1. Forum
  2. Usenet
  3. LINUX.DEBIAN.DEVEL

  • Re: Strange armel build error

    From Paul Gevers@21:1/5 to Alec Leamas on Fri Aug 16 18:10:01 2024
    To: [email protected]

    This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --------------tak4PqKFBWlFXik0Ofzi8z0q
    Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: base64

    SGksDQoNCk9uIDE2LTA4LTIwMjQgMTc6NDYsIEFsZWMgTGVhbWFzIHdyb3RlOg0KPiBBbGwg b3RoZXIgYnVpbGRzIGFyZSBPSy4gSGFzIGFueW9uZSBhIGhpbnQgYWJvdXQgd2hhdCBtaWdo dCBiZSBnb2luZyBvbiBoZXJlPw0KDQpodHRwczovL3JlbGVhc2UuZGViaWFuLm9yZy90ZXN0 aW5nL2FyY2hfcXVhbGlmeS5odG1sIGFybWVsIGNvbHVtbi4NCg0KUGF1bA0K

    --------------tak4PqKFBWlFXik0Ofzi8z0q--

    -----BEGIN PGP SIGNATURE-----

    wsB5BAABCAAjFiEEWLZtSHNr6TsFLeZynFyZ6wW9dQoFAma/d7wFAwAAAAAACgkQnFyZ6wW9dQqt bQf/aJ+mqe7FypZsHB+nHvAoXLAdEL+x4CPpcoPO80chI2GZF9lbmykCvGepknLynEhGxWLAK6Ix 5m2nRwAcffy1rDF7mJ8TMjr0wbUSLMoDYwuoH+ggFeB880rUi43PzmngosZXtOXk0D54ub/VEeSX QlBkOyyHwodT3GTv+uAK4dcDI0SeRfP6hRwPMWLi/ftoz9x8FOF/NLyvkDX3DN2c93Za/uQ3kz+a G+eMybVoMu08bs7b/NpAA4znPvR727DHlmNggm3Y/6RuaizUGkQ2+riHfTZI1i3jfNc8FeYeH81G OeFXMHkr7JFPnTl5RuB/HRym2cK5Ovq5q+tXhEdGnQ==
    =9m7F
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Timo =?utf-8?Q?R=C3=B6hling?=@21:1/5 to All on Fri Aug 16 18:20:01 2024
    Hi Alec,

    * Alec Leamas <[email protected]> [2024-08-16 17:46]: >/usr/include/c++/14/bits/atomic_futex.h:278:(.text+0x16dc): undefined reference to `std::__atomic_futex_unsigned_base::_M_futex_notify_all(unsigned int*)'
    /usr/bin/ld: /usr/include/c++/14/bits/atomic_futex.h:278:(.text+0x29d8): undefined reference to `std::__atomic_futex_unsigned_base::_M_futex_notify_all(unsigned int*)
    AFAIK the armel architecture has no lock-free atomic primitives, so
    you probably need to link against the atomic library because opencpn
    uses std::atomic somewhere in its codebase.

    Try adding

    export DEB_CXXFLAGS_MAINT_APPEND += -latomic

    to your debian/rules.


    Cheers
    Timo



    --
    ⢀⣴⠾⠻⢶⣦⠀ ╭────────────────────────────────────────────────────╮
    ⣾⠁⢠⠒⠀⣿⡁ │ Timo Röhling │
    ⢿⡄⠘⠷⠚⠋⠀ │ 9B03 EBB9 8300 DF97 C2B1 23BF CC8C 6BDD 1403 F4CA │
    ⠈⠳⣄⠀⠀⠀⠀ ╰────────────────────────────────────────────────────╯

    -----BEGIN PGP SIGNATURE-----

    iQIzBAEBCgAdFiEEmwPruYMA35fCsSO/zIxr3RQD9MoFAma/efAACgkQzIxr3RQD 9MrRTQ/+JwIcRenBD3R7NVkRiR+rNGsMadgk8pX3sbU4BYWHlDLqoShpLAUkyf6A k4Kqt3psm4Wb9cT2CuJdlyGhuWkDfpkgEmwd1YIVNWk5voWCfpEegCNN87h5XYYY exXrEJOwU5hCKv7o+W8MHjKJjYSZdW9S2HwDNVwQP6SCmeLdoWIVhpNt8WDRquIK rr+CEiqZ/rTERX6yiwWvjkSRHpDxHyUXgirrlhBEpm+licJJrXr5dQKSuxJtgn78 G5OTULIz/phFljXV7oAsG/dbYZEuESpBP/oMlB5VwJwkWgHBrU+FWekAAVgiljlc D5MV88LcPD2y2MTS5UsPkv6+Y0bfJvJ/YUVQ7VDukQx
  • From Wookey@21:1/5 to Alec Leamas on Sat Aug 17 14:50:02 2024
    On 2024-08-16 17:46 +0200, Alec Leamas wrote:

