libera/#maemo-leste/ Wednesday, 2021-09-01

freemangordonuh, abook needs modest12:41
Wizzupheh12:41
Wizzupwhat for?12:41
freemangordonsend contact12:41
freemangordonmaybe I shall replace that with some xdg thing12:41
freemangordonwho was on modest? kona?12:42
freemangordonhmm, wait12:43
freemangordonactually this is in modest repo, but not the same package12:43
Wizzupok12:43
freemangordonthis https://github.com/community-ssu/modest/tree/master/libmodest-dbus-client12:43
WizzupI think porting modest might not be too much work initially, but it's been a while since I checked12:43
freemangordonmaybe I shall jump on that12:44
freemangordonit is also a good test-case for abook12:44
Wizzup*nod*12:46
freemangordonI wonder if I shall jump on whatever was done by that 'zero' guy on github12:46
* freemangordon checks12:47
freemangordonhmm, nullmark is not here12:48
WizzupI think his documentation is useful12:49
WizzupI don't think he did a lot of coding work12:49
Wizzupit's worth a read12:49
Wizzupit sounds like porting from gvfs to gio and then porting to the newer gtkhtml (3.8 to 4.0) would mostly get us there12:50
Wizzupfrom his research12:50
freemangordonI am reading through the readme and I am not sure I understand all of it12:50
freemangordonoh, we already have a repo12:50
freemangordongood12:50
Wizzuphttps://gitlab.gnome.org/Archive/gtkhtml12:51
Wizzupif this is fast and lightweight, it might be useful for conversations too12:51
freemangordonright12:51
freemangordonbut, can;t we embed some real web engine?12:51
freemangordonbut ok, lets get to it first12:52
freemangordontinymail is first12:52
Wizzupeventually we might not want gtkhtml, but I think for now porting to 4.x makes sense12:54
freemangordonok12:54
uvosfreemangordon: what dose abook need modest for exactly12:55
uvosfreemangordon: send contact means its sends a vcard via email or what?12:55
freemangordonyes12:55
freemangordonsend vcf via email12:56
freemangordonactually it needs libmodest-dbus-client, but...12:56
uvosso use xdg-email maybe?12:57
uvoshttps://linux.die.net/man/1/xdg-email12:57
uvosseams to work fine here on desktop12:58
freemangordonthis is async, right?12:58
freemangordonno sane resultcode I guess12:58
uvoshmm maybe its not terribly usefull12:59
freemangordonalso, it seems we cannot attach more than one file :(12:59
uvosit allows you to send a email12:59
Wizzupin modest, or?12:59
Wizzuphaving modest in any case would be great12:59
uvosin whatever is registerd as mailto mime12:59
freemangordonWizzup: I am non it12:59
uvosbut the client is required by spec to ask for confirmation12:59
uvos(ie open a window with the email writen for the user to click send on or a message box)13:00
uvosit dose work fine on d4 with trojita too13:00
freemangordonWizzup: shall I keep compatibility with fremantle?13:51
Wizzupfreemangordon: in what way?13:54
WizzupI'd argue no13:54
Wizzupif it's about glib defines or something13:55
freemangordonlike g_type_class_add_private being deprecated13:55
freemangordonbut I guess I can #ifdef that13:55
WizzupI'd remove that tbh13:55
freemangordonbut that would mean we'll need to remove support for anything else but leste from the build system13:56
freemangordonand this is harder TBH13:56
Wizzupup to you then :)13:56
freemangordonok :)13:56
freemangordonBTW, it builds fine, without html renderer so far13:57
Wizzupcheck13:58
Wizzupfreemangordon: I think it builds without gvfs then14:00
freemangordonsure14:02
freemangordonI modified the build scripts14:02
freemangordonI guess I shall create a new (gio) module that replaces gnomevfs14:02
freemangordonand port modest to that14:02
Wizzupright14:03
freemangordonok, not going to fix deprecations now14:06
freemangordonlets first have it up and running14:06
tmlindsicelo, uvos: pushed out also wlroots-0.14.1 branch that can be used with sway-1.6.1, see the pmos bug for more info at https://gitlab.com/postmarketOS/pmaports/-/issues/93214:18
tmlindno luck so far with current wlroots with weston-simple-egl with dmabuf instead of EXT_image_dma_buf_import like wlroots-0.14.1 is still using14:22
tmlindi wonder if there is some dmabuf issue also involved somewhere with the pvr_egl_shim?14:23
freemangordontmlind: hmm, what exactly does not work?14:24
tmlindfreemangordon: so wlroots uses EXT_image_dma_buf_import up to 0.14.1, then drops it and uses dmabuf14:25
