Opened 10 days ago

Closed 3 days ago

#26326 closed defect (fixed)

wine error when building Firefox ESR60 Windows x86_64

Reported by: sukhbir Owned by: tbb-team
Priority: Very High Milestone:
Component: Applications/Tor Browser Version:
Severity: Normal Keywords: ff60-esr, TorBrowserTeam201806
Cc: tom Actual Points:
Parent ID: #26203 Points:
Reviewer: Sponsor:

Description

When building Firefox ESR60 for Windows x86_64 (with https://github.com/azadi/tor-browser-build-1/tree/bug-26204 applied for the fxc2 project), we get the following error:

 0:03.48 Application tried to create a window, but no driver could be loaded.
 0:04.23 TelemetryHistogramData.inc
 0:04.23 nsSTSPreloadList.h
 0:04.36 TelemetryHistogramEnums.h
 0:04.36 TelemetryProcessData.h
 0:04.36 TelemetryProcessEnums.h
 0:04.37 TelemetryScalarData.h
 0:04.37 TelemetryScalarEnums.h
 0:04.46 wine: configuration in '/var/tmp/tmp.VbPk8hyCBs/.wine64' has been updated.
 0:04.48 /usr/bin/wine /var/tmp/dist/fxc2/bin/fxc2.exe -nologo -Tvs_4_0 /var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/mlgshaders/textured-vs.hlsl -ETexturedQuadVS -VnTexturedQuadVS -Vi -DVERTEX_SHADER -Fh/tmp/tmpF6j9gD
 0:04.48 Traceback (most recent call last):
 0:04.48   File "/usr/lib/python2.7/runpy.py", line 162, in _run_module_as_main
 0:04.48     "__main__", fname, loader, pkg_name)
 0:04.48   File "/usr/lib/python2.7/runpy.py", line 72, in _run_code
 0:04.48     exec code in run_globals
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/python/mozbuild/mozbuild/action/file_generate.py", line 117, in <module>
 0:04.48     sys.exit(main(sys.argv[1:]))
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/python/mozbuild/mozbuild/action/file_generate.py", line 68, in main
 0:04.48     ret = module.__dict__[method](output, *args.additional_arguments, **kwargs)
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/genshaders.py", line 27, in main
 0:04.48     return process_manifest(output_fp, input_filename)
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/genshaders.py", line 64, in process_manifest
 0:04.48     output_fp = output_fp)
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/genshaders.py", line 97, in run_fxc
 0:04.48     proc_stdout = subprocess.check_output(argv)
 0:04.48   File "/usr/lib/python2.7/subprocess.py", line 573, in check_output
 0:04.48     raise CalledProcessError(retcode, cmd, output=output)
 0:04.48 subprocess.CalledProcessError: Command '[u'/usr/bin/wine', u'/var/tmp/dist/fxc2/bin/fxc2.exe', '-nologo', '-Tvs_4_0', '/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/mlgshaders/textured-vs.hlsl', '-ETexturedQuadVS', '-VnTexturedQuadVS', '-Vi', '-DVERTEX_SHADER', '-Fh/tmp/tmpF6j9gD']' returned non-zero exit status 53
 0:04.48 /usr/bin/wine /var/tmp/dist/fxc2/bin/fxc2.exe -nologo -Tvs_4_0_level_9_3 /var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/CompositorD3D11.hlsl -ELayerQuadVS -VnLayerQuadVS -Vi -DVERTEX_SHADER -Fh/tmp/tmpwbm6Wj
 0:04.48 Traceback (most recent call last):
 0:04.48   File "/usr/lib/python2.7/runpy.py", line 162, in _run_module_as_main
 0:04.48     "__main__", fname, loader, pkg_name)
 0:04.48   File "/usr/lib/python2.7/runpy.py", line 72, in _run_code
 0:04.48     exec code in run_globals
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/python/mozbuild/mozbuild/action/file_generate.py", line 117, in <module>
 0:04.48     sys.exit(main(sys.argv[1:]))
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/python/mozbuild/mozbuild/action/file_generate.py", line 68, in main
 0:04.48     ret = module.__dict__[method](output, *args.additional_arguments, **kwargs)
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/genshaders.py", line 27, in main
 0:04.48     return process_manifest(output_fp, input_filename)
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/genshaders.py", line 64, in process_manifest
 0:04.48     output_fp = output_fp)
 0:04.48   File "/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/genshaders.py", line 97, in run_fxc
 0:04.48     proc_stdout = subprocess.check_output(argv)
 0:04.48   File "/usr/lib/python2.7/subprocess.py", line 573, in check_output
 0:04.48     raise CalledProcessError(retcode, cmd, output=output)
 0:04.48 subprocess.CalledProcessError: Command '[u'/usr/bin/wine', u'/var/tmp/dist/fxc2/bin/fxc2.exe', '-nologo', '-Tvs_4_0_level_9_3', '/var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/CompositorD3D11.hlsl', '-ELayerQuadVS', '-VnLayerQuadVS', '-Vi', '-DVERTEX_SHADER', '-Fh/tmp/tmpwbm6Wj']' returned non-zero exit status 53
 0:04.49 backend.mk:8: recipe for target 'CompositorD3D11Shaders.h' failed
 0:04.49 make[4]: *** [CompositorD3D11Shaders.h] Error 1
 0:04.49 make[4]: *** Deleting file 'CompositorD3D11Shaders.h'
 0:04.49 make[4]: *** Waiting for unfinished jobs....
 0:04.49 backend.mk:15: recipe for target 'MLGShaders.h' failed
 0:04.49 make[4]: *** [MLGShaders.h] Error 1
 0:04.49 make[4]: *** Deleting file 'MLGShaders.h'
 0:04.49 /var/tmp/build/firefox-30544586f8a9/config/recurse.mk:100: recipe for target 'gfx/layers/export' failed
 0:04.49 make[3]: *** [gfx/layers/export] Error 2
 0:04.49 make[3]: *** Waiting for unfinished jobs....
 0:21.46 /var/tmp/build/firefox-30544586f8a9/config/recurse.mk:32: recipe for target 'export' failed
 0:21.46 make[2]: *** [export] Error 2
 0:21.46 /var/tmp/build/firefox-30544586f8a9/config/rules.mk:434: recipe for target 'default' failed
 0:21.46 make[1]: *** [default] Error 2
 0:21.46 client.mk:168: recipe for target 'build' failed
 0:21.46 make: *** [build] Error 2
 0:21.47 0 compiler warnings present.

