Thursday, 2023-08-17

*** kpo <kpo!> has joined #yocto00:03
*** tgamblin <tgamblin!~tgamblin@2001:1970:5b1f:ab00:71eb:b8e9:8589:9ca5> has quit IRC (Remote host closed the connection)00:22
*** tgamblin <tgamblin!~tgamblin@2001:1970:5b1f:ab00:d875:6297:4e81:e577> has joined #yocto00:23
*** sakoman <sakoman!> has quit IRC (Quit: Leaving.)00:45
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto00:57
*** Ablu <Ablu!~Ablu@user/Ablu> has quit IRC (Ping timeout: 245 seconds)01:13
*** Ablu <Ablu!~Ablu@user/Ablu> has joined #yocto01:15
*** pabigot <pabigot!~pab@> has quit IRC (Ping timeout: 248 seconds)01:33
*** pabigot <pabigot!> has joined #yocto01:47
*** starblue <starblue!> has quit IRC (Ping timeout: 250 seconds)01:49
*** starblue <starblue!> has joined #yocto01:51
*** sakoman <sakoman!> has joined #yocto02:21
khemRP:  is it in pseudo but other places ?02:26
*** sakoman <sakoman!> has quit IRC (Ping timeout: 246 seconds)02:26
*** sakoman <sakoman!> has joined #yocto02:26
*** sakoman <sakoman!> has quit IRC (Quit: Leaving.)02:44
*** sakoman <sakoman!~steve@> has joined #yocto02:45
*** otavio_ <otavio_!> has quit IRC (Ping timeout: 246 seconds)03:06
*** amitk <amitk!~amit@> has joined #yocto03:18
*** otavio <otavio!> has joined #yocto03:58
dvergatalabelloni: I see that build has failed and packages which has failed are still the same as previously04:37
dvergataldvergatal: so we need to get that *_package.tar.zst for one of these04:39
dvergatalabelloni: ^04:39
*** sakoman <sakoman!~steve@> has quit IRC (Quit: Leaving.)04:56
*** yates_work <yates_work!~user@> has joined #yocto05:02
*** brrm <brrm!~brrm@2a02:8071:b780:ece0::1> has quit IRC (Ping timeout: 246 seconds)05:27
*** brrm <brrm!> has joined #yocto05:29
*** valdemaras <valdemaras!~valdemara@> has joined #yocto05:47
*** alessioigor <alessioigor!~alessioig@> has joined #yocto05:50
*** Chaser <Chaser!~Chaser@user/chaser> has joined #yocto05:54
*** ray-san2 <ray-san2!~ray-san@> has joined #yocto06:10
*** frieder <frieder!> has joined #yocto06:13
*** rfuentess <rfuentess!> has joined #yocto06:15
*** kpo <kpo!> has quit IRC (Ping timeout: 245 seconds)06:22
*** xmn <xmn!> has quit IRC (Quit: ZZZzzz…)06:22
*** goliath <goliath!~goliath@user/goliath> has joined #yocto06:24
*** valdemaras <valdemaras!~valdemara@> has quit IRC (Ping timeout: 250 seconds)06:33
*** olani <olani!> has quit IRC (Ping timeout: 250 seconds)06:39
*** olani <olani!~olani@> has joined #yocto06:39
*** Schlumpf <Schlumpf!~Schlumpf@> has joined #yocto06:41
SchlumpfGood morning06:41
*** olani_ <olani_!~olani@> has joined #yocto06:42
RPkhem: I've worked out a horrible hack in next06:42
*** Kubu_work <Kubu_work!> has joined #yocto06:44
*** olani_ <olani_!~olani@> has quit IRC (Remote host closed the connection)06:58
*** olani_ <olani_!~olani@> has joined #yocto06:59
*** vladest <vladest!> has quit IRC (Ping timeout: 250 seconds)07:19
*** TroubledGuest <TroubledGuest!~TroubledG@> has joined #yocto07:23
*** flom_84 <flom_84!> has joined #yocto07:24
TroubledGuestHello everyone,07:30
TroubledGuestI've recently created a .bb recipe for the libfaketime library and I'm interested in contributing it as a patch to the OpenEmbedded layers. I'm currently trying to determine the most suitable layer for this contribution. I've considered placing it in a miscellaneous bits and pieces layer, although I haven't come across one specifically for that07:30
TroubledGuestpurpose (aside from meta-webstuff, which seems to be reserved for web-related items). Could you guide me regarding which layer would be the best fit for this type of contribution?07:30
TroubledGuestThank you for assistance :)07:30
*** pabigot <pabigot!> has quit IRC (Ping timeout: 256 seconds)07:36
LetoThe2ndyo dudX07:37
*** Schlumpf <Schlumpf!~Schlumpf@> has quit IRC (Quit: Client closed)07:40
*** flom_84 <flom_84!> has quit IRC (Ping timeout: 256 seconds)07:40
*** flom_84 <flom_84!> has joined #yocto07:44
*** Schlumpf <Schlumpf!~Schlumpf@> has joined #yocto07:48
*** flom_84 <flom_84!> has quit IRC (Max SendQ exceeded)07:48
*** vladest <vladest!~Thunderbi@> has joined #yocto07:48
*** flom_84 <flom_84!> has joined #yocto07:49
*** vladest <vladest!~Thunderbi@> has quit IRC (Client Quit)07:50
*** vladest1 <vladest1!~Thunderbi@> has joined #yocto07:50
*** pabigot <pabigot!> has joined #yocto07:52
*** vladest1 is now known as vladest07:52
*** rfuentess <rfuentess!> has quit IRC (Remote host closed the connection)07:54
*** flom_84 <flom_84!> has quit IRC (Ping timeout: 252 seconds)07:56
dvergatalRP: are you there?07:58