freemangordondmebuf like BO?14:25
freemangordonor, what is dmabuf?14:25
freemangordonbo == gbm14:25
tmlindyeah i guess the buf comes via gbm14:26
freemangordonthis should work14:26
freemangordondo you have some error message or anything that demonstrates "does not work"?14:26
uvosglamor uses EXT_image_dma_buf_import i think14:28
* freemangordon checks14:28
tmlindok14:28
tmlindEXT_image_dma_buf_import works with earlier wlroots for weston-simple-egl14:29
tmlindweston-simple-dmabuf-egl i've never seen working and it produces some pvr blob error14:29
uvosok but ofc dmabuf not working might point to deeper issues14:29
tmlindright14:29
freemangordonI don;t think dmabuf is supposed to work at all, unless I am missing something, it is GBM that should work. IIUC/IIRC dmabuf is platform specific14:30
uvossure but the pvr driver must be mangining the gbm buffers in kernel somehow14:31
tmlindwell yeah non-working dmabuf is an issue :) so making sure glamor uses EXT_image_dma_buf_import might fix pvr_egl_shim?14:31
uvosshould be dambuf there no14:31
uvos(im not sure how pvrk works ofc)14:31
freemangordonglamor uses eglCreateImageKHR14:32
freemangordonto import gbm bo14:32
freemangordonthis *is not* dmabuf afaik14:32
freemangordonI don;t think pvr blobs support EXT_image_dma_buf_import14:33
tmlindoh sorry wrong property, just a sec14:33
freemangordonI don;t remember seeing it announced14:33
uvospretty sure its implemented in terms of dmabuf14:33
uvos(in kernel)14:33
freemangordonyou mean gbm?14:33
tmlindi meant what is known to work is using EGL_WL_bind_wayland_display14:33
uvosyeah14:33
freemangordonuvos: might be14:34
freemangordonbut from user POV those are different14:34
freemangordondma_buf uses proprietary ioctl codes to map/unmap etc14:34
freemangordonafter all everything should end up in the video memory14:34
uvosglamor certenly checks for EGL_EXT_image_dma_buf_import14:35
freemangordonBTW, gbm does not even allow you to mpa14:35
uvosbut if pvr dosent expose it its not useing it ofc14:35
freemangordonuvos: modesetting uses GBM, not DMA_BUF14:35
freemangordonso is glamor14:35
tmlindsorry i pasted wrong buffer above, so what i meant is wlroots up to 0.14.1 uses EGL_WL_bind_wayland_display for weston-simple-egl and that works, dmabuf for later wlroots does not work14:35
freemangordontmlind: I am confused14:36
tmlindyes sorry14:36
uvosglamor uses it for GBM_BO_WITH_MODIFIERS14:36
uvos(its not relevant just mentioning)14:36
freemangordonuvos: which is disabled14:36
uvosright14:36
freemangordonalso, GBM_BO_WITH_MODIFIERS is still not DMA_BUF API14:37
uvosno14:37
freemangordonlook at gbm.h14:37
tmlindi guess no way to test pvr_egl_shim with EGL_WL_bind_wayland_display? probably best to compare glamor usage against kmscube etc14:37
uvoskmscube uses gbm14:37
freemangordonright14:37
tmlindand is working14:37
uvostrivial usage of gbm tends to work ok14:38
freemangordonuvos: see https://github.com/robclark/libgbm/blob/master/gbm.h#L23714:39
freemangordonbo with modifiers is basically several planes, IIRC14:39
freemangordonbut it still a BO, not DMA_BUF14:39
tmlindjust in case, if glamor is using EXT_image_dma_buf_import, and sgx blobs do not provide EGL_EXT_image_dma_buf_import_modifiers, glamor might be setting wrong default dmabuf configuration14:40
uvosits not14:40
tmlindok14:40
uvosits just a optional path14:40
freemangordonglamor is using EGL_KHR_image_pixmap14:40
freemangordontmlind: I also wonder why wlroots (or whatever) does not use native WL buffers14:41
freemangordonthose are supported by the blob14:41
freemangordondo you have a link to the code?14:41
uvosits a compositor not a client14:41
freemangordonso<14:42
freemangordon?14:42
freemangordonit can still import WL buffer, not DMA_BUF buffer14:42
freemangordonmaybe I am missing the context14:42
uvosit wants to use egl14:42
uvosProbubly14:42
tmlindi think the plan is to make wlroots support more backends, it already supports wayland and x11 at least14:42
freemangordonis this the SW in question https://github.com/swaywm/wlroots/commits/master/xwayland ?14:44
tmlindthe wlroots commit that broke weston-simple-gl: https://github.com/swaywm/wlroots/commit/4e07d4cbf9c104625d419b9123dca0ef402472e714:44
uvosfreemangordon: it wants to run ON x1114:44
