246_rc2 on one of my systems and > > > noticed that tmpfs mounted directories are significantly smaller. > > > > > > This is because with commit > > > https://github.com/systemd/systemd/commit/7d85383edbab73274dc81cc888d884bb01070bc2 > > > they have changed them to be 10% of the physical memory instead of the > > > default of 50%. > > > > > > This is a potentially breaking, or at least an unexpected behaviour > > > change, especially for people using tmpfs on /tmp for compiling. > > > > > > Maybe we should make a news item to let people know that they either > > > need to add an fstab entry with size option set, or better, create a > > > systemd local override with relevant content. > > > > Don't use /tmp for PORTAGE_TMPDIR. /tmp is meant for small temporary > > storage. If you want to compile in a tmpfs, set up a separate mount > > point for it. > > > > I don't intend to create a news item for this, but I would not object > > to someone else doing it. > > Also, the limit for /tmp is likely to change again before the 246 final > release. > > https://github.com/systemd/systemd/pull/16576 > I volunteer to write the news item. It seems other distros also have met the overly restricted /tmp size issue, due to yet another legitimate use case (see the RedHat bug referenced in the PR: https://bugzilla.redhat.com/show_bug.cgi?id=1856514). It's worth noting that above PR only resets only /tmp size, but will keep /dev/shm, /run, etc. at the lower limit. While we have to wait and see what the final form will be for systemd-246, I think it'd be useful to notify users. Systemd changing the age old convention of non configured tmpfs mount sizes (from a user's perspective non configured) it means openrc and systemd boxes will end up with different behaviours (e.g. /dev/shm will now be sized differently). Cheers, Zoltan
Re: [gentoo-dev] systems-246 changes tmpfs default size from 50% to 10% of RAM
Hi, > > Don't use /tmp for PORTAGE_TMPDIR. /tmp is meant for small temporary > storage. If you want to compile in a tmpfs, set up a separate mount > point for it. > I'm not sure I can agree with this. If we are being pedantic then we should look at FHS3.0, and it does not specify that it's intended only for small files, only that the contents are essentially ephemeral, see: https://refspecs.linuxfoundation.org/FHS_3.0/fhs-3.0.html#tmpTemporaryFiles Even if user does not use it for portage, there are other legitimate uses to have a large /tmp where user might want to store large amounts of data for faster processing and also to save a non trivial amount of SSD write cycles. Some real life example are: - Firefox stores temporary downloads there like PDF or archive files, that were selected to be opened with and application instead of permanently saving it, which can pile up if Firefox is not restarted for a longer period of time. - I pointed Hugin (panorama stitcher) to /tmp, which during work produces large amount of temporary data in the form of intermediate files. Probably similar to other multimedia related work flows. - Browsing an archive in mc (MidnightCommander) will result in large amount of data being unpacked into /tmp. - During proxy maintenance when patching a source tree, I will actually untar 2 copies into /tmp and do the patching, test compiling and diffing there. /tmp in convenient (and IMHO intended) for this kind of use, and users probably rely on it without even knowing they do. They should not be required to setup a new tmpfs mount for every use case, especially since many things implicitly assume /tmp is to be used for temporary storage. I don't see why portage cannot fit into this formula, especially since it does not use that much space relative to other potential use cases (well except for compiling LibreOffice, Firefox and friends). In today's world machines with 32-64GB of RAM are readily available, and users probably want to utilize them as much as they can (at least personally I do). Restricting /tmp to small files only does not make sense on desktop and laptop environments for at least a decade now if not more, and is probably even acceptable on home or low traffic servers too. Cheers, Zoltan
Re: [gentoo-dev] Last rites: */*: More Py2 stuff
On July 28, 2020 9:57:44 PM EDT, Gordon Pettey wrote: >That dependency is only if USE="-gnuefi". sys-boot/gnu-efi has no >Python >dependency. Instead of masking/removing refind, remove the USE flag and >force the gnu-efi dependency, or reverse the condition, add >IUSE="tianocore", and mask that USE flag. > >On Tue, Jul 28, 2020 at 7:06 PM Aaron Bauman wrote: > >> On Tue, Jul 28, 2020 at 04:55:57PM -0700, Matt Turner wrote: >> > On Tue, Jul 28, 2020 at 4:17 PM Aaron Bauman >wrote: >> > > sys-boot/refind >> > >> > How did you conclude that this package depends on Python at all? >> > >> >> Hi, Matt. It has a dependency on sys-boot/udk which was masked due to >> using py2.7 only. Hope that helps. >> >> -- >> Cheers, >> Aaron >> That is for the maintainer to decide. Hence, all the previous discussions surrounding this topic. It is a massive undertaking to remove py2.7 from the tree. -- Sent from my Android device with K-9 Mail. Please excuse my brevity.
