nix-shell shebangs were broken by 9018deab

Revert "Revert "next try for "don't abort when given unmatched '}' with 'start-condition stack underflow'. This fixes  #751"""

This reverts commit b669d3d2.

Revert "next try for "don't abort when given unmatched '}' with 'start-condition stack underflow'. This fixes  #751""

This reverts commit ed23c856. Let's
merge this *after* the 1.11.1 release.

next try for "don't abort when given unmatched '}' with 'start-condition stack underflow'. This fixes  #751"

This reverts commit 8120b6fb and fixes the regression introduced in
8d22b264.

Revert "don't abort when given unmatched '}' with 'start-condition stack underflow'. This fixes  #751"

This reverts commit 8d22b264. It
breaks Nixpkgs:

$ nix-env -qa
error: syntax error, unexpected IND_STR, expecting '}', at /home/eelco/Dev/nixpkgs-stable/pkgs/top-level/python-packages.nix:7605:8

This should get rid of a certificate warning from "nix-env -i" early
in the install script.

Mention the --max-freed option in the nix-collect-garbage man page.
Related to https://github.com/NixOS/nix/issues/609.

Similar to 00903fa7. Regardless of -K,
we now also print which output differs.

Previously files in the Nix store were owned by root or by nixbld,
depending on whether they were created by a substituter or by a
builder. This doesn't matter much, but causes spurious diffoscope
differences. So use root everywhere.

This makes it easier to investigate the non-determinism, e.g.

  $ nix-build pkgs/stdenv/linux -A stage1.pkgs.zlib --check -K
  error: derivation ‘/nix/store/l54i8wlw22656i4pk05c52ngv9rpl39q-zlib-1.2.8.drv’ may not be deterministic: output ‘/nix/store/11a27shh6n2ivi4a7s964i65ql80cf27-zlib-1.2.8’ differs from ‘/nix/store/11a27shh6n2ivi4a7s964i65ql80cf27-zlib-1.2.8-check’

  $ diffoscope /nix/store/11a27shh6n2ivi4a7s964i65ql80cf27-zlib-1.2.8 /nix/store/11a27shh6n2ivi4a7s964i65ql80cf27-zlib-1.2.8-check
  ...
  ├── lib/libz.a
  │   ├── metadata
  │   │ @@ -1,15 +1,15 @@
  │   │ -rw-r--r-- 30001/30000   3096 Jan 12 15:20 2016 adler32.o
  ...
  │   │ +rw-r--r-- 30001/30000   3096 Jan 12 15:28 2016 adler32.o
  ...

This occured when sandbox building is disabled, at least one output
exists, and at least one other output does not.

E.g.

  $ nix-build pkgs/stdenv/linux/ -A stage1.pkgs.perl --check
  nix-store: src/libstore/build.cc:1323: void nix::DerivationGoal::tryToBuild(): Assertion `buildMode != bmCheck || validPaths.size() == drv->outputs.size()' failed.

when perl.out exists but perl.man doesn't. The fix is to only check
the outputs that exist. Note that "nix-build -A stage1.pkgs.all
--check" will still give a (proper) error in this case.

This reverts commit 80ebd60e. The
reason why we cleared CFLAGS/CXXFLAGS was because otherwise we get a
default value of -O2, which interferes with the defaults set in the
Makefile. (E.g. "make OPTIMIZE=0" should not pass -O2.)

This was observed in the deb_debian7x86_64 build:
http://hydra.nixos.org/build/29973215

Calling c_str() on a temporary should be fine because the temporary
shouldn't be destroyed until after the execl() call, but who knows...

Detected by -Werror=format-security.

http://hydra.nixos.org/build/29973207

http://hydra.nixos.org/build/30001635

If repair found a corrupted/missing path that depended on a
multiple-output derivation, and some of the outputs of the latter were
not present, it failed with a message like

  error: path ‘/nix/store/cnfn9d5fjys1y93cz9shld2xwaibd7nn-bash-4.3-p42-doc’ is not valid

The --insecure flag to curl tells curl not to bother checking if the TLS
certificate presented by the server actually matches the hostname
requested, and actually is issued by a trusted CA chain.  This almost
entirely negates any benefit from using TLS in the first place.

This removes the --insecure flag to ensure we actually have a secure
connection to the intended hostname before downloading binaries.

Manually tested locally within a dev-shell; was able to download
binaries from https://cache.nixos.org without issue.

[Note: --insecure was only used for fetching NARs, whose integrity is
verified by Nix anyway using the hash from the .narinfo. But if we can
fetch the .narinfo without --insecure, we can also fetch the .nar, so
there is not much point to using --insecure. --Eelco]

Also show types when nix cannot compare values of different types.
This is also more consistent since types are already shown when comparing values of the same not comparable type.

propagate NIX_BUILD_SHELL also in pure builds document NIX_BUILD_SHELL in the nix-shell command documentation

Looks like 5a05cf40 removed usage of
environment CFLAGS and CXXFLAGS by mistake. That change broke building
of nix on fedora core 23.

For example, "${{ foo = "bar"; __toString = x: x.foo; }}" evaluates
to "bar".

With this, we can delay calling functions like mkDerivation,
buildPythonPackage, etc. until we actually need a derivation, enabling
overrides and other modifications to happen by simple attribute set
update.

POSIX compliant directory access (fixes build on Solaris)