*** prabhakarlad <prabhakarlad!> has joined #yocto07:59
*** zpfvo <zpfvo!> has joined #yocto08:01
*** leon-anavi <leon-anavi!~Leon@> has joined #yocto08:15
*** jclsn <jclsn!~jclsn@2a04:4540:6527:9600:2ce:39ff:fecf:efcd> has joined #yocto08:18
*** flom_84 <flom_84!> has joined #yocto08:24
*** florian <florian!> has joined #yocto08:27
RPdvergatal: I am now08:31
RPdvergatal: I guess the question is can we get the sstate hash of that failed build?08:33
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto08:37
dvergatalRP: yes exactly thx08:41
dvergatalRP: not the package write08:41
dvergatalbut package.tar.zst08:41
dvergatalRP: I was right the sstate is broken08:50
dvergatalls --full-time ./packages-split/flac-src/usr/src/debug/flac/1.4.3-r0/include/FLAC++/decoder.h08:52
dvergatal-rw-r--r-- 2 plobacz plobacz 12789 2023-06-15 11:52:54.000000000 +0200 ./packages-split/flac-src/usr/src/debug/flac/1.4.3-r0/include/FLAC++/decoder.h08:52
RPdvergatal: the question is then why though08:52
dvergatalRP: something has overwritten it08:53
dvergatalquestion is what?08:53
*** flom_84 <flom_84!> has quit IRC (Quit: Leaving)08:53
dvergatalmaybe previous build?08:53
RPdvergatal: I find that a bit unlikely given what i know of the code08:53
*** flom_84 <flom_84!> has joined #yocto08:53
dvergatalbut you see it does not have the milliseconds in it and I have upacked it with posix format08:54
RPdvergatal: but that doesn't mean it was overwritten08:54
dvergatalhmmmm than what?08:54
RPmore likely, not overwritten!08:54
dvergatalok so something has changed it and I really do not know what....08:55
*** florian_kc <florian_kc!> has joined #yocto08:57
dvergatalRP: hmmmm, is it possible that if the patch for sstate.bbclass is missing in the branch and sstate is shared between different builds it can modify the sstate?08:58
RPdvergatal: if you look at the code in sstate.bbclass, it creates the file and then moves into position only if the file doesn't exist08:59
*** valdemaras <valdemaras!~valdemara@> has joined #yocto08:59
RPthe builds all do share sstate, it is designed to work like that08:59
RPif your patch is breaking the checksums somehow, that is also an issue we'd have to fix as sharing sstate is expected to work09:00
RPdvergatal: and following that thought, have a look at tmp/work/core2-64-poky-linux/flac/1.4.3/temp/depsig.do_package09:01
RPdvergatal: this is the data fed to hash equivalence. Those timestamps don't contain subsecond precision09:02
RPdvergatal: so it would ignore the subseconds when marking something as hash equivalent09:03
RPthe depsig files represent the data it includes to generate the "outhash" used there for comparisions to say whether tasks had identical output09:03
dvergatalRP: one sec09:04
RPdvergatal: come to think of it, the acl and attr data isn't in there either09:09
*** flom__84__ <flom__84__!> has joined #yocto09:12
*** flom_84 <flom_84!> has quit IRC (Quit: Leaving)09:13
*** flom__84__ <flom__84__!> has quit IRC (Remote host closed the connection)09:13
*** flom__84__ <flom__84__!> has joined #yocto09:14
*** flom__84__ <flom__84__!> has quit IRC (Remote host closed the connection)09:15
*** flom__84__ <flom__84__!> has joined #yocto09:15
*** flom__84__ <flom__84__!> has quit IRC (Remote host closed the connection)09:16
*** flom__84__ <flom__84__!> has joined #yocto09:17
*** hcg <hcg!~hcg@> has joined #yocto09:19
dvergataldvergatal: it's not?>09:19
dvergatalRP: ^^09:20
dvergatalRP: I'm sorry to many ppl wanted from me atm09:20
RPdvergatal: I don't remember adding any code to handle those things there,  I could be wrong09:21
dvergatalRP: what do you mean by that? because now I don't understand you what code and where?09:22
RPdvergatal: the depsig files I'm pointing at. They don't account for millisecond data and I'm now suspecting they don't account for xattrs or acls either09:22
dvergatalyes I wasn't addint it there09:22
dvergatalso it is probably missing there09:23
*** flom__84__ <flom__84__!> has quit IRC (Quit: Leaving)09:23
*** flom__84__ <flom__84__!> has joined #yocto09:23
dvergatalRP: what location is it?09:23
RPdvergatal: you mean where is the file in the build, where is it used or where is it generated?09:23
RPor something else09:23
dvergatalRP: these depsig09:24
RPdvergatal: I told you where they are above?09:24
dvergatalok let me check it in my poky one moment09:24
dvergatalRP: this is odd, I don't have it on my frwesh poky distro which was tested on saturday09:26
RPdvergatal: it will depend on whether it was from sstate or not. You should have some depsig files in a build to look at though09:27
*** TroubledGuest <TroubledGuest!~TroubledG@> has quit IRC (Quit: Client closed)09:27
RPjust force something to rebuild (bitbake flac -C compile)09:27
