Expand on the use of :UNINTERN keyword argument, based on discussion
around issue #85.

Eric Timmons authored Jan 10, 2022 and Robert Goldman committed Jan 10, 2022

Signal a style warning if we `:import-from` using a package that does not exist and an empty symbol list. Remains an error to `:import-from` a package that does not exist with a non-empty symbol list.

Add some tests to try and ensure `:uiop/package` and `:uiop/package-local-nicknames` are not modified.

Resolve "UIOP:DEFINE-PACKAGE signals a TYPE-ERROR when trying to IMPORT-FROM a non-existent package" (#98)

The code handling other options to `define-package` used `find-package*`, 
which (by default) errors if it can't find the package it's looking for.  The
exception was the code for `:import-from`. This caused cryptic error messages
(issue #98) when trying to import from a package that did not exist.

Switched to using `find-package*`, added a new exception class, and added a test.

Only include `:local-nicknames` as an argument to `ensure-package` if it is
non-NIL.

Fixes #84

Eric Timmons authored Jun 03, 2021 and Robert Goldman committed Jun 03, 2021

It got lost in an earlier rebase of this branch.

Prior to this commit, we were trying to be far too smart when updating
a package's package-local-nicknames: we'd only call
`remove-package-local-nickname` on the set difference of the old
nicknames and the new nicknames; that is, we would not remove and then
reinstall a package-local-nickname which was already present and we
wanted to keep. The bug which caused #51 was that the set-difference
considered only the nickname, not the package to which it
referred. But rather than worrying about accurately computing the
minimal set of nicknames to remove, this commit just has
`ensure-package` remove all package-local-nicknames, then reinstall
the ones it wants. This may be either a performance improvement or
regression depending on whether computing set differences is more or
less expensive than manipulating package data structures, but my
suspicion is that no one cares either way.

Also, during the course of my debugging, I [Phoebe Goldman] added a few line-comments
to the definition of `ensure-package` so that I could tell what
various parts of its body were doing. I'd rather not write those
comments again, so I'm including them in this commit.

Tests incorporated and changelog updated.

François-René Rideau authored Apr 04, 2020 and Robert Goldman committed Apr 04, 2020

Fixes #14 as suggested by reporter Andriy Kmit.
Add support for package local nicknames to package-inferred-systems and define-package.

SBCL has been getting aggresive with checking arguments to functions. For some
reason (as of SBCL 1.4.1), setting the package source location in define-package
seems to trigger a source translation for the apply form that ends up producing
(many) compilation notes that the arguments to ensure-package in the keyword
positions are not constant, weakening keyword argument checking. We can get
around that, however, by using a funcall directly.

Tested on SBCL 1.4.4. If sb-c:source-location is evaluated during macro
expansion, then the source location will always point to asdf.lisp (inside the
define-package macro). If it is evaluated after macro expansion, it points to
the right place.

Enhance define-package so SBCL can locate the source-location of the package.

A few are kept, for backward-compatibility with systems in quicklisp.

Based on a contribution by Christian Shafmeister.

(the CLHS implies that a symbol designates a singleton list of its name).

without COPY-LIST on shared data structures. Bug found by drmeister.

Why is this somehow not affecting other implementations???

Define the :DBG macro, so it's ubiquitously available for print-debugging.
Define and export nest, so it's always there to prevent indentation drift.
Tweak restore-image again for better handling of *lisp-interaction*.
Update docstrings for define-package.
Define %process-info-pid.

Everyone should be using a recent 2.7.0 (in ANSI mode) from master for ASDF.

This is quick-build compatible and fixes lp#1230368.

To use package-system, just have foo.asd containing
	(defsystem foo :class package-system)
at the top of your quick-build hierarchy $FOODIR
for packages whose name start with "FOO/"
and ASDF will thereafter look for system "foo/bar/baz" in $FOODIR/bar/baz.lisp.

Such a file will implicitly have its own system defined;
its dependencies are computed by scanning the file,
extracting its first defpackage form,
and using the packages it uses or imports from
as a as a specification of what systems it depends on.

You can register packages as belonging to a system with
	(asdf:register-system-packages my-system '(package1 package2))
Using or importing from a package registered to a given system
will generate a dependency to the registered system.
Using or importing from a packages registered to the constant symbol T
will not generate any dependency.
Using or importing from a packages that is not registered will generate
a dependency on a system the name of which is the package name downcased.
All packages that exist at the time ASDF is initially loaded
are registered to constant symbol T.

Also, for convenience, introduce :use-reexport and :mix-reexport in
uiop/package.lisp (of course, no one can rely on it until it's mainstream,
but better late than never).

To use this style in a way compatible with older versions of ASDF 3,
you may use the asdf-package-system extension.
See lisp-interface-library for a system that uses this style this way.

Push :asdf-package-system to *features*

Another pass package upgrade.
Export FEATURE from correct package.
Use package :asdf-user when reading file for version.

Also, intern fishy symbols that are imported but missing from original package.

Also, rename ensure-pathname-absolute to ensure-absolute-pathname,
to avoid nasty messages while upgrading.
Fix upgrade test.

ENSURE-EXPORTED was exporting the wrong symbol via EXPORT*.
Also, backward compatibility with systems (i.e. cxml) that directly access
the relative-pathname slot of a system the ASDF 1 way,
instead of using system-relative-pathname.
Eliminate the last use of merge-pathnames.

Keep the obsolete slot properties for the sake of upgrade.

Also, move entry-point into system and remove class bundle-system.
I was tempted to have test-op call it, but then it's a problem
when the base system has a non-test entry-point, so we probably need
a separate test-entry-point slot.

Update the default *asdf-debug-utility* to do more of the right thing.

Update changelog.

Tweak package to avoid triggering a bug in ABCL trunk.

A better asdf.asd.

Refactor the preloaded system mechanism, use it for cl-launch 3.21.2.

Rename bundle-pathname to build-pathname, move it from bundle-system to system.

Rename the gcl<2.7 feature gcl2.6. Die on gcl2.5 and earlier.
Don't even try to compile-file asdf using gcl 2.6, just load the .lisp.

Rename asdf/compatibility to asdf/common-lisp, exporting only CL symbols.
Fix define-package to correctly handle exporting of NIL.

Also, make it work with Genera, with #'(lambda instead of (lambda,
and using lisp:loop (via a macro loop*) when destructuring is required.

Document more 2.27 changes in changelog.

If there is no :pathname and no load-pathname,
do NOT have a system pathname take defaults from *default-pathname-defaults*,
but make it NIL.