Other relevant information, proceeding this:

 0:02.27 Application tried to create a window, but no driver could be loaded.
 0:02.27 Make sure that your X server is running and that $DISPLAY is set correctly.

Child Tickets

Change History (16)

comment:1 Changed 10 days ago by sukhbir

We needed to run wine in headless mode, so I tried using xvfb:

+  Xvfb :1 &
+  export DISPLAY=:1

The only effect it seems to have is that it delays for a while and returns the same error:

 0:04.00 TelemetryHistogramData.inc
 0:04.07 TelemetryHistogramEnums.h
 0:04.11 TelemetryProcessData.h
 0:04.11 TelemetryProcessEnums.h
 0:04.11 TelemetryScalarData.h
 0:04.11 TelemetryScalarEnums.h
 5:01.26 /usr/bin/wine /var/tmp/dist/fxc2/bin/fxc2.exe -nologo -Tvs_4_0 /var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/mlgshaders/textured-vs.hlsl -ETexturedQuadVS -VnTexturedQuadVS -Vi -DVERTEX_SHADER -Fh/tmp/tmpEDeKFL
 5:01.26 /usr/bin/wine /var/tmp/dist/fxc2/bin/fxc2.exe -nologo -Tvs_4_0_level_9_3 /var/tmp/build/firefox-30544586f8a9/gfx/layers/d3d11/CompositorD3D11.hlsl -ELayerQuadVS -VnLayerQuadVS -Vi -DVERTEX_SHADER -Fh/tmp/tmpD35IoP

(Notice the gap between timestamps)

I tried with wine wineboot -i as well and that didn't work.

comment:2 Changed 10 days ago by gk

Cc: tom added
Keywords: TorBrowserTeam201806 added
Priority: MediumVery High
Status: newneeds_information

So, Tom got that working on Mozilla infra which I guess looks not that different. Tom, did you run into this when integrating the fxc2 hack?

comment:3 Changed 10 days ago by boklm

Did you try running the wine command that fails manually in the debug shell to see if that gives move details about the error?

comment:4 Changed 10 days ago by tom

I did not; but probably because we have some 'thing' that does display stuff for me so I never had to learn about xvfb or linux display stuff. =/ https://searchfox.org/mozilla-central/source/taskcluster/docker/recipes/xvfb.sh#55

comment:5 Changed 10 days ago by sukhbir

On further debugging, some progress:

export WINEDEBUG=+error
err:module:import_dll Library libwinpthread-1.dll (which is needed by L"Z:\\var\\tmp\\dist\\fxc2\\bin\\fxc2.exe") not found
err:module:LdrInitializeThunk Main exe initialization for L"Z:\\var\\tmp\\dist\\fxc2\\bin\\fxc2.exe" failed, status c0000135

I will update after fixing this.

comment:6 Changed 10 days ago by sukhbir

OK, the above was an easy fix (copying libwinpthread-1.dll) but the issue still persists. I will resume later but for anyone reading, I feel this is related to xvfb not starting, or not starting properly. Because when I hit the debug shell and run the command, it completes successfully and so does wine.