dvergatalRP: I have only for gcc-runtime, acpid, glibc, linux-libc-headers, initscripts and libgcc09:27
dvergatalRP: OK calling `bitbake flac -C compile` created it for me09:36
dvergatalRP: you are right we need to add these timestamps there as well09:37
dvergatalRP: I have found that, `report_unihash` from `bitbake/lib/bb/` is writing that depsig.do_package file, but I can't find the part responsible for the list of files with attributes09:54
*** prabhakarlad <prabhakarlad!> has quit IRC (Quit: Client closed)09:56
*** jetm <jetm!> has quit IRC (Quit: ZNC 1.7.2 -
*** philmd <philmd!> has quit IRC (Quit: ZNC 1.7.2 -
*** flom__84__ <flom__84__!> has quit IRC (Ping timeout: 246 seconds)09:58
LetoThe2ndis it just me or has auhbombing intensified?10:02
*** starblue <starblue!> has quit IRC (Ping timeout: 245 seconds)10:02
*** starblue <starblue!> has joined #yocto10:04
dvergatalLetoThe2nd: me210:04
RPLetoThe2nd: we had to rerun due to the first run being totally broken10:07
LetoThe2ndRP: ah thx10:07
*** xmn <xmn!~xmn@2600:4040:9390:8c00:5860:e2a9:13e6:4d45> has joined #yocto10:08
* RP can tell who reads the weekly status10:10
*** hcg <hcg!~hcg@> has quit IRC (Quit: Client closed)10:11
*** Guest62 <Guest62!> has joined #yocto10:14
Guest62how do i tell yocto that a recipe replaces another to prevent building the wrong version from the old recipe10:15
LetoThe2ndGuest62: PREFERRED_VERSION10:18
* LetoThe2nd can tell RP that he doesn't, though probably should10:19
*** jetm <jetm!> has joined #yocto10:21
*** jetm <jetm!> has quit IRC (Client Quit)10:22
*** hcg <hcg!~hcg@> has joined #yocto10:22
*** jetm <jetm!> has joined #yocto10:27
*** philmd <philmd!~philmd@> has joined #yocto10:30
*** valdemaras <valdemaras!~valdemara@> has quit IRC (Quit: valdemaras)10:42
*** ptsneves <ptsneves!> has joined #yocto10:44
*** frieder <frieder!> has quit IRC (Remote host closed the connection)11:02
RPdvergatal: meta/lib/oe/sstatesig.py11:03
dvergatalRP: thx, I was looking for it and couldn't find11:05
dvergatalRP: this is the line `update_hash(" %10d" % s.st_mtime)`11:06
dvergatalRP: yeah heheheh stupid fix:P changing from %10d to %f11:18
dvergatalRP: from what I'm seeing here there are other places that uses st_mtime in here but none is passing it as integer11:20
dvergatalok I'm adding a fix for that and pushing new patchset, is that ok for all of you?11:21
mcfriskRP: how likely is the qemu x86_64 and 6.4 kernel boot hang? 1/10, 1/100 or less boots?11:24
ptsnevesHey all. Does anybody know how to disable sstate cache checking at parse time in tinfoil?11:39
*** TroubledGuest <TroubledGuest!~TroubledG@> has joined #yocto11:44
*** TroubledGuest <TroubledGuest!~TroubledG@> has quit IRC (Client Quit)11:46
*** leonanavi <leonanavi!~Leon@> has joined #yocto12:06
*** lexano <lexano!~lexano@> has quit IRC (Ping timeout: 246 seconds)12:08
*** xmn_ <xmn_!~xmn@2600:4040:9390:8c00:c65:2434:df3f:83e1> has joined #yocto12:15
*** xmn_ <xmn_!~xmn@2600:4040:9390:8c00:c65:2434:df3f:83e1> has quit IRC (Client Quit)12:15
*** xmn <xmn!~xmn@2600:4040:9390:8c00:5860:e2a9:13e6:4d45> has quit IRC (Ping timeout: 246 seconds)12:17
*** xmn <xmn!~xmn@2600:4040:9390:8c00:c65:2434:df3f:83e1> has joined #yocto12:18
*** DvorkinDmitry <DvorkinDmitry!~dvorkin@> has joined #yocto12:21
DvorkinDmitryI have several apps (different versions of the bootloader0) for secondary CPU need to be build if I'm building the bootloader (bootloader1) for the main CPU. how to correctly setup the dependancy?12:24
ptsnevesDvorkinDmitry: When you say secondary and primary CPU are they a different Yocto MACHINE?12:25
DvorkinDmitryptsneves, yes. machine0 and machine112:25
ptsnevesThen take a look at this link
*** zpfvo <zpfvo!> has quit IRC (Ping timeout: 260 seconds)12:28
*** flom__84__ <flom__84__!> has joined #yocto12:29
RPmcfrisk: we've seen two and none since :/12:29
mcfriskRP: two out of how many runs?12:30
RPmcfrisk: I don't know how many runs each autobuilder run counts as so now sure. Few hundred?12:30
DvorkinDmitryptsneves, yes. I have this in my rootfs recipe (several lines, several aps). But going to name the group of apps (say, virtual/bootloader0 to build several apps under one name) and write the only one dependancy. Not sure how to make it correctly. apps are different recipes (with different build flags for same app)12:32
mcfriskRP: ok 1/100 is the rough ballpark occurence, I guess x86_64 workers running qemu x86_64 systems12:32
