mash: chrome crashes on startup, missing chrome_200_percent.pak |
|||||||||
Issue descriptionSeen on minnie, which is odd because it does not have a double-density display. [1:1:0906/123144:ERROR:memory_mapped_file.cc(52)] Couldn't open /home/chrome/chrome_200_percent.pak [1:1:0906/123144:ERROR:data_pack.cc(79)] Failed to mmap datapack ... [15423:15423:0906/123144:ERROR:memory_mapped_file.cc(52)] Couldn't open /home/chrome/chrome_200_percent.pak [15423:15423:0906/123144:ERROR:data_pack.cc(79)] Failed to mmap datapack ... [15423:15423:0906/123144:FATAL:resource_bundle.cc(140)] Check failed: false. Unable to load image with id 7383, scale=2 This doesn't happen if I manually copy over chrome_200_percent.pak. Either deploy_chrome needs to copy over that file, or we need to figure out why chrome is loading 200% resources on a 100% device.
,
Sep 6 2016
A platform that supports high DPI should have 200 assets as you can connect high DPI displays. https://cs.chromium.org/chromium/src/ui/base/resource/resource_bundle.cc?rcl=0&l=666
,
Sep 6 2016
It sounds like the deploy_chrome script isn't copying over chrome_200_resources.pak. That's really odd. (sdk veyron_minnie R55-8774.0.0) jamescook@rubella2 /w/chrome/src $ deploy_chrome --build-dir=out_veyron_minnie/Release --nostartui --mash --to=172.18.37.37 --target-dir=/home/chrome 12:27:42: INFO: Staging flags taken from USE in environment: gn 12:27:42: INFO: Testing connection to the device... 12:27:51: INFO: Copying Chrome to /home/chrome on device... Warning: Permanently added '172.18.37.37' (RSA) to the list of known hosts. sending incremental file list ./ chrome chrome-sandbox chrome-wrapper chrome_100_percent.pak ***** It should be here, but isn't ****** chrome_manifest.json chrome_mash_manifest.json I don't see any recent changes in //third_party/chromite/scripts that would explain a change in behavior.
,
Sep 6 2016
I changed the deploy code recently, this could be an artifact of that, I will investigate.
,
Sep 16 2016
I am hitting this problem as well. When you hit this error, are you using the '--target-dir=' flag to deploy_chrome (I am)?
,
Sep 16 2016
,
Sep 16 2016
Yeah, I think I was using --target-dir=/home/chrome (you can see /home/chrome in the paths above)
,
Sep 22 2016
Issue 649182 has been merged into this issue.
,
Sep 22 2016
oshima@, continuing discussion from https://chromium-review.googlesource.com/#/c/386417/ : Simple chrome extracts the USE environment from the chroot that generated the chromeos image being used to build chrome from Chrome OS. 'highdpi' appears to default to true in the ebuild so it should be set there. I am working on some GN chromite cleanup anyway, so I will look into this today.
,
Sep 22 2016
highdpi flag is deprected. I thought it's removed. With 4K display, all devices should have high dpi assets.
,
Sep 22 2016
we have at least one system setting USE=-highdpi (puppy). that should get cleaned up first.
,
Sep 22 2016
stevenjb@ -- does b/31636193 look like a dupe of this? Thanks
,
Sep 22 2016
Seems likely.
,
Sep 23 2016
Re:#13 so what's the plan, do we want to just remove the cond=C.StagingFlagSet(_HIGHDPI_FLAG)) check in chrome_util.py?
,
Sep 23 2016
re:#14 I have verified that by removing that condition in chrome_util.py, I can see chrome_200_percent.pak gets included in the "sending incremental file list" when running deploy_chrome from an incremental SimpleChrome build for samus. So if oshima@'s comment in #10 is correct we should just remove the HIGHDPI flag references in this file.
,
Sep 23 2016
,
Sep 23 2016
,
Sep 23 2016
Sorry, I should have updated this. I have a fix: https://chromium-review.googlesource.com/#/c/388030/ There was a bug in the deploy_chrome logic from the GN migration. That code is no longer needed so I removed it and verified that the highdpi flag is getting set. Hopefully the CQ will behave today and this will land (I will chump it if I need to), then we can roll chromite and pick up this + sadrul's change.
,
Sep 26 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/chromite/+/4a6b53c6f2960a61bf8d2686244a0bc6aa1e1737 commit 4a6b53c6f2960a61bf8d2686244a0bc6aa1e1737 Author: Steven Bennetts <stevenjb@chromium.org> Date: Thu Sep 22 20:13:40 2016 cros_chrome_sdk: Elim no longer needed (and incorrect) USE parsing BUG= chromium:644408 TEST=deploy_chrome includes high dpi assets Change-Id: I5bafa56846008029149c170a8f644617d01bdf83 Reviewed-on: https://chromium-review.googlesource.com/388030 Tested-by: Steven Bennetts <stevenjb@chromium.org> Reviewed-by: Haixia Shi <hshi@chromium.org> Reviewed-by: Mike Frysinger <vapier@chromium.org> [modify] https://crrev.com/4a6b53c6f2960a61bf8d2686244a0bc6aa1e1737/cli/cros/cros_chrome_sdk.py
,
Sep 27 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/chromite/+/f9fd92d7cb31bbdefaf3294523c1a746e3465922 commit f9fd92d7cb31bbdefaf3294523c1a746e3465922 Author: Steven Bennetts <stevenjb@chromium.org> Date: Tue Sep 27 15:30:38 2016 SimpleChrome: deploy_chrome: Make libwidevine optional BUG= chromium:644408 ,chromium:612248 TEST=SimpleChromeWorkflow works Change-Id: Idc2906afb52dfa739788cab9f9d38d5cc14484b5 Reviewed-on: https://chromium-review.googlesource.com/390012 Reviewed-by: Steven Bennetts <stevenjb@chromium.org> Reviewed-by: Luis Hector Chavez <lhchavez@google.com> Tested-by: Steven Bennetts <stevenjb@chromium.org> [modify] https://crrev.com/f9fd92d7cb31bbdefaf3294523c1a746e3465922/lib/chrome_util.py
,
Sep 27 2016
Any idea when chromite will roll? I hit the 200% assets thing again yesterday afternoon.
,
Sep 27 2016
Working on it now... should have happend yesterday but there were CQ problems.
,
Sep 28 2016
,
Dec 9 2016
ChromeOs 8872.70.0/55.0.2883.87
,
Feb 26 2018
|
|||||||||
►
Sign in to add a comment |
|||||||||
Comment 1 by msw@chromium.org
, Sep 6 2016