Age | Commit message (Collapse) | Author | Files | Lines | |
---|---|---|---|---|---|
2012-01-19 | * Add some debug output to print the derivation name once it's known. | Eelco Dolstra | 1 | -1/+4 | |
This makes it easier to pinpoint the source of a crash. | |||||
2012-01-19 | * Add a test case for comparing derivations. | Eelco Dolstra | 2 | -0/+11 | |
2012-01-18 | * Don't distribute Store.cc, it's generated automatically by xsubpp. | Eelco Dolstra | 1 | -1/+3 | |
2012-01-18 | * Pass the appropriate flags to GCC when building the Perl bindings. | Eelco Dolstra | 1 | -1/+2 | |
Without these, Nix fails on 32-bit Linux with Perl 5.14, with a rather unhelpful error message: Not a CODE reference at /nix/store/n6kpbacn6nn7i3i735v8j3di8aqyl07v-perl-5.14.2/lib/perl5/5.14.2/i686-linux-thread-multi/DynaLoader.pm This is likely because the lack of -D_FILE_OFFSET_BITS=64 causes various Perl structures to not match what the Perl interpreter expects. | |||||
2012-01-17 | * Added a command ‘nix-store --print-env $drvpath’ that prints out the | Eelco Dolstra | 2 | -0/+39 | |
environment of the given derivation in a format that can be sourced by the shell, e.g. $ eval "$(nix-store --print-env $(nix-instantiate /etc/nixos/nixpkgs -A pkg))" $ NIX_BUILD_TOP=/tmp $ source $stdenv/setup This is especially useful to reproduce the environment used to build a package outside of its builder for development purposes. TODO: add a nix-build option to do the above and fetch the dependencies of the derivation as well. | |||||
2012-01-13 | * nix-build: put the temporary derivation symlink in a temporary | Eelco Dolstra | 3 | -52/+25 | |
directory rather than the current directory. * nix-build: --drv-link now implies --add-drv-link. | |||||
2012-01-12 | * ‘-module’ causes the generated library to have an .so extension on | Eelco Dolstra | 1 | -1/+0 | |
Mac OS X instead of .dylib, so don't do that. | |||||
2012-01-11 | tests: Run `download-using-manifests' via libtool so that dlopening works. | Ludovic Courtès | 3 | -1/+8 | |
2012-01-11 | perl: Build libNixStore.la with `-module'. | Ludovic Courtès | 1 | -1/+2 | |
2012-01-11 | * Figure out the extension of dynamic libraries; don't hard-code .so. | Eelco Dolstra | 2 | -2/+7 | |
2012-01-07 | * Don't create thunks for simple constants (integers, strings, paths) | Eelco Dolstra | 5 | -133/+193 | |
and allocate them only once. * Move Value and related functions into value.hh. | |||||
2012-01-05 | * Doh! | Eelco Dolstra | 1 | -1/+1 | |
2012-01-05 | * Add a -I flag to the Perl bindings to nix-build and some other | Eelco Dolstra | 7 | -11/+15 | |
scripts. * Include the version and architecture in the -I flag so that there is at least a chance that a Nix binary built for one Perl version will run on another version. | |||||
2012-01-04 | * Remove dead code. | Eelco Dolstra | 2 | -4/+0 | |
2012-01-04 | * Don't use dynamic_cast, it's very slow. "nix-instantiate | Eelco Dolstra | 3 | -24/+23 | |
/etc/nixos/nixos -A system" spent about 10% of its time in dynamic_cast. | |||||
2012-01-04 | * Merge the multiple-outputs-sandbox branch (svn merge --reintegrate | Eelco Dolstra | 21 | -125/+343 | |
^/nix/branches/multiple-outputs-sandbox). Multiple output support still isn't complete, but it wasn't complete in the trunk either, so it doesn't hurt. | |||||
2012-01-04 | * Check whether the outputName attribute works. | Eelco Dolstra | 1 | -2/+2 | |
2012-01-04 | * Let --disable-gc work. | Eelco Dolstra | 1 | -2/+2 | |
2012-01-04 | * currentOutput -> outputName. "current" implies some temporal | Eelco Dolstra | 1 | -1/+1 | |
aspect. | |||||
2012-01-04 | * Export the original input attributes of the derivation in | Eelco Dolstra | 1 | -1/+4 | |
‘drvAttrs’. This will simplify the implementation of functions such as ‘overrideDerivation’ in Nixpkgs, which need to filter out any added attributes such as outPath. | |||||
2012-01-04 | * Simplify the implementation of "derivation" a bit: lift out the | Eelco Dolstra | 1 | -18/+11 | |
common attribution so that they're evaluated only once, etc. Note that the default output is now the first element of the "outputs" attribute, rather than the first element of the sorted list of outputs. This seems more user-friendly. | |||||