ptsnevesDvorkinDmitry: So create a packagegroup with all those apps you want and then refer to only the packagegroup dependency when needed. If the packagegroup is not completely a good fit(because bootloaders are not often packages) you can emulate such behavior in many ways.12:34
RPmcfrisk: something like that, yes. they're kvm accelerated12:36
*** paulg <paulg!> has quit IRC (Remote host closed the connection)12:40
mcfriskRP: alright, I'll prepare some test runs too after some qemu/slirp/testimage fixes..12:41
DvorkinDmitryptsneves, oh, I thought about package group. not a bad idea. but could you tell me another ways? just curious12:42
*** zpfvo <zpfvo!> has joined #yocto12:43
*** amitk_ <amitk_!~amit@> has joined #yocto12:49
ptsnevesDvorkinDmitry: speculating but i can imagine a recipe with the do_build task defining all the multiconfig dependencies. Then nobody outside that recipe would need to know those details and you could just add (R)DEPENDS to the image/recipe needing those apps.12:49
dvergatalabelloni: RP: I have send new patchset including this fix for floating point number in timestamp and additional minor changes in patches e.g. changes of status etc.12:50
DvorkinDmitryptsneves, yes. that's what I have now in my image recipe: do_image[mcdepends] += "mc:..app1 mc:..app2". I had the same in one "meta" recipe. Now looking for the way to butify the construct12:53
RPdvergatal: thanks. Don't we need the acl/xattr info there as well in the depsig files?12:58
dvergatalin attributes?12:58
RPdvergatal: also, for readability that field was fixed width12:58
dvergataldvergatal: wait i'll check12:59
dvergatalRP: it looks fine13:00
*** hcg <hcg!~hcg@> has quit IRC (Quit: Client closed)13:00
dvergatalRP: you want a screenshoot?13:00
dvergatalRP: regarding modes we probably need to extend it to something else than just e.g. drwxr-xr-x13:01
dvergatalRP: we will probably need to discuss it how to store it13:01
*** vvmeson <vvmeson!> has quit IRC (Quit: Konversation terminated!)13:06
dvergatalRP: because right now I can't imagine how to store acls/xattrs for each file without + at the end of modes and writting the output from getfacl will be probably painfull13:07
*** 029AALG7R <029AALG7R!> has joined #yocto13:10
RPI guess a screenshot or a pastebin or share of a file might help me see it.13:13
dvergatalok pastebin13:14
RPwe do need to record the acl/xattr data there13:14
RPJPEW: any suggestions on how to add this to depsig?13:14
JPEWHmm, I'm not to familiar with ACL TBH. I'll take a closer look after I walk the kids to school13:17
*** flom__84__ <flom__84__!> has quit IRC (Quit: Leaving)13:21
dvergatalRP: and another one with timestamps including milliseconds
*** amitk_ <amitk_!~amit@> has quit IRC (Ping timeout: 256 seconds)13:24
dvergatalOK I'm going on a bike be probably in the evening13:24
dvergatalabelloni: RP: let me know when the build starts13:25
*** kpo <kpo!~kpo@> has joined #yocto13:25
*** flom84 <flom84!> has joined #yocto13:29
*** flom84 <flom84!> has quit IRC (Remote host closed the connection)13:30
RPdvergatal: ok, timestamps look ok but we still need to sort the acl/xattr issue13:30
*** paulg <paulg!> has joined #yocto13:35
*** flom84 <flom84!> has joined #yocto13:35
*** 029AALG7R is now known as vmeson13:35
*** amitk_ <amitk_!~amit@> has joined #yocto13:37
jonmasonzeddii: you saw 724ba6751532055db75992fc6ae21c3e322e94a7 in v6.5?  Looks like all of the arm32 dts have moved under vendor named directories13:41
jonmasoneasy enough to work around, but wanted to make you aware13:41
*** sakoman <sakoman!> has joined #yocto13:41
zeddiijonmason. I did! but in my 6.5 testing with qemu I only did 64bit on qemu, so didn't notice much13:41
jonmasonI'll push a patch to fix qemuarmv5 in linux-yocto-dev today13:42
zeddiicool. I didn't test it yet, so that is appreciated.13:42
jonmasonthe beenfit of having dev-kernel as part of meta-arm ci ;-)13:42
*** fuzzybear86 <fuzzybear86!> has joined #yocto13:44
jonmasonzeddii: thinking aloud, this is going to make having 2 kernels a PITA after v6.413:44
zeddiiyah, since the kernel version is not an OVERRIDE value13:45
zeddiiper-machine works though, for any set variables.13:45
zeddiibut again, no version. hmm.13:45
jonmasonyeah, that'll probably mean that all of the arm32's are going to force a version for the next release13:46
jonmasonthat or have a bbappend with a DT location override :/13:46
* zeddii ponders13:47
*** flom84 <flom84!~flom84@user/flom84> has quit IRC (Quit: Leaving)13:49
*** flom84 <flom84!~flom84@user/flom84> has joined #yocto13:49
jonmasonzeddii: looking through other fails of dev-kernel, qemuarm needs the frame buffer fix for 6.5 as well.  Looks like my patch has no response upstream :(13:52
