about summary refs log tree commit diff
AgeCommit message (Collapse)AuthorFilesLines
2004-06-22 * Refactoring.Eelco Dolstra4-20/+25
2004-06-21 * Missing files added to `make dist'.Eelco Dolstra1-1/+2
2004-06-21 * Remove debug output.Eelco Dolstra1-1/+0
2004-06-21 * Adapted nix-pull to use the new substitute mechanism.Eelco Dolstra8-58/+99
2004-06-21 * Acquire a lock on the output path when running a substitute. AlsoEelco Dolstra1-0/+22
delete obstructing invalid paths.
2004-06-21 * Ugh, nasty Heisenbug due to an uninitialiased variable. The bugEelco Dolstra1-0/+1
only caused a crash if the program was *not* invoked with a high verbosity level.
2004-06-21 * Remove obstructing invalid store paths add[Text]ToStore().Eelco Dolstra1-0/+6
2004-06-21 * Wrap calls to registerSubstitute() in a single transaction toEelco Dolstra3-11/+13
improve throughput. * Don't build the `substitute-rev' table for now, since it caused Theta(N^2) time and log file consumption when adding N substitutes. Maybe we can do without it.
2004-06-21 * Disable calls to fsync() since Berkeley DB's DB_TXN_WRITE_NOSYNCEelco Dolstra1-0/+8
flag doesn't seem to work as advertised.
2004-06-20 * Re-enable support for substitutes in the normaliser.Eelco Dolstra10-108/+444
* A better substitute mechanism. Instead of generating a store expression for each store path for which we have a substitute, we can have a single store expression that builds a generic program that is invoked to build the desired store path, which is passed as an argument. This means that operations like `nix-pull' only produce O(1) files instead of O(N) files in the store when registering N substitutes. (It consumes O(N) database storage, of course, but that's not a performance problem). * Added a test for the substitute mechanism. * `nix-store --substitute' reads the substitutes from standard input, instead of from the command line. This prevents us from running into the kernel's limit on command line length.
2004-06-20 * README for the Emacs mode (written a while ago, but forgot to commit).Eelco Dolstra1-0/+10
2004-06-20 * Refactoring.Eelco Dolstra3-41/+52
2004-06-19 * Re-enable build hooks.Eelco Dolstra2-75/+327
2004-06-18 * Big refactoring. Move to a much more explicitly state machine basedEelco Dolstra5-766/+874
approach. This makes it much easier to add extra complexity in the normaliser / realiser (e.g., build hooks, substitutes).
2004-06-18 * This is also useful.Eelco Dolstra3-1/+263
2004-06-18 * Shared (garbage collecting) pointers. Copied from Boost.Eelco Dolstra12-17/+1268
2004-06-15 * Refactoring.Eelco Dolstra3-83/+57
2004-06-08 * Cleanup.Eelco Dolstra1-10/+17
2004-06-04 * A Nix mode for Emacs.Eelco Dolstra6-2/+118
2004-05-18 * Drain the output of the build hook to show error messages. UglyEelco Dolstra1-4/+26
hack.
2004-05-18 * Don't set the rpath here --- it's not portable.Eelco Dolstra1-2/+2
2004-05-18 * Go back to the old way of generating the system name, and allow itEelco Dolstra1-3/+6
to be specified in configure (using `--with-system=SYSTEM').
2004-05-18 * setpgrp() is not POSIX (and on Mac OS X it's different than onEelco Dolstra1-1/+1
Linux), so use setpgid().
2004-05-14 * execl() requires a terminating 0.Eelco Dolstra1-10/+21
* When a fast build wakes up a goal, try to start that goal in the same iteration of the startBuild() loop of run(). Otherwise no job might be started until the next job terminates.
2004-05-13 * Distributed builds and load balancing now seem to work pretty well.Eelco Dolstra2-39/+101
(Though the `build-remote.pl' script has a gigantic race condition).
2004-05-13 * Set the executable bit.Eelco Dolstra1-0/+0
2004-05-13 * Load balancing. `build-remote.pl' will only execute up to aEelco Dolstra4-11/+34
per-machine maximum number of parallel jobs on a remote machine.