2012-01-03 | * Drop the inefficient "Path" suffix in output attribute names. | Eelco Dolstra | 2 | -7/+5 | |
2012-01-03 | * Move the implementation of the ‘derivation’ primop into a separate | Eelco Dolstra | 4 | -26/+35 | |
file. | |||||
2012-01-03 | * Sync with the trunk. | Eelco Dolstra | 35 | -243/+242 | |
2012-01-03 | * Ignore missing manifest symlinks. | Eelco Dolstra | 1 | -0/+1 | |
2012-01-03 | * FreeBSD tar defaults to the tape device instead of stdout. | Eelco Dolstra | 1 | -1/+1 | |
2012-01-03 | * Forgot to add. | Eelco Dolstra | 1 | -0/+13 | |
2012-01-03 | * Add a test for nix-channel. | Eelco Dolstra | 13 | -75/+103 | |
* Refactor the nix-channel unpacker a bit. | |||||
2012-01-03 | * Use Nix::Config. | Eelco Dolstra | 4 | -22/+15 | |
2012-01-03 | * Refactoring: Get rid of a few subdirectories in corepkgs/, and some | Eelco Dolstra | 15 | -84/+52 | |
other simplifications. * Use <nix/...> to locate the corepkgs. This allows them to be overriden through $NIX_PATH. * Use bash's pipefail option in the NAR builder so that we don't need to create a temporary file. | |||||
2011-12-30 | * Follow our own coding conventions. | Eelco Dolstra | 3 | -65/+61 | |
2011-12-30 | * Remove a redundant dot (Nicolas Pierron). | Eelco Dolstra | 1 | -1/+1 | |
2011-12-30 | * Support multiple outputs in nix-store (specifically the ‘--query’ | Eelco Dolstra | 1 | -43/+63 | |
and ‘--realise’ actions). | |||||
2011-12-30 | * Oops. | Eelco Dolstra | 1 | -1/+1 | |
2011-12-30 | * Move topoSortPaths() out of gc.cc. | Eelco Dolstra | 2 | -36/+36 | |
2011-12-30 | * Reject a build if there is a cycle among the outputs. This is | Eelco Dolstra | 4 | -9/+24 | |
necessary because existing code assumes that the references graph is acyclic. | |||||
2011-12-30 | * Fix doc error. | Eelco Dolstra | 1 | -1/+1 | |
2011-12-30 | * Sync with the trunk. | Eelco Dolstra | 9 | -30/+70 | |
2011-12-25 | * Make sure that lock files are cleaned up properly when building | Eelco Dolstra | 1 | -7/+7 | |
through the build hook. | |||||
2011-12-22 | * Oops. | Eelco Dolstra | 1 | -1/+1 | |
2011-12-22 | * In the garbage collector, delete invalid paths before deleting | Eelco Dolstra | 3 | -12/+45 | |
unreachable paths. This matters when using --max-freed etc.: unreachable paths could become reachable again, so it's nicer to keep them if there is "real" garbage to be deleted. Also, don't use readDirectory() but read the Nix store and delete invalid paths in parallel. This reduces GC latency on very large Nix stores. | |||||
2011-12-22 | * Drop unnecessary call to canonPath() (nixStore is already canonical). | Eelco Dolstra | 1 | -1/+1 | |
2011-12-22 | * ‘--disable-shared’ is no longer supported. Fortunately it's not | Eelco Dolstra | 1 | -2/+2 | |
needed for the coverage analysis. | |||||
2011-12-21 | * Another case of lock file permissions being too liberal. | Eelco Dolstra | 1 | -1/+1 | |
2011-12-21 | * Security: make sure the lock files used by build-remote.pl are not | Eelco Dolstra | 1 | -5/+5 | |
readable to other users. Otherwise, any user can open the lock file for reading and lock it, thus DoSing the remote build mechanism. | |||||
2011-12-21 | * Use a lock to ensure that only one build-remote instance can copy a | Eelco Dolstra | 1 | -1/+10 | |
closure to a given machine at the same time. This prevents the case where multiple instances try to copy the same missing store path to the target machine, which is very wasteful. | |||||
2011-12-21 | * The ‘foo.drvPath’ feature was already broken in read-only mode. | Eelco Dolstra | 2 | -4/+5 | |
Since it's rarely used and fixing it is too much work right now, just document it. | |||||
2011-12-21 | Make the reference cycle in the cyclic outputs test indirect | Shea Levy | 1 | -3/+4 | |
2011-12-21 | Revert previous commit | Shea Levy | 1 | -9/+0 | |
It doesn't detect indirect references | |||||
2011-12-21 | Detect and reject mutually-referential outputs | Shea Levy | 1 | -0/+9 | |
There is probably a more efficient way to do this. |