Re: [gentoo-dev] Last rites: */*: More Py2 stuff
That dependency is only if USE="-gnuefi". sys-boot/gnu-efi has no Python dependency. Instead of masking/removing refind, remove the USE flag and force the gnu-efi dependency, or reverse the condition, add IUSE="tianocore", and mask that USE flag. On Tue, Jul 28, 2020 at 7:06 PM Aaron Bauman wrote: > On Tue, Jul 28, 2020 at 04:55:57PM -0700, Matt Turner wrote: > > On Tue, Jul 28, 2020 at 4:17 PM Aaron Bauman wrote: > > > sys-boot/refind > > > > How did you conclude that this package depends on Python at all? > > > > Hi, Matt. It has a dependency on sys-boot/udk which was masked due to > using py2.7 only. Hope that helps. > > -- > Cheers, > Aaron >
Re: [gentoo-dev] systems-246 changes tmpfs default size from 50% to 10% of RAM
On 7/28/20 2:50 PM, Zoltan Puskas wrote: > Hi, > > I've upgraded to and running systemd-246_rc2 on one of my systems and > noticed that tmpfs mounted directories are significantly smaller. > > This is because with commit > https://github.com/systemd/systemd/commit/7d85383edbab73274dc81cc888d884bb01070bc2 > they have changed them to be 10% of the physical memory instead of the > default of 50%. > > This is a potentially breaking, or at least an unexpected behaviour > change, especially for people using tmpfs on /tmp for compiling. This could also affect any Gentoo admin running a MySQL/MariaDB database with the default settings as it will sometimes write temporary tables to ${EPREFIX}/tmp. Brian signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] Last rites: */*: More Py2 stuff
On Tue, Jul 28, 2020 at 04:55:57PM -0700, Matt Turner wrote: > On Tue, Jul 28, 2020 at 4:17 PM Aaron Bauman wrote: > > sys-boot/refind > > How did you conclude that this package depends on Python at all? > Hi, Matt. It has a dependency on sys-boot/udk which was masked due to using py2.7 only. Hope that helps. -- Cheers, Aaron signature.asc Description: PGP signature
Re: [gentoo-dev] Last rites: */*: More Py2 stuff
On Tue, Jul 28, 2020 at 4:17 PM Aaron Bauman wrote: > sys-boot/refind How did you conclude that this package depends on Python at all?
[gentoo-dev] Last rites: */*: More Py2 stuff
# Aaron Bauman (2020-07-28) # More Py2 only stuff. Plz see -dev ML for discussions # Remove bindings, port to Py3, etc # Removal in 30 days app-accessibility/epos app-admin/conkyforecast app-admin/github-backup-utils app-admin/syslog-summary app-arch/cfv app-arch/ipkg-utils app-backup/bareos app-backup/genbackupdata app-cdr/cdcover app-crypt/openssl-blacklist app-crypt/ssh-multiadd app-misc/mswinurl_launcher app-misc/mtail app-mobilephone/wammu app-office/kexi app-office/lyx app-text/fbless app-text/sgmltools-lite dev-cpp/icnc dev-lang/ispc dev-lang/spark dev-libs/qrosspython dev-python/cddb-py dev-python/flup dev-python/google-apputils dev-python/id3-py dev-python/mox dev-python/pmw dev-python/pyid3lib dev-python/pylzma dev-python/pyode dev-python/pyogg dev-python/pyrex dev-python/python-fchksum dev-python/pythonutils dev-python/pyvorbis dev-python/sphinxtogithub dev-tex/abntex dev-tex/crosstex dev-util/bam dev-util/doxy-coverage dev-util/tailor dev-vcs/cvs2svn dev-vcs/git-bz dev-vcs/gitinspector dev-vcs/gitstats dev-vcs/svnmailer games-action/openclonk games-emulation/fceux games-emulation/m64py games-emulation/mupen64plus games-strategy/0ad media-gfx/alembic media-gfx/cptutils media-gfx/uniconvertor media-libs/ganv media-libs/slv2 media-plugins/vamp-aubio-plugins media-sound/codecgraph media-sound/edna media-sound/exaile media-sound/jack media-sound/moosic media-sound/patchage media-sound/positron net-analyzer/linkchecker net-analyzer/pbgpp net-fs/nfstest net-im/spectrum2 net-irc/quasselgrep net-misc/pssh net-misc/putty net-misc/ris-linux net-nntp/sabnzbd net-print/pkpgcounter net-proxy/hatop