*** rfuentess <rfuentess!> has joined #yocto13:52
jonmasonall of that should get it CI green for me13:53
jonmasonexcept for some meta-arm BS that needs debugging13:53
zeddiijonmason. I'll do that cherry pick. leave that with me.13:53
jonmasonok, let me get that qemuarmv5 patch out13:54
*** Guest31 <Guest31!~Guest31@> has joined #yocto13:57
*** flom84 <flom84!~flom84@user/flom84> has quit IRC (Quit: Leaving)14:00
vvnhey, when a library have several versions that can be installed simultaneously, but programs expect a fixed name, e.g. is installed but program foo expects Should I add PACKAGES += "${PN}-foo" which installs a symlink?14:01
*** flom84 <flom84!~flom84@user/flom84> has joined #yocto14:01
*** Xagen <Xagen!~Xagen@> has joined #yocto14:03
*** flom84 <flom84!~flom84@user/flom84> has quit IRC (Remote host closed the connection)14:04
*** flom84 <flom84!~flom84@user/flom84> has joined #yocto14:04
JPEWRP, dvergatal: It should be possible to add xattrs and ACLs to OEOuthashBasic, but you'd want to read them from python as invoking a program to do it will be too slow; Python doesn't have native support this though, only with additional modules14:11
*** flom84 <flom84!~flom84@user/flom84> has quit IRC (Ping timeout: 250 seconds)14:15
*** Schlumpf <Schlumpf!~Schlumpf@> has quit IRC (Quit: Client closed)14:15
RPJPEW: that does make it sound tricky :(14:16
JPEWIt might be possible to randomly sort our the syscalls, or copy in the library code; I'm not a fan of either but I doubt it will fly to add pip dependencies to bitbake either :)14:17
RPJPEW: different kinds of pain14:18
*** Guest62 <Guest62!> has quit IRC (Quit: Client closed)14:19
JPEWBoth the existing libraries for it are C-python bindings, so copying in the code isn't an option (they aren't pure-python)14:19
JPEWbut they are both active project14:20
RPJPEW: I'm guessing they link to libraries too14:20
RPJPEW: I know xattr and acl from their recipes14:23
JPEWAh, right14:23
RPJPEW: I guess we at least have recipes for those bits14:24
JPEWWe could I suppose write python native bindings to those libraries14:25
JPEWIt would be better than syscalls IMHO14:25
RPyes, this isn't trivial stuff14:25
JPEWMaybe anyway; I've not looked at the library API at all, so someone correct me if I'm wrong14:26
RPJPEW: I've not in a long time and I can't remember anything14:26
JPEWEh, we've all got more important things to remember than esoteric APIs. I still have to look up almost every python API if it's been more than an hour since I last used it :)14:27
*** ray-san2 <ray-san2!~ray-san@> has quit IRC (Remote host closed the connection)14:28
*** goliath <goliath!~goliath@user/goliath> has quit IRC (Quit: SIGSEGV)14:30
*** belsirk <belsirk!~rfuentess@> has joined #yocto14:30
*** rfuentess <rfuentess!> has quit IRC (Read error: Connection reset by peer)14:32
*** amitk_ <amitk_!~amit@> has quit IRC (Quit: Lost terminal)14:32
*** rfuentess <rfuentess!> has joined #yocto14:34
*** belsirk <belsirk!~rfuentess@> has quit IRC (Ping timeout: 256 seconds)14:36
vvnOr can a user specify a different installation prefix for a specific package?14:36
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto14:37
vvnif there are,, etc. and I want to install them simultaneously, can I install them both and choose a different prefix for them? (e.g. /opt/foo/1.0 and /opt/foo/2.0 respectively)14:37
vvn(there are use cases where a system needs several versions of the same library for various applications)14:38
mcfriskvvn: yes but all files must be in different paths also for packages which are not no rootfs, e.g. dev, dbg, headers..14:43
*** beneth <beneth!> has quit IRC (Quit: Gateway shutdown)14:44
*** beneth <beneth!> has joined #yocto14:46
*** vladest <vladest!~Thunderbi@> has quit IRC (Ping timeout: 258 seconds)14:48
*** fuzzybear86 <fuzzybear86!> has quit IRC (Quit: Client closed)14:52
DvorkinDmitryhow to write EXTRA_IMAGEDEPENDS += "myrecipe" in multiconfig machine configuration, like machine1.conf: EXTRA_IMAGEDEPENDS += "mc:machine1:machine2:myrecipe:do_deploy"14:54
vvnDvorkinDmitry: EXTRA_IMAGEDEPENDS expects a package name, not a task14:55
vvnwhat is the use case exactly?14:55
DvorkinDmitryvvn, yes. I need to run the recipe deploy for the recipe (built for machine1) before creating the image for machine214:56
DvorkinDmitryi.e. build the firmware for secondary CPU and deploy it before creating the image for main CPU14:57
DvorkinDmitryand setup the dependancy in machine config14:57
