This reverts commit 0696b0ef78.
Okay, now finally, let's get this straight. We actually *want*
preferLocalBuild, *because* we have improved the source splitup in
c92dbffeac.
The idea is to use local builds in order to prevent the source being
pushed to a remote machine, splitted up there (and thus copied again)
and then being copied *again* FROM the remote machine.
"DOH!" - as @edolstra or @rbvermaa would call it... and good d^Hnight.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This reverts commit 26f024626c.
I actually wasn't reading the "remove" in the commit message, so sorry
for the brainfart/noise.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This reverts commit fdb5cf8107.
The reason I'm reverting this is that the implications this had on the
IO load of Hydra are fixed by c92dbffeac.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
So far we've done the source code split up by using the generic
unpackPhase and copying it all over into the different outputs.
However, this had the problem of generating the I/O load of about three
times the size of the source tree: First at fetchurl of the tarball
(although it's not as much because it's compressed), second at
unpackPhase and third at installPhase.
Now we don't use installPhase anymore and directly unpack into the
output paths, which unfortunately becomes quite a bit more complex
because we need to transform the paths of the tar file on the fly.
I've also tried using GNU Tar's --to-command option to even untar *and*
patch it at the same time, but forking for every single file in the
tarball gets REALLY slow and also gets even more complex than this two
stage approach because you need to make sure that the patch file is
applied correctly, for example for files that don't yet exist but are to
be created by the patch file.
We're using --anchored and --no-wildcards-match-slash here to prevent
accidentally excluding files we don't want to exclude. One example is
something like v8/tools/gyp/v8.gyp.
So the current approach is some compromise between complexity and speed
and should hopefully get rid of the Hydra build timeouts by lowering I/O
load.
See here for examples of builds having this issue:
http://hydra.nixos.org/build/19045023http://hydra.nixos.org/build/19044973http://hydra.nixos.org/build/19044968http://hydra.nixos.org/build/19045019
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Overview of the updated versions:
stable: 40.0.2214.91 -> 40.0.2214.115
beta: 41.0.2272.16 -> 41.0.2272.64
dev: 41.0.2272.16 -> 42.0.2305.3
Introduces 42.0.2305.3 as the new dev version, which no longer requires
our user namespaces sandbox patch. Thanks to everyone participating in
https://crbug.com/312380 for finally having this upstream.
In the course of supporting the official namespace sandbox (that's what
the user namespace sandbox is called), a few things needed to be fixed
for version 42:
* Add an updated nix_plugin_paths.patch, because the old
one tries to patch the path for libpdf, which is now natively included
in Chromium.
* Don't copy libpdf.so to libexec path for version 42, it's no longer
needed as it's completely built-in now.
* Disable SUID sandbox directly in the source instead of going the easy
route of passing --disable-setuid-sandbox. The reason is that with
the command line flag a nasty nagbar will appear.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This brings a new stable version 40.0.2214.91 along with a beta update
to version 41.0.2272.16, the dev channel is still stuck at version
41.0.2272.12 and within the next days will jump to version 42.
For this reason, I've done some cheating here and brought the beta
channel in par with the dev channel, because dev is older than beta on
OmahaProxy.
Here's an overview of the channel upgrades:
stable: 39.0.2171.65 -> 40.0.2214.91 [1]
beta: 40.0.2214.10 -> 41.0.2272.16 [1] [2] [3]
dev: 41.0.2224.3 -> 41.0.2272.16 [1] [2] [3]
[1]: We needed to patch in locations of lib{pci,udev}.so, because
Chromium tries to load them at runtime. For version 41 startup will
fail if it is unable to load libudev, but it also has the advantage
that this fixes GPU detection using libpci in the stable version,
which in turn could fix a few bugs on NixOS.
[2]: The upstream Debian package for the binary plugins now uses XZ
compression for the enclosed data tarball.
[3]: Chromium 41 needs {shapshot,natives}_blob.bin in order to start up,
so let's cp it among with the .pak files to avoid adding a
conditional for version 40.
The release annoucement of the stable channel update can be found here:
http://googlechromereleases.blogspot.de/2015/01/stable-update.html
Note that this release contains 62 security fixes(!) and I'm hereby
apologizing for the delay of this update.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Writing the gid_map is already non-fatal, but the actual sandbox process
still tries to setresgid() to nogroup (usually 65534). This however
fails, because if user namespace sandboxing is present, the namespace
doesn't have CAP_SETGID at this point.