uvosfreemangordon: not the other way around14:44
uvoslike weston dose14:44
uvosor am i missunderstanding tmlind14:44
* freemangordon checks the commit14:45
tmlindyeah it can run on x11 or on wayland14:45
uvoson wayland? it is the comp. it runs on egl/gbm/kms14:46
tmlindso freemangordon: so is glamor doing the GL_CLAMP_TO_EDGE calls?14:46
uvosno14:46
freemangordonyes14:46
freemangordon:)14:46
uvoswe patched that out14:46
freemangordonno14:46
freemangordonit turned out it is another issue14:46
freemangordonso I reverted later on14:46
freemangordonjust a sec14:47
uvosoh ok14:47
tmlindthe CLAMP_TO_EDGE stuff is needed for sure, but i think the real reason wlroots broke is because renderer->egl->procs.eglQueryWaylandBufferWL no longer gets called, and maybe dmabuf now gets wrong params14:48
tmlindthen in addition to CLAMP_TO_EDGE, some extra trickery was needed for wlroots, see this commit from jonathan that i updated: https://github.com/tmlind/wlroots/commit/2ec83ab4f11b1f574d565bd04bbef2e55f154d7614:49
freemangordonuvos: actually, the issue was that GL_CLAMP_TO_EDGE was not set14:50
freemangordonand some insane default was used, leading to a broken rendering14:50
uvosok14:50
tmlindfreemangordon: also check glamor for GL_EXT_unpack_subimage usage, and compared to the wlroots commit 2ec83ab4f11b above14:50
uvosim honestly pretty hazy on all of this considering the time delta14:50
freemangordontmlind: sorry, can;t do this ATM, I am on modest :)14:51
tmlindfreemangordon: ok just few suggestions on what might cause the issues :)14:52
freemangordontmlind: glTexSubImage2D is not supported on KHRImage backed texture14:53
uvosGL_EXT_pack_subimage is optional in glamor14:55
freemangordongimme 5 minutes to grok the commit14:55
uvosglTexSubImage2D(GL_TEXTURE_2D, 0,14:55
uvos                                    x1 - box->x1, y1 - box->y1,14:55
uvos                                    x2 - x1, 1,14:55
freemangordonuvos: unfortunately no14:55
uvos                                    f->format, f->type,14:55
uvos                                    bits + ofs);14:55
uvoswell it checks for it and dosent bail14:55
freemangordonit tries to use it and thats the reason why it does not fully work with pvr blobs14:56
uvostmlind: glTexSubImage2D is not supported on KHRImage backed texture <-- your sure that isent valid14:56
freemangordonit segfaults in the blob14:56
uvos?14:56
tmlindGL_TEXTURE_2D needs to be handled line at a time if no GL_EXT_unpack_subimage14:56
freemangordonuvos: yes, lemme find the specs14:56
uvostmlind: thats whats happening thera14:57
uvosiirc14:57
uvosit goes for (; y1 < y2; y1++, ofs += byte_stride)14:57
uvos*iiuc rather14:58
uvosi tried compearing the wlroots changes to glamor before14:58
uvosand could not find what glamor is doing different/wrong14:58
tmlindok14:58
uvosand then i went and tried to check this statement "glTexSubImage2D is not supported on KHRImage backed texture" but came up blank15:00
uvossince you can look at the specs for all of these things but its pretty opaque what can be used with what15:00
freemangordon"There is no support for most of the functions that manipulate     other texture targets (e.g. you cannot use gl*Tex*Image*() functions with     TEXTURE_EXTERNAL_OES)"15:00
uvosfreemangordon: any luck on that?15:00
uvosfreemangordon: where is this?15:00
freemangordonhttps://www.khronos.org/registry/OpenGL/extensions/OES/OES_EGL_image_external.txt15:01
freemangordonthis is pretty fuzzy though15:01
freemangordonbecause we use glEGLImageTargetTexture2DOES, but with GL_TEXTURE_2D, not GL_TEXTURE_EXTERNAL_OES15:02
freemangordonhowever, using GL_TEXTURE_2D should fail on gles2, IIUC, but it does not with the blobs15:03
freemangordonbut, later on it segfaults in the blob when you try to glTexSubImage2D15:03
freemangordonI am not sure I can explain it any better15:04
freemangordonso, the workaround I was on is to create a texture backed FBO to copy the original texture to and to use glTexSubImage2D on that15:04
freemangordonthat was working pretty much ok on d4, however, when I tried on n900, it was hanging the whole device, and this is where I said f**k it and stopeed :)15:05
tmlindheh15:06
tmlindfreemangordon: care to post that workaround somewhere?15:07
freemangordontmlind: so, you you know how *exactly* wlroots fails?15:07
