* doc/stdenv: rewrite manual build procedure to be closer to an auto-build
This is based on
<https://jade.fyi/blog/building-nix-derivations-manually/> plus some
more original research.
The previous version of this section did not work for your choice of
simple Haskell package, e.g. haskellPackages.hscolour, due to things
like `compileBuildDriverPhase` and other custom phases that it
does not address at all.
It seems more correct to use genericBuild in development to harmonize it
with what is actually done.
I feel a little bit like I am committing a sin by suggesting using the
experimental CLI in the manual (afaict I am the first to do this), but I
have given the old version of the command, and there are justifiable
reasons to do it:
* The noted limitations with env-vars are fixed. The one with the
non-empty temp directory was one I ran into myself and oh boy was that
not fun to debug.
* Additionally the outputs are set *before* sourcing `setup.sh`: there
is an issue with nix-shell where the original version of `$out` winds
up in `NIX_LDFLAGS` due to _addRpathPrefix, which means that resulting
executables may not run properly.
It is sad that `nix develop` propagates a wrong value of `SHELL` to
builders, though. It is equally sad that `nix-shell` is essentially
abandoned upstream, with undocumented and not insignificant differences
from `nix develop`.
For the exact script differences:
17e6b85d05/src/nix-build/nix-build.cc (L516-L551)db026103b1/src/nix/get-env.sh
Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
* nixpkgs manual: add an alternative example in stdenv-separateDebugInfo
This change gets rid of the indirect reference to `nix-env -i` usage
and shows how to achieve the same goal with a shell expression.
Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
The text was originally added [0] following an apparently incomplete
research on how everything plays together. In fact, Nix propagates
`outputs` to the corresponding nested derivations, and there is some
messy behavior in Nixpkgs that only seems to propagate
`meta.outputsToInstall` in `buildEnv`[1].
This change moves the hints on how to use NixOS specifics to NixOS
module documentation (which is hopefully easier to find through
search.nixos.org), describes the default behavior in Nixpkgs (updating
a the link to the source), and removes the confusing mention of
`nix-env`.
the last of them should not be there to begin with. we don't want
beginners to use `nix-env`, as this is known to run them into trouble
eventually.
[0]: https://github.com/NixOS/nixpkgs/pull/76794
[1]: 1774d07242/pkgs/build-support/buildenv/default.nix (L66)
* Updates meta.chapter.md with a reference link to the usage of the package description field instead of referring to nix-env
---------
Co-authored-by: Valentin Gagarin <valentin.gagarin@tweag.io>
also updates nixdoc to 2.3.0. the nixdoc update is not a separate commit
because that would leave the manual build broken for one commit,
potentially breaking bisects and rebases.
pandoc recognizes `::: note` admonitions, nixos-render-docs only
recognizes `::: {.note}`. surprisingly pandoc also emits the correct
docbook tags for `[](#xref)`s, so we can use that too.
* doc/stdenv/meta.chapter.md: document meta.badPlatforms
We don't have any documentation for the `meta.badPlatforms` attribute.
This commit adds documentation for it.
There has been a longstanding ambiguity between `broken` and
`badPlatforms`, which seem to serve overlapping purposes.
This commit adds to the documentation two examples of constraints
which cannot be expressed by `platforms` and `badPlatforms`.
This commit also mentions `NIXPKGS_ALLOW_BROKEN=1` for overriding
`broken`.
without stable ids on headings we cannot generate stable links to these
headings. nrd complains about this, but the current docbook workflow
does not.
a few generated ids remain, mostly in examples and footnotes. most of
the examples are generated by nixdoc (which has since gained MD export
functions, and the MD export does generate IDs).
Closes #16182
This improves the error message
Error: _assignFirst found no valid variant!
which occurred when the set of outputs was not sufficient to set
the various outputDev, outputBin, etc variables. Specifically, this
would mean that "out" is not among the outputs, which is valid for
a derivation.
This changes the message to something like
error: _assignFirst: could not find a non-empty variable to assign to outputDev. The following variables were all unset or empty: dev out.
If you did not define an "out" output, make sure to define all the specific required outputs: define an output for one of the unset variables.
While this isn't a full explanation of what stdenv can and can not do,
I think it's vast improvement over the 0 bits of information that it
used to provide. This at least gives a clue as to what's going on, and
even suggests a fix, although probably multiple such fixes are required
in an instance where someone starts with a no-out derivation from scratch
(and decide to persist).