Fortunately, the effective GID is already 65534, so we just need to
check whether the target gid matches and only(!) setresgid() if it
doesn't.
So if someone would run a SUID version of the sandbox, it would still
work nonetheless without a negative impact on security.
Fixes#5730, thanks to @wizeman for reporting and initial debugging.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Introduces environment variables to set plugin base paths. The schema
for these is like NIX_CHROMIUM_PLUGIN_PATH_<N>. Where <N> is the path
type we want to change, the supported (full) variable names are:
* NIX_CHROMIUM_PLUGIN_PATH_ALL
* NIX_CHROMIUM_PLUGIN_PATH_PEPPERFLASH
* NIX_CHROMIUM_PLUGIN_PATH_FILEFLASH
* NIX_CHROMIUM_PLUGIN_PATH_PDF
* NIX_CHROMIUM_PLUGIN_PATH_FILE_EFFECTS
* NIX_CHROMIUM_PLUGIN_PATH_NACL
* NIX_CHROMIUM_PLUGIN_PATH_PNACL
* NIX_CHROMIUM_PLUGIN_PATH_WIDEVINE
Whereas NIX_CHROMIUM_PLUGIN_PATH_ALL is the plugin base path for every
path which is not set explicitly, so by setting ..._ALL and not setting
..._WIDEVINE, the widevine plugin will be searched in the directory
specified using ..._ALL.
Right now, the only plugin where this is used is widevine, and it still
doesn't properly work yet.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
With this commit, the following new upstream versions are introduced:
stable: 38.0.2125.101 -> 39.0.2171.65
beta: 39.0.2171.19 -> 40.0.2214.10
dev: 40.0.2182.3 -> 41.0.2224.3
We can now remove missing_alg_import.patch, because version 39 is nom
stable and thus fixes the missing include directive upstream.
However, starting with version 40, we hit a few bugs with system
protobuf, so we're disabling it for every version >=40 to avoid
runtime/startup errors.
Here is the stable channel announcement for version 39 on the official
blog:
http://googlechromereleases.blogspot.de/2014/11/stable-channel-update_18.html
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
With this commit, the following new upstream versions are introduced:
stable: 37.0.2062.94 -> 37.0.2062.120
beta: 38.0.2125.24 -> 38.0.2125.101
dev: 39.0.2138.3 -> 39.0.2171.7
All channels built fine on my machine and were tested against a few
sites.
The missing_alg_import.patch now is no longer needed for version 39, so
we're not applying it there anymore.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
With this commit, the following new upstream versions are introduced:
stable: 36.0.1985.125 -> 37.0.2062.94
beta: 37.0.2062.58 -> 37.0.2062.94
dev: 38.0.2107.3 -> 38.0.2125.8
All channels built fine on my machine and were tested against a few
sites.
Stable and beta channel now contain the same release, because version
37 hit the stable channel. For release notes, please have a look at the
announcement:
http://googlechromereleases.blogspot.de/2014/08/stable-channel-update_26.html
Of course we're also dropping all version 36 specific crap, such as the
architecture-specific target suffix for builds, which now is no longer
needed.
The gyp flag use_mojo=0 is no longer needed, as it was a workaround
concerning version 37.0.2054.3 only.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
beta: 37.0.2062.44 -> 37.0.2062.58 (builds fine, tested)
dev: 38.0.2101.0 -> 38.0.2107.3 (builds fine, tested)
Drop patch for fixing angle build for the dev version, because it was
applied upstream already.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
With this commit, the following new upstream versions are introduced:
stable: 35.0.1916.153 -> 36.0.1985.125
beta: 36.0.1985.84 -> 37.0.2062.44
dev: 37.0.2054.3 -> 38.0.2101.0
All builds were successfully tested on my machine, however in order to
update the beta and dev channels, a few additional modifications were
necessary:
* Don't update address_input_strings.grdp anymore because this has been
done/fixed upstream and was relevant in version 37.0.2054.3 _only_.
* No need to fix references to /usr/bin/gcc in version 38 anymore.
* Constrain patch for Angle (introduced in 4cbedd7) to version 37 only,
because it already has been applied upstream in version 38.