    From another perspective: what is the right thing to do in a situation like this? Trying to hunt down the problem, and thus causing all sorts of noise like this message? This is what the policy says, but still...

    Or just exclude that architecture i. e., list all archs but armel?

    You should at least inform the relevant debian-ports list (less noisy
    than debian-devel, but as you can see debian-devel worked well in this
    case so that's not an unreasonable choice), and not just turn the
    architecture off. This can be via a cc:ed bug or just a mail. Bugs
    help track issues so future maintainers can find their similar problem (arch-specific issues often appear in multiple packages, needing
    similar solutions).

    If you get no response after a couple of weeks and need to do an
    upload then it is fair enough to disable the architecture until a fix
    arrives, but idealy prod again and wait a while if you can.

    One of the good things about debian is that we support a range of
    architectures (to the best of our abilities). Packagers are not
    expected to know how to fix arch-specific issues but they are expected
    to ask someone who might if they can help, and not just degrade debian
    by removing packages from arches without at least filing a bug + asking.

    All IMHO of course, and recognising that porter responses can be both
    slow and nonexistent, but I do think it's important that we try keep
    debian as consistent as possible across architectures, and don't just
    reach for the 'disable' button. For someone on a particular arch that
    is the same as the 'remove from archive' button in effect.

    Wookey
    --
    Principal hats: Debian, Wookware, ARM
    http://wookware.org/

    -----BEGIN PGP SIGNATURE-----

    iQIzBAABCgAdFiEER4nvI8Pe/wVWh5yq+4YyUahvnkcFAmbAmrUACgkQ+4YyUahv nkcR/RAAtRfOsO2z9VDHSMS9shSlXRe7UDDZ4i/TkxPJSZ09uLvN9TfNB/rFyJfS +Ed5R9ge1d8CpUIQ0hB6OS+6516yAghnjMW0acXi1awsm8r4wrDPuxVHLZL7GNRm oTNyfM06o3SsqknGgBp1MOp0LpAA/RcbW7aPAo/8BUepwyJZd2vl+YTjIPGSNwKm 3qDGdcSA8mG/KVZ9ZZfabboGVE3/bcNooaeZuRyvy1MubCElP+CcIhjn6h4TV8WB +GIYBYVQ1sK73obPpOapZnz96buVM+QU0gQncn3RvpnJpJSDBNzAWaD/5jybXjQ2 F8P2J+cgnL2HSkDuZBHrKA0Xd0Da9RiFEn5GM1xVlSwF/q8U+8Zzmkg6EMgjhBbT xFXKD7Cu++hMv+jFCcDPyOjQtNxp+/S+iIkkjnG2XErAkrKMLVQbzbFcR1Jj6sAj oOqv0FRxtwLxkpLcorEPFytAwPRRzHHWS/yFHgEV4VZ4t4tdDYtUcYum9Nz5IEa7 eP17JlhpgzcmejkuNLjy3L3E+CjNDVTYAY55C62XpbWIYBajWw5ySJGbnFySE0i4 JjeSL4ERaCCyeN3iTSV3J9YMMKyj8j/MTGfovRCXnxWOesMl3uIR8h2Dx9hxnC99 Lc9Gmqb8YrQcQsmfX4PlWXBB1EQuY3iZjf02v37gRzkkIod1Krw=
    =P/Oe
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Wookey@21:1/5 to Alec Leamas on Sun Aug 18 06:20:01 2024
    On 2024-08-17 17:58 +0200, Alec Leamas wrote:


    Fair enough. But TBH, i just cannot wait "a couple of weeks" for a possible reply; there are users waiting for the backports as I write.

    Fair enough. If you can't wait then you can't wait.