(If you see #comment:4, you will notice that the docker script says, "Up to 5 attempts will be made to start xvfb with a short delay # between retries", so clearly something is going on here.)

comment:7 Changed 10 days ago by gk

Status: needs_informationnew

comment:8 Changed 9 days ago by sukhbir

OK xvfb is done though I am not happy with it but that's for later. The next error:

 4:55.01 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:24:1: error: multiple storage classes in declaration specifiers
 4:55.01  static INLINE void mv_bias(int refmb_ref_frame_sign_bias, int refframe,
 4:55.01  ^~~~~~
 4:55.01 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:34:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE void vp8_clamp_mv2(int_mv *mv, const MACROBLOCKD *xd) {
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:48:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE void vp8_clamp_mv(int_mv *mv, int mb_to_left_edge,
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:60:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE unsigned int vp8_check_mv_bounds(int_mv *mv, int mb_to_left_edge,
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:86:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE uint32_t left_block_mv(const MODE_INFO *cur_mb, int b) {
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:98:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE uint32_t above_block_mv(const MODE_INFO *cur_mb, int b,
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:110:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE B_PREDICTION_MODE left_block_mode(const MODE_INFO *cur_mb,
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:128:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE B_PREDICTION_MODE above_block_mode(const MODE_INFO *cur_mb, int b,
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/config/rules.mk:773: recipe for target 'alloccommon.o' failed
 4:55.02  static INLINE B_PREDICTION_MODE left_block_mode(const MODE_INFO *cur_mb,
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/findnearmv.h:128:1: error: multiple storage classes in declaration specifiers
 4:55.02  static INLINE B_PREDICTION_MODE above_block_mode(const MODE_INFO *cur_mb, int b,
 4:55.02  ^~~~~~
 4:55.02 /var/tmp/build/firefox-30544586f8a9/config/rules.mk:773: recipe for target 'alloccommon.o' failed
 4:55.02 make[4]: *** [alloccommon.o] Error 1
 4:55.02 /var/tmp/build/firefox-30544586f8a9/config/recurse.mk:73: recipe for target 'media/libvpx/target' failed
 4:55.02 make[3]: *** [media/libvpx/target] Error 2
 4:55.02 make[3]: *** Waiting for unfinished jobs....

This is probably related to https://bugzilla.mozilla.org/show_bug.cgi?id=1378529

comment:9 Changed 9 days ago by sukhbir

Furthermore:

 5:08.66 /var/tmp/build/firefox-30544586f8a9/obj-mingw/mozilla-config.h:33:0: note: this is the location of the previous definition
 5:08.66  #define HAVE_PTHREAD_H 1
 5:08.66 
 5:08.67 In file included from /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vp8/common/generic/systemdependent.c:16:0:
 5:08.67 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vpx_ports/x86.h:113:1: error: multiple storage classes in declaration specifiers
 5:08.67  static INLINE uint64_t xgetbv(void) {
 5:08.67  ^~~~~~
 5:08.67 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vpx_ports/x86.h:166:1: error: multiple storage classes in declaration specifiers
 5:08.67  static INLINE int x86_simd_caps(void) {
 5:08.67  ^~~~~~
 5:08.67 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vpx_ports/x86.h:224:1: error: multiple storage classes in declaration specifiers
 5:08.67  static INLINE unsigned int x86_readtsc(void) {
 5:08.67  ^~~~~~
 5:08.67 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vpx_ports/x86.h:242:1: error: multiple storage classes in declaration specifiers
 5:08.67  static INLINE uint64_t x86_readtsc64(void) {
 5:08.67  ^~~~~~
 5:08.67 /var/tmp/build/firefox-30544586f8a9/media/libvpx/libvpx/vpx_ports/x86.h:307:1: error: multiple storage classes in declaration specifiers
 5:08.67  static INLINE unsigned int x87_set_double_precision(void) {
 5:08.67  ^~~~~~
 5:08.70 /var/tmp/build/firefox-30544586f8a9/config/rules.mk:773: recipe for target 'systemdependent.o' failed
 5:08.70 make[4]: *** [systemdependent.o] Error 1
 5:08.70 /var/tmp/build/firefox-30544586f8a9/config/recurse.mk:73: recipe for target 'media/libvpx/target' failed
 5:08.70 make[3]: *** [media/libvpx/target] Error 2