vvnmcfrisk: what about a bbclass that moves the installed files to a different prefix (e.g. /opt/${BPN}/${PV})? So that if one wants to install multiple versions of the same package (assuming the different recipes exist), they can add foo_%.bbappend to inherit multiversion?14:59
vvnDvorkinDmitry: you can create a package which has some_task[mcdepends] += "mc:machine1:machine2:myrecipe:do_deploy" then add EXTRA_IMAGEDEPENDS += "my-firmware"15:02
*** fuzzybear8 <fuzzybear8!> has joined #yocto15:04
paulgmy moron move of the morning. Making a typo PREFFERED_VERSION and wondering why it wasn't taking effect...15:08
ptsnevespaulg: Dont worry i spent most of my day looking for why myvar["key}"] was returning nothing15:10
*** alessioigor <alessioigor!~alessioig@> has quit IRC (Quit: alessioigor)15:10
*** alessioigor <alessioigor!~alessioig@> has joined #yocto15:10
*** Guest31 <Guest31!~Guest31@> has quit IRC (Ping timeout: 246 seconds)15:18
dvergatalRP: JPEW: ok from my side I will talk on that tomorrow with my team mates15:18
*** fuzzybear8 <fuzzybear8!> has quit IRC (Quit: Client closed)15:19
dvergatalRP: we still have an issue to solve with setting the order of postins-useradd-${PN} calls in case of more DEPENDS in chain15:20
*** vladest <vladest!> has joined #yocto15:28
*** rfuentess <rfuentess!> has quit IRC (Remote host closed the connection)15:30
*** Haxxa <Haxxa!> has quit IRC (Quit: Haxxa flies away.)15:31
*** Haxxa <Haxxa!> has joined #yocto15:31
DvorkinDmitryvvn, I've created packagegroup-mytask should I have something like RDEPENDS:${PN} += "mc:mach1:mach2:my-firmware:do_deploy" inside, or syntax is different?15:41
DvorkinDmitryvvn, oh,m your solution is different... I'm lost. If I can do like I'm trying to do? with packagegroup? or better to use some "fake" recipe?15:43
*** paulg <paulg!> has quit IRC (Read error: Connection reset by peer)15:46
*** paulg <paulg!> has joined #yocto15:47
*** vladest <vladest!> has quit IRC (Remote host closed the connection)15:48
*** amitk_ <amitk_!~amit@> has joined #yocto15:48
*** vladest <vladest!> has joined #yocto15:50
ptsnevesDvorkinDmitry: you cannot put tasks in RDEPENDS as mentioned by vvn earlier :)15:56
*** ptsneves <ptsneves!> has quit IRC (Ping timeout: 246 seconds)16:04
*** alessioigor <alessioigor!~alessioig@> has quit IRC (Quit: alessioigor)16:06
*** alessioigor <alessioigor!~alessioig@> has joined #yocto16:06
JPEWRP: Ah, I think we are in luck; ACL on Linux are implemented using a system.posix_acl_access extended attribute; I think we can perhaps just query that to get the ACL instead of using libacl16:24
JPEWmm, except its some weird binary encoding16:28
*** alimon <alimon!~alimon@2806:10b7:2:b99c:2c32:cfff:fe8e:de1f> has quit IRC (Remote host closed the connection)16:30
yates_workcan someone please point me to the "gnome-layer"?16:33
JPEWyates_work: I think it's in meta-openembdded16:33
yates_workJPEW: thanks - let me have a look16:34
*** Haxxa <Haxxa!> has quit IRC (Quit: Haxxa flies away.)16:38
vvnDvorkinDmitry: either create a package and add it to your image with EXTRA_IMAGEDEPENDS for example, or pull the task in your image recipe directly with do_rootfs[mcdepends] += "mc:mach1:mach2:my-firmware:do_deploy" then add a function to ROOTFS_POSTPROCESS_COMMAND to copy the files where you want16:41
*** xmn <xmn!~xmn@2600:4040:9390:8c00:c65:2434:df3f:83e1> has quit IRC (Quit: ZZZzzz…)16:43
*** florian_kc <florian_kc!> has quit IRC (Ping timeout: 244 seconds)16:44
yates_workwhen you enter a layer in the BBLAYERS variable in bblayers.conf, does bitbake search that layer and all sublayers, or do you have to explicitly put the sublayer of interest there?16:45
yates_workdo i have to put this whole path? meta-openembedded/meta-oe/dynamic-layers/gnome-layer16:46
smurrayyates_work:  dynamic layers don't need to be explicitly added, just add meta-oe.  But the recipes under dynamic-layers/gnome-layer won't be visible unless meta-gnome is also in the stack16:47
SaurJPEW: If two different builds use a common hashequiv server, what happens if they do not share the sstate? I.e., if one builds a recipe and updates the HE server, what happens when the second build wants to build the same application and finds the information in the HE server, but the corresponding sstate information does not exist?16:47
JPEWIt will be an sstate miss an rebuild; I think it will populate sstate with the missing name16:48
SaurOk, so basically as if the HE server had not existed in the first place.16:48
JPEWWell, sort of. If other downstream hashes exist, it will be able to use them still16:49
*** goliath <goliath!~goliath@user/goliath> has joined #yocto16:49
yates_worksmurray: thanks.16:54
yates_workis the order in bblayers important?