    To make it more interesting, the simple -latomic fix doesn't seem to cut it

    That's a pity, it sounds plausible. I'll try to take a look.

    To me, the reasonable approach would be to inform the porter list (need to figure out which) and then disable armel for now. As soon as there is a solution I could and should upload it.

    That is indeed a totally reasonable approach.

    Wookey
    --
    Principal hats: Debian, Wookware, ARM
    http://wookware.org/

    -----BEGIN PGP SIGNATURE-----

    iQIzBAABCgAdFiEER4nvI8Pe/wVWh5yq+4YyUahvnkcFAmbBdG0ACgkQ+4YyUahv nkcX5w/+OB4G/Y/CB2K00PgaptZi6dOaYH050iPLVR+1xVu9iPTYPPCmvo9W4zeu dST8maSACAbSndRd5NkaeYKCiAGks7lc/G39B5/0yUVWV+vXi3mdK+867crZEYbN W3hr4dYVpKUQB2SICWdiulhsC5/O0NKharR+euGeARVvShVwTQwDrzHyFqObxRf9 yOvBaZaUpuANvm4/Dj8BjaLb5Gs1XsASTst73EbGzvXDocUZpb4R0InqV/2jv2dU ATfC23yKQVHjpv0ix2oI6d+zx1BO5iIMXNPZ2scYa5GjoNGZAO4EDTf5yIzo0JvC zLbPuLVJfTIOe4DBantp2MJ/+pms2jCFqTqxeowGia2ecra2fMmvggtD9FbWrS6v UxkmQhn5+v0vbhDHRmdL93SX0u0bKDVxHXnVoTEPyn5dA1HCe4YppRvPZT8zyEdQ xunRXZYXXs3/0PWe4rvzp3nMXtRtte5BOiLHZ2CIXLoZK8yNbYWHLkM/+zcXNgsS wEZE9f8bY72cmHz76BqYXfaFDfSJ9q3I0slC5pHvkKx5Qd17Eav+1z2fXBvkpOmy 5Mh0+H7/AZyQRgtQhOvZBFwSZyO3zMCpTNbJDwpeAQmP+EHxBc9j9cYEsHMnxXzL vtIgfy3RgFA067dlDJ/FOXK18NqD3AxOmOzOlOEBvssOmeaRxrU=
    =1D93
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Stephen Kitt@21:1/5 to Alec Leamas on Sun Aug 18 09:10:01 2024
    On Fri, 16 Aug 2024 17:46:45 +0200, Alec Leamas <[email protected]> wrote:
    Or just exclude that architecture i. e., list all archs but armel?

    If you can’t fix the build, you don’t need to exclude the architecture — you
    can ask for removal of the armel package in testing. That will allow the package to migrate even if armel is missing.

    Regards,

    Stephen

    -----BEGIN PGP SIGNATURE-----

    iQIzBAEBCgAdFiEEnPVX/hPLkMoq7x0ggNMC9Yhtg5wFAmbBnQMACgkQgNMC9Yht g5zl5Q/8DOO7mXxTtgEYHVV14egfxp3dnVT14bSd7dODRtwywTM3xkbezTzkB7Kn kRMidJ1ITQA2EEikoSfge8rYEFaxzm7Z+gUnPvNqvX+GIjd0qUfAIlr2yNyfNTsC syWT/tpxnp2taLsnRGcyN35qlsRSnKktW6OR7wqnck0VCuQsKmTX3cL0f/oU/Jjp EIrSS8CZPREgrMVmvjn7vycwta4upluyDUzbpONJbiR3EarSOO5CfeFNy+ZL71kp 3mqfcev1bl1sp8BQ/Xhzh+I1lqTC19SLyQPOXVu9hQBp0ECW6+Oyil6uxjGxlfVr dIGikLDASU6qaHsvdSBIvCNc7/iEcee0HoqvJ53ntNuNpF9iJXZTZYpZJTVO9R+q xEJ6BGPIYxpC0PiD6P8LIK9drocnzoQqHMLoLXdxtsEc5c3nxBzaGjwa4YroeG33 Dcum4yxMdxrVefdiYm6jJgm58O/2HwHk3ybq8OKHWciEbOeP2arcLIhcQUncP3I5 k7B3Ymbh/k4tbEIUAM6Z6tFdvb2fCWmg4/iHG2hlAB/ZRrpx04oUiHNypla4aRFi BLq/s2oTv00WdDN/gv9quxzQbSwDuruztWNBkJqIA7nG61XCjPYkZwu7Lu9dg0M8 7B5KXRiivQFeeWJdTEtC8prswt2ndY3aYAhiZDFcINUp7eF0VZc=
    =b1D8
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Adam D. Barratt@21:1/5 to Andrey Rakhmatullin on Sun Aug 18 11:40:01 2024
    On Sun, 2024-08-18 at 14:23 +0500, Andrey Rakhmatullin wrote:
    On Sun, Aug 18, 2024 at 11:02:03AM +0200, Alec Leamas wrote:
    Hi Stephen,

