Commit graph

306 commits

Author SHA1 Message Date
Rok Garbas dad754843a cygwin: looks like stdout/stdin are reserved words 2015-05-13 09:37:56 +02:00
Eelco Dolstra 60340ce3e2 Implement caching of fetchurl/fetchTarball results
ETags are used to prevent redownloading unchanged files.
2015-04-09 11:42:04 +02:00
Eelco Dolstra a596c525ad Add base64 encoder/decoder 2015-02-10 11:33:33 +01:00
Eelco Dolstra e0def5bc4b Use libsodium instead of OpenSSL for binary cache signing
Sodium's Ed25519 signatures are much shorter than OpenSSL's RSA
signatures. Public keys are also much shorter, so they're now
specified directly in the nix.conf option ‘binary-cache-public-keys’.

The new command ‘nix-store --generate-binary-cache-key’ generates and
prints a public and secret key.
2015-02-04 17:10:31 +01:00
Eelco Dolstra 0d1dafa0c4 Simplify parseHash32 2015-02-03 18:56:47 +01:00
Eelco Dolstra db2ec59903 Simplify printHash32 2015-02-03 18:35:24 +01:00
aszlig 8b88d25cda libutil: Limit readLink() error to only overflows.
Let's not just improve the error message itself, but also the behaviour
to actually work around the ntfs-3g symlink bug. If the readlink() call
returns a smaller size than the stat() call, this really isn't a problem
even if the symlink target really has changed between the calls.

So if stat() reports the size for the absolute path, it's most likely
that the relative path is smaller and thus it should also work for file
system bugs as mentioned in 93002d69fc58c2b71e2dfad202139230c630c53a.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Tested-by: John Ericson <Ericson2314@Yahoo.com>
2015-01-02 12:53:42 +01:00
aszlig bbd45ac80f libutil: Improve errmsg on readLink size mismatch.
A message like "error: reading symbolic link `...' : Success" really is
quite confusing, so let's not indicate "success" but rather point out
the real issue.

We could also limit the check of this to just check for non-negative
values, but this would introduce a race condition between stat() and
readlink() if the link target changes between those two calls, thus
leading to a buffer overflow vulnerability.

Reported by @Ericson2314 on IRC. Happened due to a possible ntfs-3g bug
where a relative symlink returned the absolute path (st_)size in stat()
while readlink() returned the relative size.

Signed-off-by: aszlig <aszlig@redmoonstudios.org>
Tested-by: John Ericson <Ericson2314@Yahoo.com>
2015-01-02 12:53:42 +01:00
Eelco Dolstra 4e0607369e Pedantry 2014-12-14 01:51:14 +01:00
Eelco Dolstra 8bdff8c100 Merge branch 'cygwin-master' of https://github.com/ternaris/nix 2014-12-14 01:49:14 +01:00
Eelco Dolstra f52b6c944e Fix some memory leaks 2014-12-12 15:01:16 +01:00
Eelco Dolstra 28f22b4653 Ensure we're writing to stderr in the builder
http://hydra.nixos.org/build/17862041
2014-12-12 14:35:44 +01:00
Eelco Dolstra ad790022fd Doh 2014-12-12 13:41:00 +01:00
Eelco Dolstra 851b47bd7d Don't do vfork in conjunction with setuid 2014-12-10 18:01:01 +01:00
Eelco Dolstra 0e8fc118b3 Use vfork 2014-12-10 17:25:39 +01:00
Eelco Dolstra ad332e1718 Revert "Use posix_spawn to run the pager"
This reverts commit d34d2b2bbf.
2014-12-10 13:48:50 +01:00
Marko Durkovic 629e8da3aa Explicitly include required C headers 2014-12-09 13:00:59 +01:00
Eelco Dolstra d34d2b2bbf Use posix_spawn to run the pager
In low memory environments, "nix-env -qa" failed because the fork to
run the pager hit the kernel's overcommit limits. Using posix_spawn
gets around this. (Actually, you have to use posix_spawn with the
undocumented POSIX_SPAWN_USEVFORK flag, otherwise it just uses
fork/exec...)
2014-12-05 20:34:41 +01:00
Eelco Dolstra 976df480c9 Add a primop for regular expression pattern matching
The function ‘builtins.match’ takes a POSIX extended regular
expression and an arbitrary string. It returns ‘null’ if the string
does not match the regular expression. Otherwise, it returns a list
containing substring matches corresponding to parenthesis groups in
the regex. The regex must match the entire string (i.e. there is an
implied "^<pat>$" around the regex).  For example:

  match "foo" "foobar" => null
  match "foo" "foo" => []
  match "f(o+)(.*)" "foooobar" => ["oooo" "bar"]
  match "(.*/)?([^/]*)" "/dir/file.nix" => ["/dir/" "file.nix"]
  match "(.*/)?([^/]*)" "file.nix" => [null "file.nix"]

The following example finds all regular files with extension .nix or
.patch underneath the current directory:

  let

    findFiles = pat: dir: concatLists (mapAttrsToList (name: type:
      if type == "directory" then
        findFiles pat (dir + "/" + name)
      else if type == "regular" && match pat name != null then
        [(dir + "/" + name)]
      else []) (readDir dir));

  in findFiles ".*\\.(nix|patch)" (toString ./.)
2014-11-25 11:47:06 +01:00
Eelco Dolstra a3e5c99d66 nix-daemon: Call exit(), not _exit()
This was preventing destructors from running. In particular, it was
preventing the deletion of the temproot file for each worker
process. It may also have been responsible for the excessive WAL
growth on Hydra (due to the SQLite database not being closed
properly).

