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).
The nixpkgs manual contains references to both sri hash and explicit
sha256 attributes. This is at best confusing to new users. Since the
final destination is exclusive use of sri hashes, see nixos/rfcs#131,
might as well push new users in that direction gently.
Notable exceptions to sri hash support are builtins.fetchTarball,
cataclysm-dda, coq, dockerTools.pullimage, elixir.override, and
fetchCrate. None, other than builtins.fetchTarball, are fundamentally
incompatible, but all currently accept explicit sha256 attributes as
input. Because adding backwards compatibility is out of scope for this
change, they have been left intact, but migration to sri format has been
made for any using old hash formats.
All hashes have been manually tested to be accurate, and updates were
only made for missing upstream artefacts or bugs.
Promote the `maintainers = with maintainers; [ ]` syntax as that is most common
in nixpkgs, and remove the `nix-env` example which doesn't work like that anymore.
A tricky thing about FreeBSD is that there is no stable ABI across
versions. That means that putting in the version as part of the config
string is paramount.
We have a parsed represenation that separates name versus version to
accomplish this. We include FreeBSD versions 12 and 13 to demonstrate
how it works.
This change mimics existing strip{All,Debug}List variables to
allow special stripping directories just for Target.
The primary use case in mind is gcc where package has to install
both host and target ELFs. They have to be stripped by their own
strip tools accordingly.
Co-authored-by: Rick van Schijndel <Mindavi@users.noreply.github.com>
Co-authored-by: Sandro <sandro.jaeckel@gmail.com>
This commit clarifies that the meaning of the `meta.sourceProvenance`
field is independent of and unaffected by the value of the
`meta.license` field. This is based on the intent of the RFC author
as expressed here:
https://github.com/NixOS/nixpkgs/pull/161098#issuecomment-1081270201
This clarification is added for two reasons:
1. If in the future there should be some disagreement about what
`sourceProvenance` to assign to a package, this may help resolve
the disagreement. Any interpretation of `sourceProvenance` which
is influenced by the `meta.license` is clearly an incorrect
interpretation.
2. If it should turn out that it is impossible to disentangle
`sourceProvenance` from `meta.license`, this would indicate the
need for changes to the `sourceProvenance` scheme. That change
might be as simple as replacing the sentence added by this commit
with some other sentence explaining how the two fields influence
each other.
This commit implements the recommendation made in the paragraph of
this comments which begins with "Please say this explicitly...":
https://github.com/NixOS/nixpkgs/pull/161098#issuecomment-1081309089
this is not an actual sync, but rather the manual taking the leading role.
right now it does not make sense to actually change `patch-shebangs.sh`
as that would cause a rebuild of the entire universe.
we should figure out how to keep them aligned with minimal effort both
in terms of maintenance as well as navigation for readers.
This commit describes the "->" notation for dependency types in
greater detail, and uses g++ to provide examples of all six cases
(although the host->target and target->target examples are a bit
artificial).
It also adds three more rows to the table for the "->*" dependency
types for non-compiler-like packages; these dependency types were
already present in the documentation but the "*" was not really
explained.
Lastly, this commit adds a hyperlink to the table from the place where
it is mentioned in the "specifying dependencies" chapter.
This confused the hell out of me, as I didn't spot the
> The following flags are disabled by default ...
when reading about `pie`, because that sentence was hidden in the
previous hardening flag's section.
Also explain that `pie` hardening is on by default on musl.
With removeUnknownConfigureFlags, it's impossible to express a package
that needs --enable-static, but will not accept --disable-shared,
without overriding the result of removeUnknownConfigureFlags _again_
in pkgs/top-level/static.nix.
It would be much better (and more in line with the rest of Nixpkgs) if
we encoded changes needed for static builds in package definitions
themselves, rather than in an ever-expanding list in static.nix. This
is especially true when doing it in static.nix is going to require
multiple overrides to express what could be expressed with stdenv
options.
So as a step in that direction, and to fix the problem described
above, here I replace removeUnknownConfigureFlags with a new stdenv
option, dontAddStaticConfigureFlags. With this mechanism, a package
that needs one but not both of the flags just needs to set
dontAddStaticConfigureFlags and then set up configureFlags manually
based on stdenv.hostPlatform.isStatic.
We are still using Pandoc’s Markdown parser, which differs from CommonMark spec slightly.
Notably:
- Line breaks in lists behave differently.
- Admonitions do not support the simpler syntax https://github.com/jgm/commonmark-hs/issues/75
- The auto_identifiers uses a different algorithm – I made the previous ones explicit.
- Languages (classes) of code blocks cannot contain whitespace so we have to use “pycon” alias instead of Python “console” as GitHub’s linguist
While at it, I also fixed the following issues:
- ShellSesssion was used
- Removed some pointless docbook tags.
Added the following ids to avoid possible id conflicts from ids auto-generated from titles:
- setup-hook-perl
- setup-hook-python
- setup-hook-pkg-config
- setup-hook-automake
- setup-hook-autoconf
- setup-hook-libxml2
- setup-hook-gdk-pixbuf