    On 18/08/2024 09:04, Stephen Kitt wrote:

    [...]
    If you can’t fix the build, you don’t need to exclude the architecture — you
    can ask for removal of the armel package in testing. That will
    allow the
    package to migrate even if armel is missing.

    This looks to me like a sound solution in this case. After all,
    opencpn is a
    full-fledged GUI leaf package without reverse deps and zero users
    on armel
    hardware. But then again, how is this done?

    If by this you mean asking for a removal then reportbug
    release.debian.org

    Not for an architecture-specific removal. Those happen in unstable and
    then propagate to testing, so it's ftp.debian.org.

    Regards,

    Adam

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Andrey Rakhmatullin@21:1/5 to Alec Leamas on Sun Aug 18 11:30:01 2024
    On Sun, Aug 18, 2024 at 11:02:03AM +0200, Alec Leamas wrote:
    Hi Stephen,

    On 18/08/2024 09:04, Stephen Kitt wrote:
    On Fri, 16 Aug 2024 17:46:45 +0200, Alec Leamas <[email protected]> wrote:
    Or just exclude that architecture i. e., list all archs but armel?
    If you can’t fix the build, you don’t need to exclude the architecture — you
    can ask for removal of the armel package in testing. That will allow the package to migrate even if armel is missing.

    This looks to me like a sound solution in this case. After all, opencpn is a full-fledged GUI leaf package without reverse deps and zero users on armel hardware. But then again, how is this done?

    If by this you mean asking for a removal then reportbug release.debian.org

    --
    WBR, wRAR

    -----BEGIN PGP SIGNATURE-----

    iQJhBAABCgBLFiEEolIP6gqGcKZh3YxVM2L3AxpJkuEFAmbBvZgtFIAAAAAAFQAP cGthLWFkZHJlc3NAZ251cGcub3Jnd3JhckBkZWJpYW4ub3JnAAoJEDNi9wMaSZLh 2r8P+wZ2L/jaQdP+o8tN3ZLQybGn/Nj/O4LnM4LC5ecFuWhvHHgzEJJc2dWm9sI2 DqRmKubv9KHcYyDp1LtKcVVfcC2TvrsRwPKoAN6yItng04c/TZdX3nUQV+W3m9yU W+btrIug0n5qLP2R5ygFC1ACn34vTarS8u0LO3xeLng915O+DoR7rLOq9GkYZsrT werYG++mWQvoy8LeXalmw2YHRiOQVVelBaR472Y4KBZkT9cq6r1BaQgtqxgz8igz wXOXlp6VKGRmu8OwKu6xX+qVBDDLYXgMs7UU9AZLmQf+EaPizMAQTBMMyfBe6Vkc 2ArpG/ajMLoY1ajEDsSJcfsRtU0KmHFwN4aixErVBEO2VKgGsyplEfh42rypWh3O d9d8aXxoOf+Tsf5Uh5dXKoCy6YZWFOTS/bQwWPBS2sFsd/aI8HazzOGjsZQoOpNt KoctdlOBRoDQXnMftP2KD6xlXXgP+iIs+sn4iI+Yr7ZlXAdmcoppw4n+yvGwhXqr CSQ9cATOwvoq6S3VHdj5JRSG1f7ZBs3yaMMVoWNTNLok8D8UfT7ZlUT/322eE028 1q+kmVEvckKOHT/jJjpSEDqyjUk9tEmGjOsjiU5azDxHS8T0K21HY5shSvfoSv8F ukLW7WhLUccANeNB8TEaVx8k8iijgXy7MhZqzCu0r30FO2FP
    =fHsz
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)