freemangordontmlind: it was a test program, lemme check if it is in the repo15:07
tmlindweston-simple-egl tries to use some invalid fd and gets: fcntl64(49282, F_DUPFD_CLOEXEC, 0)      = -1 EBADF (Bad file descriptor)15:08
tmlindthat's the dmabuf breakage, weston-simple-shm and weston-simple-damage work just fine as does sway15:08
tmlindso where's the segfault in pvr_egl_shim then? after the GL_TEXTURE_2D call?15:10
freemangordonI need to boot my d4, I think I have the latest sources there15:10
freemangordoncan't recall, sorry15:10
freemangordonoh, no, it was not in the shim15:10
freemangordonbut in glamor15:10
* tmlind has been building countless iterations on d4 of wlroots and sway versions to bisect the issues..15:10
freemangordonit is glamor that tries to use glTexSubImage2D15:11
tmlindok15:11
freemangordonsorry, have to do some RL work, ttyl15:11
tmlindlater15:11
tmlindno idea where weston-simple-egl pulls that bad fd from.. sounds like some bogus dmabuf config15:14
konaFreemangordon yes, currently porting tinymail to gio so I can build modest15:19
freemangordonkona: good, though I was able to build TM without this module, shall I push my changes?16:47
freemangordonI added leste as another platform which doesn;t build gnomevfs module16:48
konaSure, can you show me which repo you are using?16:54
konaIs it https://github.com/community-ssu/tinymail16:56
freemangordontmlind: https://github.com/freemangordon/pvr_egl_shim/blob/master/glamor_test_points.c16:58
freemangordonkona: sec16:58
freemangordonkona: https://github.com/maemo-leste/tinymail16:58
konaGreat!16:58
freemangordontmlind: this test program basically copies BO texture to a normal one and then does glTexSubImage2D, IIRC17:00
freemangordonkeep in mind I did that a while ago so I don;t remember the details17:00
freemangordonbut basically it should do more or less the same as what some code path in glamor does17:00
freemangordonoh, lemme push the makefile as well17:01
freemangordondone17:02
freemangordonkona: https://github.com/maemo-leste/tinymail/commit/31e878ce1817d7a3548654fc59910121813ae9b617:12
konaFreemangordon: thanks!17:15
uvosfreemangordon: i miirored it here http://uvos.xyz/maserati/testapps/17:16
freemangordonkona: keep in mind this is missing gnomevfs module17:18
uvosfreemangordon: ah you found it17:18
freemangordonmhm17:20
uvosso i gues doing the same thing in glamor_upload_boxes should make glamor work no?17:21
uvosi gues to be efficant you would need selectively do so depending on how the x11 Pixmap was created17:22
freemangordonright17:25
freemangordonkona: going to add wpeditor to leste repos, it is needed by modest17:28
lelfreemangordon created a repository: https://github.com/maemo-leste/wpeditor17:30
konafreemangordon: ok!17:45
konafreemangordon: is https://github.com/maemo-leste/gtk already in apt repos, or should I build myself?17:54
uvosthats in repos17:55
konathanks17:55
uvoskona: https://maedevu.maemo.org/pkgweb/17:56
uvosyou can look there for stuff17:56
freemangordonkona: you can use https://github.com/maemo-leste/wpeditor untill we add it to jenkins18:03
tmlindfreemangordon: thanks for pushing out the test program18:08
konafreemangordon: i am trying to build that branch now and configure reports No package 'gtk+-2.0' found18:09
konaoh, libgtk, *facepalm*18:10
* freemangordon is afk, bbl18:11
Wizzupkona: yeah check debian/control for package names you need18:12
uvosdisableing ofono via rc-update del with the cellular metapackage installed makes the maemo stack hang a short while after boot22:29
uvosh-d hangs22:30
uvosand you have to kill it22:30
uvosthen the status menu remains hanged and you have to kill that too22:30
uvosand you have to kill icd to get networking back22:30
uvosso is suspect the icd status menu item is at fault22:31
uvosthe fact that this failure mode is possible is concering22:35
uvosbut thats what you get when you make the terrible decision to make the status items plugins instead of seperate processies22:35
freemangordonhmm, modest needs calendar-ui-widgets-0-dev22:58
freemangordonI will have to RE that22:58
freemangordonth library that is22:59
freemangordonWizzup: what provides calendar applet?23:10
freemangordonoh, it is in extras23:14

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!