[00:00] As long as they're not Xfce apps (seems to be the general consensus) [00:00] no, even if they are, they should support window managers that aren't xfce [00:01] but they should also support not-CSD [00:02] Right [00:05] It probably warrants an extended discussion and reconsideration of the issues... not sure how long ochosi is away, probably don't want to push in either direction until he reappears [00:09] Ah, from the email thread it seems GTK_CSD=0 should just revert it all to be classic Xfce? [00:09] Yep [00:09] So basically, just changing the default. If there's a non-CSD option, I don't care what the default is. [00:09] doesn't work on header bars, at least not on anything i have installed [00:10] In fact, I think it's logical to make the libxfce4ui headerbar implementation respect that [00:10] yes [00:11] Hrm, I purged catfish from everywhere, I can't test it. [00:11] although automatic conversion is still a bit problematic if you go full headerbars like eg gedit [00:11] cos then you have this big ugly toolbar and no menu bar [00:16] this is what headeraway does: https://i.imgur.com/HSUBgTq.png [00:17] screenshot is from before xfwm supported them at all and so you got two decorations [00:17] Gotcha [00:18] I could likely use that as I never found an alternative for simple-scan (I've replaced or purged every other CSD/headerbar using application.) [00:19] i doubt it still works without fix ups... it's a preload hack and i made it 6 years ago [00:20] So, in theory we could somehow plug that into libxfce4ui, and it would be available for Xfce components, or anything GTK3 in an Xfce session? [00:21] it would only work for xfce components that explicitly used it instead of using gtk header bar directly [00:21] Ah, gotcha. So the ideal usecase would still be using it in preload [00:22] also, confirmed it no longer works [00:22] Bummer. [00:22] has no effect on nautilus anyway [00:22] (skanlite pulls in too much KDE) [00:23] i wouldn't recommend shipping gtk preload hacks in a real distribution [00:24] No, it's not the best idea. I think Lubuntu tried that and it bit them. [00:24] third party stuff is somebody else's problem [00:24] i think ubuntu did it at one point :) [00:25] Oh my. [00:25] the hud/dbus menu thing [00:25] Which was my absolute favorite ubuntu feature [00:25] maybe im remembering wrong [00:25] is that a /s? [00:26] I really thought the hud was awesome [00:26] But yeah, that was a patched GTK that enabled that [00:26] it was awesome if you already memorized the contents of every menu in an app [00:27] discoverability for new users was terrible... worse than ribbon, or the thing that hid menu items that you rarely used [00:39] GTK_CSD=0 LD_PRELOAD=libgtk3-nocsd.so.0 /usr/bin/remmina looks off. :3 [00:40] i'm still on 18.04 so i can't really even test this stuff properly [00:40] all my xfce is still gtk2 [00:42] The latter bit is fixable. :3 [00:45] i'll probably install +1 after christmas [00:45] ten stay on it for two years :) [00:45] You *could* add the xubuntu-dev/staging PPA, but I doubt you like PPAs. [00:45] i love ppas [00:45] ...Welp. [00:46] i have that ppa installed already [00:46] and also experimental [00:46] Shouldn't you have all GTK3 then? [00:46] no idea [00:47] hmm.. my parole is gtk3 [00:48] but GTK_CSD has absolutely no effect on it [00:49] it's more the third party stuff... like my simple-scan has no CSD [00:49] That might have been patched out, lucky you. [00:49] neither does atril [00:49] We switched to Atril because evince did. [00:50] oh its a mate fork? [00:51] Yep. [00:51] engrampa, atril, mate-calc. === brainwash_ is now known as brainwash === brainwash_ is now known as brainwash [10:46] ali1234, Unit193, I created a new proposal that might help with the Xfce conversation. I've already shot this along to the core devs, but what do you think? https://wiki.xfce.org/releng/4.16/roadmap/general_ui/csd_support [12:59] bluesabre: works for me. can you add the frozen client bug to the list of problems too please? [12:59] this one: https://gitlab.gnome.org/GNOME/gtk/issues/215 [13:03] i think that one is important because it is not a subjective design issue, CSD is objectively worse in that situation. and because GNOME have declared it unfixable.