*** alimon <alimon!~alimon@2806:10b7:2:b99c:2c32:cfff:fe8e:de1f> has joined #yocto16:56
*** michele_ <michele_!~michele@> has quit IRC (Read error: Connection reset by peer)16:59
yates_workswapping meta-oe and meta-gnome did not resolve the errors17:02
yates_workthis is under kirkstone, if it matters17:04
yates_workok got it17:11
yates_worki had to add several other layers under meta-openembedded.17:11
yates_workquestion about packages: when i build the default core-image-sato, it includes the dnf package manager. is there a way to point it to the built rpms under poky/build/tmp so that e.g. a new package can be installed with dnf?17:18
yates_workpretty sure the answer is yes. let me google dnf17:19
* khem checks RP's hack 17:23
khemI think this is best we can do in given scenario17:26
khemyates_work: use `bitbake-layers add-layer` command after you have checked out the layer repos17:30
*** Chaser <Chaser!~Chaser@user/chaser> has quit IRC (Quit: Chaser)17:32
khemlayerindex-fetch and layerindex-show-depends are also something you can use17:34
*** rty <rty!> has joined #yocto17:37
rtyhow do I preserve do_compile log for a recipe, even if it compiled successfully?17:37
*** Chaser <Chaser!~Chaser@user/chaser> has joined #yocto17:38
JaMarty: they are preserved by default in WORKDIR/temp/log.do_compile17:40
rtyonly if compilation failed17:41
rtyor maybe for more reasons. but if it finishes the recipe (like, packages, etc.), then do_compile is not there17:42
khemit wont be there if the package was used from sstate or rm-work was inherited in distro17:43
*** leon-anavi <leon-anavi!~Leon@> has quit IRC (Quit: Leaving)17:44
khemoh it might still be there with rm-work enabled17:44
*** leonanavi <leonanavi!~Leon@> has quit IRC (Remote host closed the connection)17:44
*** florian_kc <florian_kc!> has joined #yocto17:45
rtyI'll try with cleanall17:45
rtyI got the log after cleanall, thanks17:58
*** zpfvo <zpfvo!> has quit IRC (Quit: Leaving.)17:58
*** zpfvo <zpfvo!> has joined #yocto18:04
*** zpfvo <zpfvo!> has quit IRC (Remote host closed the connection)18:10
khem-ccompile might have done it too18:11
dvergatalJPEW: I already have a working code :D18:12
dvergatalJPEW: thx to you18:12
JPEWAh, me too18:12
dvergatalJPEW: ooo you have already applied it to the
dvergatalJPEW: or just an example?18:14
JPEWYa. I'll post it after lunch18:14
*** amitk_ <amitk_!~amit@> has quit IRC (Ping timeout: 256 seconds)18:17
*** amitk_ <amitk_!~amit@> has joined #yocto18:18
*** alessioigor <alessioigor!~alessioig@> has quit IRC (Quit: alessioigor)18:21
*** kanavin <kanavin!~Alexander@2a02:2454:29b:c600:5c43:16ff:3eac:2a77> has quit IRC (Remote host closed the connection)18:24
*** kanavin <kanavin!~Alexander@2a02:2454:29b:c600:5c43:16ff:3eac:2a77> has joined #yocto18:24
dvergatalJPEW: enjoy your meal18:26
*** GNUmoon <GNUmoon!~GNUmoon@gateway/tor-sasl/gnumoon> has quit IRC (Ping timeout: 246 seconds)18:31
*** GNUmoon <GNUmoon!~GNUmoon@gateway/tor-sasl/gnumoon> has joined #yocto18:34
*** rty <rty!> has quit IRC (Quit: Client closed)18:36
*** flom84 <flom84!~flom84@user/flom84> has joined #yocto18:51
JPEWRP: I put the ACL and XATTR support in bitbake, but if you want it in oe-core instead, let me know18:55
*** Chaser <Chaser!~Chaser@user/chaser> has quit IRC (Quit: Chaser)18:59
*** amitk_ <amitk_!~amit@> has quit IRC (Remote host closed the connection)19:00
*** Chaser <Chaser!~Chaser@user/chaser> has joined #yocto19:03
JPEWmmm, probably need to filter out ACL_USER_OBJ, ACL_GROUP_OBJ, and ACL_OTHER from the ACLs in the file; they are duplicated by the stat mode, and we skip some fields there19:05
JPEW(with the added benefit that most of the outhashes won't change if we do that because there aren't ACLs or xattrs)19:06
*** davidinux <davidinux!~davidinux@> has joined #yocto19:20
yates_workkhem: i am aware of that utility, but prefer to just type it in directly to the bblayers.conf - is there anything wrong with that?19:23
yates_worki.e., edit bblayers.conf and type19:23
