From: patrick keshishian Subject: Re: mozilla.port.mk: no --enable-official-branding for Tor Browser To: Caspar Schutijser Cc: Tim van der Molen , ports@openbsd.org Date: Sat, 8 Aug 2026 16:30:23 -0700 Hi, On Sat, Dec 27, 2025 at 9:06 AM Caspar Schutijser wrote: > > On Mon, Dec 22, 2025 at 04:42:06PM +0100, Caspar Schutijser wrote: > > On Fri, Dec 19, 2025 at 01:55:32PM +0100, Tim van der Molen wrote: > > > > > > 7.7 ATM. I will try to update to -current this coming week. > > > > > > > > > > > > > > > Sorry for the long delay. Took me longer to move to -current than > > > > > I had anticipated. > > > > > > > > > > I can confirm that the issue of the file-chooser window size > > > > > does not exist in -current, as you mentioned. > > > > > > > > > > > > > > > > I just wanted to point out that the file-chooser window size > > > > issue seems to have returned. > > > > > > > > At least the last two updates I've done (sysupgrade -s followed > > > > by pkg_add -u) > > > > > > > > Currently running: > > > > > > > > 7.8-current (GENERIC.MP) #155: Mon Dec 15 10:01:15 MST 2025 > > > > > > > > $ pkg_info | grep tor- | cut -d ' ' -f1 > > > > tor-0.4.8.21 > > > > tor-browser-15.0.3 > > > > > > > > Not a huge deal, as I'm not relying on tor-browser, but thought > > > > it was worth mentioning. > > > > > > I'm also seeing this. It started after the latest tor-browser update. > > > > Thanks for the heads-up. I see it with 15.0.3 too. I'll now build the > > previous version (15.0.1) again to see whether the problem indeed does > > not appear with that version. If so, I'll see if I can spot anything > > in the diff between 15.0.1 and 15.0.3. > > 15.0.1 also shows the problematic behavior with the file chooser dialog > on my system. Just to be sure, I had a brief look at the diff between > 15.0.1 and 15.0.3 (which is bigger than one would expect) but I didn't > immediately spot a change that would be obviously responsible in there. > So not sure what is going on. > > I now doubt it is really caused by a code change between 15.0.1 > and 15.0.3. Rather, I think "something else" causes this behavior, > and we were just lucky that the behavior was gone for a while. > > Not sure when I have time for this, but I plan to dig around in > the diff between Tor Browser 15.0.3 and the corresponding > Firefox version, and see if I can find out if there is > a change that upstream made to Firefox that could be the cause. > > If someone is willing to help, that would be appreciated. Here are some > pointers: > $ git clone https://gitlab.torproject.org/tpo/applications/tor-browser.git # Warning, big repository > $ git log|shortlog|diff FIREFOX_140_6_0esr_BUILD1..tor-browser-140.6.0esr-15.0-1-build5 > > If people have other ideas for where to look, that would be appreciated > too. > > Caspar The file-chooser window problem has been persistent for some time now. I have been updating the snapshot roughly once a week (via `sysupgrade -s' followed by `pkg_add -u'). The update on July 31st continued to demonstrate the file-chooser issue. I just did a `sysupgrade -s' and `pkg_add -u' (8/8) not expecting anything different, but noticed the file-chooser in tor-browser is now in correct