comment:10 Changed 9 days ago by sukhbir

The xvfb change above is pretty simple (in firefox/build);

 rm -f configure
 rm -f js/src/configure
 
+Xvfb :1 -screen 0 800x600x16 &
+export DISPLAY=:1.0

(I am not sure if the resolution is needed...)

I also made a few changes for wine:

+export WINEARCH=win64
+export HOME=/var/tmp/home
+mkdir -p $HOME
+WINEROOT=$HOME/.wine/drive_c

and

   windows-x86_64:
     var:
+      pre_pkginst: dpkg --add-architecture i386
       martools_filename: mar-tools-win64.zip
+      arch_deps:
+        - wine
+        - libwine:i386
+        - xvfb

This works ... mostly:

 2:01.68 err:process:__wine_kernel_init boot event wait timed out
 2:01.92 err:process:__wine_kernel_init boot event wait timed out
 4:01.75 err:process:__wine_kernel_init boot event wait timed out
 4:03.16 err:process:__wine_kernel_init boot event wait timed out

This doesn't seem to affect calling fxc2 though, because those commands complete successfully.

These will all go in the patch so you can ignore them for now but I am documenting them here in case you want to try the 32bit build or help improve the above.

comment:11 in reply to:  10 Changed 9 days ago by boklm

Replying to sukhbir:

The xvfb change above is pretty simple (in firefox/build);

 rm -f configure
 rm -f js/src/configure
 
+Xvfb :1 -screen 0 800x600x16 &
+export DISPLAY=:1.0

(I am not sure if the resolution is needed...)

Maybe you could use xvfb-run, with something like:

[% IF c("var/windows") %]xvfb-run[% END %] ./mach build

I also made a few changes for wine:

+export WINEARCH=win64
+export HOME=/var/tmp/home
+mkdir -p $HOME
+WINEROOT=$HOME/.wine/drive_c

and

   windows-x86_64:
     var:
+      pre_pkginst: dpkg --add-architecture i386
       martools_filename: mar-tools-win64.zip
+      arch_deps:
+        - wine
+        - libwine:i386
+        - xvfb

Do we need libwine:i386 for the 64bit build?

comment:12 Changed 6 days ago by sukhbir

I am now building on top of tor-browser-60.0.1esr-8.0-1 with tjr's and fxc2 patches. Here is what my mozconfig looks like:

diff --git a/projects/firefox/mozconfig-windows-x86_64 b/projects/firefox/mozconfig-windows-x86_64
index 669c124..7ae803c 100644
--- a/projects/firefox/mozconfig-windows-x86_64
+++ b/projects/firefox/mozconfig-windows-x86_64
@@ -20,10 +20,12 @@ ac_add_options --enable-verify-mar
 
 # We can't build the sandbox code with mingw-w64:
 # https://bugzilla.mozilla.org/show_bug.cgi?id=1042426
-ac_add_options --disable-sandbox
+#ac_add_options --disable-sandbox
 # Let's make sure no preference is enabling either Adobe's or Google's CDM.
 ac_add_options --disable-eme
 ac_add_options --disable-crashreporter
 ac_add_options --disable-maintenance-service
 ac_add_options --disable-webrtc
+ac_add_options --disable-stylo
 ac_add_options --disable-tests
+ac_add_options --disable-accessibility

Two errors:

(1):

 4:18.25 /var/tmp/build/firefox-dd0fed4bd34c/security/certverifier/Buffer.cpp: In function 'bool mozilla::operator==(const Buffer&, const Buffer&)':
 4:18.25 /var/tmp/build/firefox-dd0fed4bd34c/security/certverifier/Buffer.cpp:15:73: error: 'memcmp' was not declared in this scope
 4:18.25      (a.length() == b.length() && memcmp(a.begin(), b.begin(), a.length()) == 0);

Simple fix for this (did anyone else get this? I wonder why not...):

diff --git a/security/certverifier/Buffer.cpp b/security/certverifier/Buffer.cpp
index a36bfeabc13f..407d45a41253 100644
--- a/security/certverifier/Buffer.cpp
+++ b/security/certverifier/Buffer.cpp
@@ -5,6 +5,7 @@
  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
 
 #include "Buffer.h"
+#include <cstring>
 
 namespace mozilla {

(2):

22:11.45 update.locale
22:13.80 cp: cannot stat 'instgen/helper.exe': No such file or directory
22:13.80 /var/tmp/build/firefox-dd0fed4bd34c/toolkit/mozapps/installer/windows/nsis/makensis.mk:99: recipe for target 'uninstaller' failed

I am guessing this is #26205. EDIT: Sorry, #26319.