* Drop user namespaces patch for version 31 up until version 35,
because version 36 is already in stable.
* Don't try to build bundled Clang and/or even build using Clang.
* Remove obsolete patchPhase commands that are specific to version 35
and older.
While testing the dev version 38 I came accross a font rendering issue
which needs to be addressed ASAP (perhaps related to #3187), however the
browser works otherwise.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
After refactoring the updater we no longer did properly propagate the
exit code from the nix-prefetch-url call to the main script. So if the
newest version could not be fetched it didn't even bother to try the
previous release and we would end up with an empty hash.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
With this commit, the following new upstream versions are introduced:
stable: 35.0.1916.114 -> 35.0.1916.153
beta: 35.0.1916.86 -> 36.0.1985.67
dev: 36.0.1964.2 -> 37.0.2054.3
All builds successfully tested on my machine, however in order to update
the beta and dev channels, a few additional modifications were
necessary:
* Update/rebase USER_NS sandbox patch for version 36 and higher.
* Create address_input_strings.grdp before running gyp in version 37.
* Remove an empty string leftover from 0517041.
* Add patch for building bundled Angle for version 37.
The patch for Angle is to remove reliance on git being present during
build and is from https://chromium-review.googlesource.com/202048 but
with own modifications to remove/fix Windows-specific parts within the
patch file.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This fixes build for version 36, which i accidentally broke in commit
f6e31fadd8.
The reason this happened, was that my Hydra didn't pick up the latest
commit and I actually tested and built the parent commit instead of the
update commit.
So, this commit is the real "builds fine, tested" for all channels.
Also, the sandbox client initalization has moved into
setuid_sandbox_client.cc, so we need to move the lookup of the
CHROMIUM_SANDBOX_BINARY_PATH environment variable there.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
The system attribute was already there in the function head of the
shared update helper but it actually wasn't used and thus later the
import of <nixpkgs> was done using builtins.currentSystem instead of the
system attribute inherited from the source derivation.
Now we correctly propagate the attribute, so that even when running a
64bit kernel you can run a 32bit Chromium with binary plugins.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This is to ensure that nothing unexpected will be after the merge of the
refactoring branch, and also my own autoupdate machinery is expecting
this location, so there really is no reason to change it now.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
beta: 34.0.1847.60 -> 35.0.1916.47 (builds fine, tested)
dev: 35.0.1883.0 -> 36.0.1941.0 (builds fine, tested)
For the new version 36, we needed to rebase our user namespaces sandbox
patch, because http://crbug.com/312380 is preparing for an upstream
implementation of the same functionality.
Also, we need to add ply and jinja2 to the depends on version 36. This
is done unconditionally, because I want to avoid cluttering up the
expressions with various versionOlder checks.
The sandbox binary had to be fixed as well and we no longer use system
zlib, as - who might have guessed it - it's a fast moving target at
Chromium as well.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
It doesn't make sense to do the splitting of the source code on a remote
machine, so don't try to do it.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This now uses the Debian package from the sources derivation instead of
hardcoding it, so we finally should have proper PepperAPI plugin support
without crashing plugins and whatnot.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
This cases the Debian binaries to be fetched from Google's official APT
repository. If we aren't able to find a package from the APT repository,
it's very likely that it already got deleted upstream and we need to
fallback to mirrors instead.
Unfortunately, we can't use mirrors for updating, because Google doesn't
sign the Debian packages themselves and only the release files.
We're going to hook it into a Chromium updater soon, making the sha256
hashes publicly available, so if it is missing, we can still put the
sha256 manually into sources.nix, without risking anything by blindly
fetching from one of the provided mirrors.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
The updater is now splitted between a shellscript and a Nix expression
file which contains helpers and lookup functions to reconstruct all
information needed in order to fetch the source tarballs.
This means, that the sources.nix now doesn't contain URLs and only
versions and the corresponding SHA256 hashes. Of course, right now this
sounds like it's unnecessary, but we're going to fetch binaries soon so
it's a good idea to not unnecessarily clutter up sources.nix.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Next, we're going to refactor update.sh and the first step is to ensure
that we keep everything related to sources into its own subdirectory to
not clutter up the main directory too much.
Signed-off-by: aszlig <aszlig@redmoonstudios.org>