net-wireless/gr-baz net-wireless/gr-doa net-wireless/gr-foo net-wireless/gr-ntsc net-wireless/gr-ntsc-rc net-wireless/gr-ppm-wiegand net-wireless/gr-rds net-wireless/gr-rftap net-wireless/gr-specest net-wireless/mousejack net-wireless/rfcat sci-biology/seqan sci-biology/shrimp sci-chemistry/eden sci-chemistry/pymol-plugins-caver sci-chemistry/numbat sci-chemistry/pymol sci-chemistry/sparky sci-libs/dealii sci-libs/gmsh sci-misc/gato sys-boot/refind sys-boot/udk sys-cluster/pbs-python sys-fs/rarfs sys-fs/traydevice www-apache/mod_scgi www-apps/scgi www-misc/nx_util x11-libs/flowcanvas x11-misc/dsx x11-misc/pypanel -- Cheers, Aaron signature.asc Description: PGP signature
[gentoo-dev] systems-246 changes tmpfs default size from 50% to 10% of RAM
Hi, I've upgraded to and running systemd-246_rc2 on one of my systems and noticed that tmpfs mounted directories are significantly smaller. This is because with commit https://github.com/systemd/systemd/commit/7d85383edbab73274dc81cc888d884bb01070bc2 they have changed them to be 10% of the physical memory instead of the default of 50%. This is a potentially breaking, or at least an unexpected behaviour change, especially for people using tmpfs on /tmp for compiling. Maybe we should make a news item to let people know that they either need to add an fstab entry with size option set, or better, create a systemd local override with relevant content. Cheers, Zoltan
Re: [gentoo-dev] systems-246 changes tmpfs default size from 50% to 10% of RAM
On Tue, Jul 28, 2020 at 3:13 PM Mike Gilbert wrote: > > On Tue, Jul 28, 2020 at 2:50 PM Zoltan Puskas wrote: > > > > Hi, > > > > I've upgraded to and running systemd-246_rc2 on one of my systems and > > noticed that tmpfs mounted directories are significantly smaller. > > > > This is because with commit > > https://github.com/systemd/systemd/commit/7d85383edbab73274dc81cc888d884bb01070bc2 > > they have changed them to be 10% of the physical memory instead of the > > default of 50%. > > > > This is a potentially breaking, or at least an unexpected behaviour > > change, especially for people using tmpfs on /tmp for compiling. > > > > Maybe we should make a news item to let people know that they either > > need to add an fstab entry with size option set, or better, create a > > systemd local override with relevant content. > > Don't use /tmp for PORTAGE_TMPDIR. /tmp is meant for small temporary > storage. If you want to compile in a tmpfs, set up a separate mount > point for it. > > I don't intend to create a news item for this, but I would not object > to someone else doing it. Also, the limit for /tmp is likely to change again before the 246 final release. https://github.com/systemd/systemd/pull/16576
Re: [gentoo-dev] systems-246 changes tmpfs default size from 50% to 10% of RAM
On Tue, Jul 28, 2020 at 2:50 PM Zoltan Puskas wrote: > > Hi, > > I've upgraded to and running systemd-246_rc2 on one of my systems and > noticed that tmpfs mounted directories are significantly smaller. > > This is because with commit > https://github.com/systemd/systemd/commit/7d85383edbab73274dc81cc888d884bb01070bc2 > they have changed them to be 10% of the physical memory instead of the > default of 50%. > > This is a potentially breaking, or at least an unexpected behaviour > change, especially for people using tmpfs on /tmp for compiling. > > Maybe we should make a news item to let people know that they either > need to add an fstab entry with size option set, or better, create a > systemd local override with relevant content. Don't use /tmp for PORTAGE_TMPDIR. /tmp is meant for small temporary storage. If you want to compile in a tmpfs, set up a separate mount point for it. I don't intend to create a news item for this, but I would not object to someone else doing it.