Last edited 6 days ago by sukhbir (previous) (diff)

comment:13 in reply to:  12 Changed 6 days ago by tom

Replying to sukhbir:

Two errors:

(1):

 4:18.25 /var/tmp/build/firefox-dd0fed4bd34c/security/certverifier/Buffer.cpp: In function 'bool mozilla::operator==(const Buffer&, const Buffer&)':
 4:18.25 /var/tmp/build/firefox-dd0fed4bd34c/security/certverifier/Buffer.cpp:15:73: error: 'memcmp' was not declared in this scope
 4:18.25      (a.length() == b.length() && memcmp(a.begin(), b.begin(), a.length()) == 0);

Simple fix for this (did anyone else get this? I wonder why not...):

diff --git a/security/certverifier/Buffer.cpp b/security/certverifier/Buffer.cpp
index a36bfeabc13f..407d45a41253 100644
--- a/security/certverifier/Buffer.cpp
+++ b/security/certverifier/Buffer.cpp
@@ -5,6 +5,7 @@
  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
 
 #include "Buffer.h"
+#include <cstring>
 
 namespace mozilla {

Firefox uses a unified build system that smashes several .cpp files into a single one. If you have any local patches that edit any file in the directory, it's likely it changed the order of unified builds and thus it lost the include that it was getting from another file. We generally don't pay much mind to these types of errors but feel free to send a patch upstream.

comment:14 in reply to:  12 Changed 6 days ago by gk

Replying to sukhbir:

(2):

22:11.45 update.locale
22:13.80 cp: cannot stat 'instgen/helper.exe': No such file or directory
22:13.80 /var/tmp/build/firefox-dd0fed4bd34c/toolkit/mozapps/installer/windows/nsis/makensis.mk:99: recipe for target 'uninstaller' failed

I am guessing this is #26205. EDIT: Sorry, #26319.

You are actually right with #26205. At least that's the ticket I had in mind for this issue. Could be that #26319 is actually a duplicate but I am not sure but IIRC the bug occurred even before the packaging step. I just modified the Makefile to work around that issue back then which is okay for now I think even though we might want to have a proper patch eventually.

comment:15 in reply to:  12 Changed 6 days ago by gk

Replying to sukhbir:

I am now building on top of tor-browser-60.0.1esr-8.0-1 with tjr's and fxc2 patches. Here is what my mozconfig looks like:

diff --git a/projects/firefox/mozconfig-windows-x86_64 b/projects/firefox/mozconfig-windows-x86_64
index 669c124..7ae803c 100644
--- a/projects/firefox/mozconfig-windows-x86_64
+++ b/projects/firefox/mozconfig-windows-x86_64
@@ -20,10 +20,12 @@ ac_add_options --enable-verify-mar
 
 # We can't build the sandbox code with mingw-w64:
 # https://bugzilla.mozilla.org/show_bug.cgi?id=1042426
-ac_add_options --disable-sandbox
+#ac_add_options --disable-sandbox
 # Let's make sure no preference is enabling either Adobe's or Google's CDM.
 ac_add_options --disable-eme
 ac_add_options --disable-crashreporter
 ac_add_options --disable-maintenance-service
 ac_add_options --disable-webrtc
+ac_add_options --disable-stylo
 ac_add_options --disable-tests
+ac_add_options --disable-accessibility

Note that right now enabling the sandbox for 64bit Windows will crash, see: https://bugzilla.mozilla.org/show_bug.cgi?id=1461421 and https://bugzilla.mozilla.org/show_bug.cgi?id=1466192. So, I guess we leave this off for now until either of those bugs is fixed (I think having the sandbox on trumps having jemalloc used if we need to make a choice for now).

comment:16 Changed 3 days ago by gk

Resolution: fixed
Status: newclosed

The wine error is solved with commit 57693e01837f0fc7c60bb056e5b752f5f67a987b on tor-browser-build's master branch.

Note: See TracTickets for help on using tickets.