**** BEGIN LOGGING AT Sun Feb 04 03:00:01 2018 Feb 04 07:14:01 Are Wikiwide and Oksana the same person? Or are they related? Feb 04 07:14:25 brolin_empey: yes, Wikiwide is Oksana Feb 04 07:14:43 Why? Feb 04 07:16:19 Anyway, Wikiwide is no longer here but I wanted to tell them that they were actually discussing touch panel technology, not display technology. A touch panel is often used without a display, such as for a (capacitive) trackpad. Feb 04 07:16:48 brolin_empey: she isnt around right now but she will read your message soon, oksana is connected but away Feb 04 07:17:37 ah, yeah interesting Feb 04 07:17:45 Touch is often confused with touch Feb 04 07:17:49 displays* Feb 04 07:18:10 there's touch tablets too Feb 04 07:23:22 Juesto: ACK. Feb 04 10:54:46 http://talk.maemo.org/showthread.php?t=100192 Feb 04 11:36:00 Wizzup: was about to ask about this new devuan n900 test etc Feb 04 11:36:05 * enyc looks Feb 04 11:36:24 oooooooooooooh bugtrackers and all Feb 04 11:46:25 Wizzup: to what extent is this system funcitonaly-limited by the n900 ram size? Can stuff be built with "thumb2" isa similarly? I would like to suggest using the same "flopswap" (on sd-card, not wearing out internal-memory!) system as is available in cssu... Feb 04 12:17:36 enyc: yes, thumb can be done Feb 04 12:17:46 enyc: we also use zram Feb 04 12:44:33 Wizzup: 'can be done' -> whats' involved to amke this work reliably for most package builds, though? Feb 04 12:45:54 enyc: turning a knod in the autobuilder Feb 04 12:46:00 *knob Feb 04 12:46:54 freemangordon: the 'power' kernel required/included tweaks/workarounds to make thumb2 work reliably, does the 4.15/whatever kernel have the same workarounds actually enabled? Feb 04 12:47:35 enyc: thumb errata workaround was upstreamed long time ago, so yes. Feb 04 12:47:57 excellent Feb 04 12:48:18 (i got the mirpession things like that had sometimes to be 'enabled', not always on-by-default in kconfig .... Feb 04 12:48:30 because, e.g. sometimes causes some slowdown, or that kind of thing Feb 04 12:48:36 http://lists.infradead.org/pipermail/linux-arm-kernel/2013-March/159347.html Feb 04 12:48:53 well, it has to be enabled in kernel config Feb 04 12:49:26 but iirc it is enabled by default in omap2plus_defconfig Feb 04 12:49:46 and we have it enabled in config we're using to build n900 kernel Feb 04 12:50:44 enyc: https://github.com/maemo-leste/n9xx-linux/blob/pvr-wip/arch/arm/configs/rx51_defconfig#L498 Feb 04 12:54:38 Linux/arm 4.15.0-rc6 Feb 04 12:55:02 mhm Feb 04 12:55:33 that was the last one 3 weeks ago :) Feb 04 12:56:02 hrrm... KPTI on armhf nedeed I wonder? Feb 04 12:56:10 this new image, is armhf in nature, rather than armel ??? Feb 04 12:56:26 yes, armhf, but we'll build armel soon Feb 04 12:56:38 whys' armel wanted/needed ? Feb 04 12:56:43 again, a knob in jenkins Feb 04 12:56:57 enyc: for ABI compatibility with fremantle binaries Feb 04 12:57:15 freemangordon: hrrm, which means duplicate libraries and more ram/disk and all that, on an install? Feb 04 12:57:24 [can you multiarch armhf and armel on same install??] Feb 04 12:57:31 why should we? Feb 04 12:57:39 we'll have separate images Feb 04 12:58:14 i thought, armhf faster, use cpu features properly, and all that.... whcih for 'new' system makes sense... Feb 04 12:58:27 the problem is, if you have binary-only fremantle binaries you want bincompat with? Feb 04 12:58:46 I would think, multiarch (might) be a useful compromise Feb 04 12:58:59 to get 'best of both worlds' Feb 04 12:59:35 I don;t think it is a good idea on 256 MB RAM device Feb 04 12:59:48 kk, for reason I meantioned (duplicates in ram and all) Feb 04 12:59:53 yeah Feb 04 13:00:03 does that mean duplicate C-ilbabry and all? Feb 04 13:00:10 afaik yes Feb 04 13:00:28 thus, we'll provide separate armel image Feb 04 13:01:55 ... but the armhf image will allow installing the armel multiarch, if desired (e.g. to run some particular application, even if not 'core system' etc...) ?? Feb 04 13:03:02 enyc: right now there are bot armhf and armel binaries in the repos Feb 04 13:03:06 *both Feb 04 13:03:14 right, so it should be do-able, in theory Feb 04 13:03:18 mhm Feb 04 13:10:20 Question is, what ''blockers'' in core-maemo-system are there, that will need armel? Feb 04 21:08:03 all the closed blobs basically **** ENDING LOGGING AT Mon Feb 05 03:00:02 2018