size/configuration, however, the font glyphs are displayed as unknown glyphs (I assume that is what the rectangles represent -- see attached screen grab). This is the `pkg_add -u' output: # pkg_add -u quirks-7.215:updatedb-0p0->0p0: ok quirks-7.215 signed on 2026-08-07T21:48:48Z quirks-7.214->7.215: ok adwaita-icon-theme-50.0:py3-packaging-26.2->26.3: ok adwaita-icon-theme-50.0:glib2-2.88.2->2.88.3: ok adwaita-icon-theme-50.0:shared-mime-info-2.5.1->2.5.1: ok adwaita-icon-theme-50.0:gdk-pixbuf-2.44.7->2.44.7: ok adwaita-icon-theme-50.0:cairo-1.18.4->1.18.4: ok adwaita-icon-theme-50.0:.libs-harfbuzz-14.1.0+harfbuzz-14.2.1->harfbuzz-14.3.0 (adwaita-icon-theme-50.0:.libs-harfbuzz-14.1.0+harfbuzz-14.2.1->harfbuzz-14.3.0: ok adwaita-icon-theme-50.0:pango-1.58.0->1.58.2: ok adwaita-icon-theme-50.0:librsvg-2.62.3p0v0->2.62.3p0v0: ok adwaita-icon-theme-50.0->50.0: ok .libs1-aom-3.9.1p0+aom-3.13.3->aom-3.14.1: ok at-spi2-core-2.60.5->2.60.6: ok dconf-0.40.0p2->0.40.0p2: ok desktop-file-utils-0.28p0->0.28p0: ok ffmpeg-8.1.2v1:.libs-x265-3.6+.libs-x265-4.1p0+x265-4.2->x265-4.3: ok ffmpeg-8.1.2v1:libass-0.17.5->0.17.5: ok .libs-ffmpeg-6.1.3v1+.libs1-ffmpeg-4.4.5v1+ffmpeg-8.1.2v1->ffmpeg-8.1.2v1 (proce.libs-ffmpeg-6.1.3v1+.libs1-ffmpeg-4.4.5v1+ffmpeg-8.1.2v1->ffmpeg-8.1.2v1: ok gnuplot-5.4.8p0->5.4.8p0: ok gtk+3-3.24.52->3.24.52: ok league-fonts-0.1p3:league-script-number-one-1.001pre1p0->1.001pre1p0 (processing.libs1-tb-browser-14.5.8+tb-browser-15.0.19->tb-browser-15.0.19: ok tb-noscript-13.6.31.1984->13.6.31.1984: ok tor-browser-15.0.19->15.0.19: ok Running tags: ok Read shared items: ok # echo $? 0 Just a reminder that Stuart Henderson's post on a similar topic hinted at a pango related cause: https://marc.info/?l=openbsd-misc&m=173689221602591&w=2 I see pango updated in `pkg_add -u' output. I hope some of this provides some hints as to the source of the problem. Thanks for reading, --patrick Following is a dmesg diff from July 31st and Aug 8th snapshots: --- 20260731-extract.txt Sat Aug 8 15:23:47 2026 +++ 20260808-extract.txt Sat Aug 8 15:24:06 2026 @@ -1,7 +1,7 @@ -OpenBSD 8.0-beta (GENERIC.MP) #73: Fri Jul 31 09:46:32 MDT 2026 +OpenBSD 8.0-beta (GENERIC.MP) #83: Sat Aug 8 10:51:27 MDT 2026 deraadt@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP real mem = 16725286912 (15950MB) -avail mem = 16194736128 (15444MB) +avail mem = 16194732032 (15444MB) random: good seed from bootblocks mpath0 at root scsibus0 at mpath0: 256 targets @@ -29,7 +29,7 @@ mtrr: Pentium Pro MTRR support, 8 var ranges, 88 fixed cpu0: apic clock running at 199MHz cpu0: mwait min=64, max=64, IBE cpu1 at mainbus0: apid 1 (application processor) -cpu1: AMD E-350 Processor, 1596.65 MHz, 14-01-00, patch 05000029 +cpu1: AMD E-350 Processor, 1596.66 MHz, 14-01-00, patch 05000029 cpu1: smt 0, core 1, package 0, type P ioapic0 at mainbus0: apid 2 pa 0xfec00000, version 21, 24 pins, remapped acpimcfg0 at acpi0 @@ -147,166 +147,3 @@ radeondrm0: PALM radeondrm0: 1366x768, 32bpp wsdisplay0 at radeondrm0 mux 1: console (std, vt100 emulation), using wskbd0 wsdisplay0: screen 1-5 added (std, vt100 emulation) -drm:pid38288:evergreen_packet3_check *WARNING* vbo resource seems too big (524289) for the bo (524288)