Apparently broken by accident in
8e9140cfde.
2014-11-19 17:09:27 +01:00
Shea Levy 6062b12160 Fix build on gcc < 4.7 2014-10-20 12:15:50 -04:00
Shea Levy f040159f77 Revert "Drop support for pre-c++11 compilers."
The breakage this fixed can be worked around without removing support.

This reverts commit 84a13dc576.
2014-10-20 11:33:48 -04:00
Shea Levy 84a13dc576 Drop support for pre-c++11 compilers.
In particular, gcc 4.6's std::exception::~exception has an exception
specification in c++0x mode, which requires us to use that deprecated
feature in nix (and led to breakage after some recent changes that were
valid c++11).

nix already uses several c++11 features and gcc 4.7 has been around for
over 2 years.
2014-10-18 22:44:59 -04:00
Eelco Dolstra 3f8576a6ab Remove some duplicate code 2014-10-03 22:37:51 +02:00
Eelco Dolstra 104e55bb7f nix-env: Add regular expression support in selectors
So you can now do things like:

  $ nix-env -qa '.*zip.*'
  $ nix-env -qa '.*(firefox|chromium).*'
2014-10-03 21:29:40 +02:00
Eelco Dolstra 3800f441e4 createDirs(): Handle ‘path’ being a symlink
In particular, this fixes "nix-build -o /tmp/result" on Mac OS X
(where /tmp is a symlink).
2014-10-03 16:53:28 +02:00
Eelco Dolstra 0342eb1705 Remove bogus comment 2014-09-19 15:07:22 +02:00
Eelco Dolstra 809ca33806 Use PR_SET_PDEATHSIG to ensure child cleanup 2014-08-21 15:31:43 +02:00
Eelco Dolstra 11849a320e Use proper quotes everywhere 2014-08-20 18:03:48 +02:00
Eelco Dolstra 373fad75e1 Add some color 2014-08-20 16:50:17 +02:00
Eelco Dolstra 392430b2c4 nix-store -l: Automatically pipe output into $PAGER 2014-08-20 15:12:58 +02:00
Eelco Dolstra 4d73e2e893 Get rid of "killing <pid>" message for unused build hooks 2014-08-04 17:27:45 +02:00
Eelco Dolstra 4c4b219c07 Call commonChildInit() before doing chroot init
This ensures that daemon clients see error messages from the chroot
setup.
2014-08-01 19:29:03 +02:00
Eelco Dolstra eea0401d7a Eliminate redundant copy 2014-08-01 17:30:51 +02:00
Eelco Dolstra daf3f2c11f Make readDirectory() return inode / file type 2014-08-01 17:14:28 +02:00
Eelco Dolstra 50dc1f5b71 Restore default SIGPIPE handler before invoking ‘man’
Fixes NixOS/nixpkgs#3410.
2014-07-31 10:31:17 +02:00
Eelco Dolstra 0fae20c362 Use pthread_cancel instead of a signal
Signal handlers are process-wide, so sending SIGINT to the monitor
thread will cause the normal SIGINT handler to run. This sets the
isInterrupted flag, which is not what we want. So use pthread_cancel
instead.
2014-07-24 11:47:51 +02:00
Eelco Dolstra aa1560ca07 Fix bogus pass by reference
http://hydra.nixos.org/build/12711659
2014-07-24 09:58:07 +02:00
Eelco Dolstra 24c6d992c6 More debugging 2014-07-24 01:21:56 +02:00
Eelco Dolstra 1eb0af7ed5 Add some assertions 2014-07-24 00:16:06 +02:00
Eelco Dolstra d0fad1779f nix-daemon: Simplify stderr handling 2014-07-23 19:37:40 +02:00
Eelco Dolstra 49fe9592a4 nix-daemon: Use a thread instead of SIGPOLL to catch client disconnects
The thread calls poll() to wait until a HUP (or other error event)
happens on the client connection. If so, it sends SIGINT to the main
thread, which is then cleaned up normally. This is much nicer than
messing around with SIGPOLL.
2014-07-23 19:21:00 +02:00
Eelco Dolstra fdee1ced43 startProcess: Make writing error messages from the child more robust 2014-07-23 19:11:26 +02:00
Eelco Dolstra 2e77bd70fa Better fix for strcasecmp on Darwin 2014-07-18 12:54:30 +02:00
Eelco Dolstra 8ddffe7aac Ugly hack to fix building on old Darwin
http://hydra.nixos.org/build/12580878
2014-07-17 23:57:17 +02:00
Eelco Dolstra 2304a7dd21 Get rid of a compiler warning 2014-07-16 16:32:26 +02:00
Eelco Dolstra 985f1595fe Be more strict about file names in NARs 2014-07-16 16:30:50 +02:00
Eelco Dolstra 276a40b31f Handle case collisions on case-insensitive systems
When running NixOps under Mac OS X, we need to be able to import store
paths built on Linux into the local Nix store. However, HFS+ is
usually case-insensitive, so if there are directories with file names
that differ only in case, then importing will fail.

The solution is to add a suffix ("~nix~case~hack~<integer>") to
colliding files. For instance, if we have a directory containing
xt_CONNMARK.h and xt_connmark.h, then the latter will be renamed to
"xt_connmark.h~nix~case~hack~1". If a store path is dumped as a NAR,
the suffixes are removed. Thus, importing and exporting via a
case-insensitive Nix store is round-tripping. So when NixOps calls
nix-copy-closure to copy the path to a Linux machine, you get the
original file names back.

Closes #119.
2014-07-16 16:02:05 +02:00
Eelco Dolstra d0eb970fb4 Fix broken Pid constructor 2014-07-10 21:48:21 +02:00
Eelco Dolstra 8e9140cfde Refactoring: Move all fork handling into a higher-order function
C++11 lambdas ftw.
2014-07-10 16:58:09 +02:00