Age | Commit message (Collapse) | Author | Files | Lines |
|
Fixes #1173.
|
|
Also templatize readInt() to work for various integer types.
|
|
http://hydra.nixos.org/build/49490928
|
|
|
|
|
|
This is necessary because we have multiple SQLite databases (e.g. the
binary cache cache).
|
|
"build-max-jobs" and the "-j" option can now be set to "auto" to use
the number of CPUs in the system. (Unlike build-cores, it doesn't use
0 to imply auto-configuration, because a) magic values are a bad idea
in general; b) 0 is a legitimate value used to disable local
building.)
Fixes #1198.
|
|
|
|
We set build-cores automatically, not build-max-jobs. (The commit
message for de4cdd0d47adc70a4db12397a42c18ee50b4e662 also got this
wrong.)
|
|
Issue #1254.
|
|
unpack-channel.nix: fix 'mv' corner case
|
|
unpack-channel.nix fails if the tarball contains a directory named the
same as the channel:
mv: cannot move 'nixpkgs' to a subdirectory of itself, '.../nixpkgs'
This commit fixes that by not moving the directory if it already has the
correct name.
|
|
The nix-shell fix in 668fef2e4f1c7758e2a55e355b4826014d5b7ba9 revealed
that we had some --pure tests that incorrectly depended on PATH from
config.nix's mkDerivation being overwritten by the caller's PATH.
http://hydra.nixos.org/build/49242478
|
|
|
|
|
|
|
|
Need to remember that std::map::insert() and emplace() don't overwrite
existing entries...
This fixes a regression relative to 1.11 that in particular triggers
in nested nix-shells.
Before:
$ nativeBuildInputs=/foo nix-shell -p hello --run 'hello'
build input /foo does not exist
After:
$ nativeBuildInputs=/foo nix-shell -p hello --run 'hello'
Hello, world!
|
|
... and use this in Downloader::downloadCached(). This fixes
$ nix-build https://nixos.org/channels/nixos-16.09-small/nixexprs.tar.xz -A hello
error: cannot import path ‘/nix/store/csfbp1s60dkgmk9f8g0zk0mwb7hzgabd-nixexprs.tar.xz’ because it lacks a valid signature
|
|
|
|
http://hydra.nixos.org/build/49130529
|
|
This fixes
unable to download ‘https://cache.nixos.org/nar/077h8ji74y9b0qx7rjk71xd80vjqp6q5gy137r553jlvdlxdcdlk.nar.xz’: HTTP error 200 (curl error: Failure when receiving data from the peer)
|
|
|
|
|
|
This reverts commit 81c53fe8e56f4a4ce10088fe2d7b6a524a6dc126. This
check appears to be stricter than we need (it broke a bunch of
platforms that previously did build:
http://hydra.nixos.org/eval/1331921#tabs-now-fail).
|
|
|
|
|
|
Disabled hardened build because it makes the linker fail with messages like
relocation R_X86_64_PC32 against undefined symbol `BZ2_bzWriteOpen' can not be used when making a shared object; recompile with -fPIC
See https://fedoraproject.org/wiki/Changes/Harden_All_Packages.
|
|
http://hydra.nixos.org/build/49031196/nixlog/2/raw
|
|
|
|
|
|
Document toString better
|
|
|
|
This allows <nix/fetchurl.nix> to fetch private Git/Mercurial
repositories, e.g.
import <nix/fetchurl.nix> {
url = https://edolstra@bitbucket.org/edolstra/my-private-repo/get/80a14018daed.tar.bz2;
sha256 = "1mgqzn7biqkq3hf2697b0jc4wabkqhmzq2srdymjfa6sb9zb6qs7";
}
where /etc/nix/netrc contains:
machine bitbucket.org
login edolstra
password blabla...
This works even when sandboxing is enabled.
To do: add unpacking support (i.e. fetchzip functionality).
|
|
Also get rid of Settings::processEnvironment(), it appears to be
useless.
|
|
|
|
|
|
Some sites (e.g. BitBucket) give a helpful 401 error when trying to
download a private archive if the User-Agent contains "curl", but give
a redirect to a login page otherwise (so for instance
"nix-prefetch-url" will succeed but produce useless output).
|
|
Add netrc-file support
|
|
This adds support for s3:// URIs in all places where Nix allows URIs,
e.g. in builtins.fetchurl, builtins.fetchTarball, <nix/fetchurl.nix>
and NIX_PATH. It allows fetching resources from private S3 buckets,
using credentials obtained from the standard places (i.e. AWS_*
environment variables, ~/.aws/credentials and the EC2 metadata
server). This may not be super-useful in general, but since we already
depend on aws-sdk-cpp, it's a cheap feature to add.
|
|
Unbreak 32-bit builds by always implicitly including config.h
|
|
nix-daemon: Don't splice with len=SIZE_MAX
|
|
Currently, 'nix-daemon --stdio' is always failing for me, due to the
splice call always failing with (on a 32-bit host):
splice(0, NULL, 3, NULL, 4294967295, SPLICE_F_MOVE) = -1 EINVAL (Invalid argument)
With a bit of ftracing (and luck) the problem seems to be that splice()
always fails with EINVAL if the len cast as ssize_t is negative:
http://lxr.free-electrons.com/source/fs/read_write.c?v=4.4#L384
So use SSIZE_MAX instead of SIZE_MAX.
|
|
|
|
Because config.h can #define things like _FILE_OFFSET_BITS=64 and not
every compilation unit includes config.h, we currently compile half of
Nix with _FILE_OFFSET_BITS=64 and other half with _FILE_OFFSET_BITS
unset. This causes major havoc with the Settings class on e.g. 32-bit ARM,
where different compilation units disagree with the struct layout.
E.g.:
diff --git a/src/libstore/globals.cc b/src/libstore/globals.cc
@@ -166,6 +166,8 @@ void Settings::update()
_get(useSubstitutes, "build-use-substitutes");
+ fprintf(stderr, "at Settings::update(): &useSubstitutes = %p\n", &nix::settings.useSubstitutes);
_get(buildUsersGroup, "build-users-group");
diff --git a/src/libstore/remote-store.cc b/src/libstore/remote-store.cc
+++ b/src/libstore/remote-store.cc
@@ -138,6 +138,8 @@ void RemoteStore::initConnection(Connection & conn)
void RemoteStore::setOptions(Connection & conn)
{
+ fprintf(stderr, "at RemoteStore::setOptions(): &useSubstitutes = %p\n", &nix::settings.useSubstitutes);
conn.to << wopSetOptions
Gave me:
at Settings::update(): &useSubstitutes = 0xb6e5c5cb
at RemoteStore::setOptions(): &useSubstitutes = 0xb6e5c5c7
That was not a fun one to debug!
|
|
At least in the main Makefile we have:
GLOBAL_CXXFLAGS += -std=c++14 -g -Wall
|
|
|
|
|
|
This is useful for nix-copy-closure.
|
|
For example, SSH stores could be trusted.
|
|
|