*** flom84 <flom84!~flom84@user/flom84> has quit IRC (Ping timeout: 246 seconds)19:37
*** old_boy <old_boy!~old_boy@> has joined #yocto19:38
old_boyhow can I extend machine.conf in a different layer other than the meta-bsp layer?19:39
old_boyin my meta-bsp layer I have require which I don't want to do that in meta-bsp layer as those are distro specific, so wanted to know how can I add to machine.conf and from a different layer if that different layer is getting included...19:41
RPJPEW: at a quick glance the patches look like a good start, thanks!19:47
RPJPEW: I wasn't expecting to see patches, particularly so quickly :)19:48
RPkhem: I didn't really like the hack but I don't know what else we can do19:48
JPEWI got nerd sniped ;)19:48
*** Chaser <Chaser!~Chaser@user/chaser> has quit IRC (Quit: Chaser)19:51
*** xmn <xmn!~xmn@2600:4040:9390:8c00:219e:a991:ce72:1aaa> has joined #yocto19:56
RPJPEW: Again. Sorry! :)19:58
RPJPEW: it is an interesting problem19:58
JPEWEh, it's fine. I've always wondered how to interface with a library using ctypes19:58
RPJPEW: It isn't the most pleasant thing to do but not that hard either...19:59
*** old_boy <old_boy!~old_boy@> has quit IRC (Quit: Client closed)20:00
dvergatalRP: JPEW: I'm running it locally, probably it will finish tomorrow :P20:14
mischiefanyone using cve-check.bbclass? i think the usage of CVSS v3 scores may be broken. they show up as zeroes, at least in kirkstone.20:18
*** old_boy <old_boy!~old_boy@> has joined #yocto20:35
abelloniwe get check-layer-nightly failures for meta-oe20:45
RPSaur: did you have a chance to work out if my updates to the SRCPV changes address most of your concerns?21:18
* RP thinks the patch is clean on the autobuilder now21:19
*** Wouter0100670440 <Wouter0100670440!> has quit IRC (Quit: The Lounge -
*** Wouter0100670440 <Wouter0100670440!> has joined #yocto21:22
khemabelloni: I think its build node problem see21:31
khemyates_work: its fine to edit manually, but then you have to resolve layer dependencies manually21:32
abellonikhem: right but there is also the upstream-status errors or are they usually ignored?21:41
khemAssertionError: 432 != 0 : Found following patches with malformed or missing upstream status:21:43
khemthats just meta-oe and you have 10 other layers and no one fixes it. I can not do it all alone21:44
khemI have sent request to mailing list with details no one responded with single fix :)21:45
RPI think that is specific to that distro and only triggers when the layer has parsing errors21:48
khemthe error was due to libfaketime recipe patch that I had staged in master-next22:10
khemI have punted that now.22:10
khemThere has been some troublesome patches lately22:11
khemRP:  the issue although is due to sstate re-use and some host OS using unwinding and some not22:11
khemI never delved deep into the root cause though22:12
khemI think we should build host tools statically linked including toolchain for best results22:12
*** Xagen <Xagen!~Xagen@> has quit IRC (Ping timeout: 248 seconds)22:16
*** DvorkinDmitry <DvorkinDmitry!~dvorkin@> has quit IRC (Quit: KVIrc 5.0.0 Aria
yates_workfor a qemu/rpm package build, i've found a bunch of repomd.xml files in the tmp folder - which one do i point the qemu /etc/yum.repo.d/blah.repo metalink to?22:31
*** kpo <kpo!~kpo@> has quit IRC (Ping timeout: 245 seconds)22:32
*** kevinrowland <kevinrowland!~kevinrowl@> has joined #yocto23:09
kevinrowlandfolks, there's a file at `/etc/ssl/certs/ca-certificates.crt` in my rootfs and I don't know who put it there. `oe-pkgdata-util find-path` can't find it, and I can't find it by manually grepping `tmp/pkgdata` either. I don't spot anything in `poky/` that might generate it (as a rootfs post-processing step maybe?). does anyone know how it got there?23:10
*** Kubu_work <Kubu_work!> has quit IRC (Quit: Leaving.)23:14
*** belgianguy <belgianguy!> has joined #yocto23:24
kevinrowlandokay I found a postinst in `` which probably does it23:31
*** florian_kc <florian_kc!> has quit IRC (Ping timeout: 258 seconds)23:44
khemkevinrowland: right that postinst is run offline so its a post rootfs step23:53
kevinrowlandthanks for confirming khem :)23:54

Generated by 2.17.2 by Marius Gedminas - find it at!