2004-05-13 * Perform all tests.Eelco Dolstra1-4/+2
2004-05-13 * The build hooks used to implement distributed builds can now be runEelco Dolstra9-144/+498
in parallel. Hooks are more efficient: locks on output paths are only acquired when the hook says that it is willing to accept a build job. Hooks now work in two phases. First, they should first tell Nix whether they are willing to accept a job. Nix guarantuees that no two hooks will ever be in the first phase at the same time (this simplifies the implementation of hooks, since they don't have to perform locking (?)). Second, if they accept a job, they are then responsible for building it (on the remote system), and copying the result back. These can be run in parallel with other hooks and locally executed jobs. The implementation is a bit messy right now, though. * The directory `distributed' shows a (hacky) example of a hook that distributes build jobs over a set of machines listed in a configuration file.
2004-05-12 * Use `-j0'.Eelco Dolstra1-8/+1
2004-05-12 * A switch `-j NUMBER' to set the maximum number of parallel jobs (0 =Eelco Dolstra5-9/+30
no limit). * Add missing file to distribution.
2004-05-12 * Broken test.Eelco Dolstra1-3/+3
2004-05-12 * Pass to the build hook all sorts of information useful forEelco Dolstra1-6/+41
distributing a build action to another machine. In particular, the paths in the input closures, the output paths, and successor mapping for sub-derivations.
2004-05-12 * An quick and dirty hack to support distributed builds.Eelco Dolstra9-13/+92
2004-05-11 * True parallel builds. Nix can now run as many build jobs inEelco Dolstra12-302/+689
parallel as possible (similar to GNU Make's `-j' switch). This is useful on SMP systems, but it is especially useful for doing builds on multiple machines. The idea is that a large derivation is initiated on one master machine, which then distributes sub-derivations to any number of slave machines. This should not happen synchronously or in lock-step, so the master must be capable of dealing with multiple parallel build jobs. We now have the infrastructure to support this. TODO: substitutes are currently broken.
2004-05-11 * Ignore interrupt signals while handling an exception.Eelco Dolstra1-4/+15
* Ignore EINTR in reads and writes.
2004-05-04 * A test to verify that Nix executes build jobs in parallel, ifEelco Dolstra5-3/+81
possible. This test fails right now because this hasn't been implemented right now. Yes, I'm doing Test-Driven Development! ;-)
2004-05-04 * 1000th revision!Eelco Dolstra4-2/+50
* A test to verify that locking of output paths (caused by concurrent invocations of Nix) works correctly.
2004-05-04 * Another test.Eelco Dolstra6-2/+66
2004-05-04 * Grrr. TESTS are not included in EXTRA_DIST.Eelco Dolstra1-1/+1
2004-05-04 * Allow the location of the store etc. to be specified usingEelco Dolstra9-13/+86
environment variables. * Started adding some automatic tests. * Do a `make check' when building RPMs.
2004-04-26 * Bump the version number.Eelco Dolstra1-1/+1
2004-04-26 * Fixed URL.Eelco Dolstra1-2/+2
2004-04-26 * Only add `-preRELEASE' to the version if STABLE != 1.Eelco Dolstra2-4/+40
* Documented release procedures.
2004-04-23 * Pass SYSTEM through config.h, and allow spaces.Eelco Dolstra3-2/+6
2004-04-22 * Look for GC roots in @localstatedir@/nix/gcroots.Eelco Dolstra2-20/+58
2004-04-22 * Don't create $(localstatedir)/nix/profiles if --disable-init-stateEelco Dolstra2-3/+1
is specified.
2004-04-21 * Dist error.Eelco Dolstra1-1/+1
2004-04-21 * Channels. These allow you to stay current with an evolving set ofEelco Dolstra10-52/+194
Nix expressions. To subscribe to a channel (needs to be done only once): nix-channel --add \ http://catamaran.labs.cs.uu.nl/dist/nix/channels/nixpkgs-unstable This just adds the given URL to ~/.nix-channels (which can also be edited manually). To update from all channels: nix-channel --update This fetches the latest expressions and pulls cache manifests. The default Nix expression (~/.nix-defexpr) is made to point to the conjunction of the expressions downloaded from all channels. So to update all installed derivations in the current user environment: nix-channel --update nix-env --upgrade '*' If you are really courageous, you can put this in a cronjob or something. You can subscribe to multiple channels. It is not entirely clear what happens when there are name clashes between derivations from different channels. From nix-env/main.cc it appears that the one with the lowest (highest?) hash will be used, which is pretty meaningless.
2004-04-21 * Do initialise state (the DB etc.) when doing a `make install',Eelco Dolstra4-8/+18
unless `--disable-init-state' is passed to configure.