A package catalog now supplies information about a package's modules
and dependencies, so propagate it when copying a catalog, make the
information accessible via `raco pkg catalog-show', and so on.
The safe-for-space pass could add clearing operations on "typed"
stack positions, which are known to contain a fixnum, flonum, or
extflonum. Non-clearing references, however, were not annotated to
indicate that clearing references were present, since clearing is
not expected on typed positions.
Along the lines of not expecting clearing, the bytecode validator's
encoding of the stack doesn't accomodate both "has a type" and "claims
never to be cleared", so it couldn't detect the bytecode compiler
bug. (Also, this problem didn't show up in the HOSC paper's model of
the validator, because the model pre-dates type tracking.)
Fix the bytecode compiler's space-safety pass so that it never inserts
clearing operations for typed stack positions. Then, the validator can
simply reject any attempt to clear a typed position.
Also, annotate applications generated by lifting as safe-for-space
tail calls.
Merge to v5.3.4
Also, more consistently propagate a given checksum, which can
happen more through the `pkg-install' export from `pkg/lib'
than through `raco pkg'.
Also, report to the user when consulting GitHub or downloading
a checksum.
- build the sub projections only once and build them before getting
the blame objects (instead of after getting the actual values)
- added context information to the blame objects
- added an optional #:min-count argument to insist on a minimum length
for the stream
related to PR 13709
A constant result for foreign-thread use of a callback allows a
callback to return without synchronizing with the Racket thread.
A constant result is thus useful when a callback's work can simply
be skipped if the callback is applied in the "wrong" OS thread.
This termonology change affects lots of function names from `pkg/lib'
and `pkg/db' (former `pkg/pnr-db'), and it also affects some `raco
pkg' commands.
Existing package installations that are marked as 'pnr in a
local configuration are converted automatically to 'catalog, but any
existing "indexes" configuration must be changed to "catalogs".
Bug introduced by 7a8c2ff063: a tree can be deep enough that the
representation of the path to the current item can be too big to
fit into 32 bits. It will always fit in 64 bits, though.
Merge to 5.3.4
... for the purpose of "populate 'compiled' directories" --- but only if
the user has write permission for the package directory.
This change may or may not be a good idea. The idea is that installed
packages generally should be treated in the same way as the main
"collects" tree (e.g., avoiding debugging instrumentation), but if you
happen to be developing a package, then you want it treated like things
that are not in the main "collects" tree. So, how do you pick? Maybe
opening a file in the package is a good way to pick.
If DrRacket decides to skip a file for "populate compiled",
then there may exist a file in "compiled/drracket", anyway,
or there may be such a file for some depenency of the skipped
file. Before this patch, that situation was considered to be a
broken installation, and things would go bad in the likely
case that the "compiled/drracket" files were out of date.
To avoid that problem, parameterize `used-compiled-file-paths'
to drop the DrRacket "populate compiled" target while loading
the skipped file.
This change sets up a more selective "populate compiled" where
a package's modules might switch between eligible and
ineligible for compilation by DrRacket.
Make the installed-package database lock reentrant, change some
functions to take the lock, and fix the documentation on when a
lock is expected to be taken outside of functions.
The code was assuming that the record type was
MX even though this could be false. Also adjusted
some code to also make it easier to test.
Note: the behavior when MX records are missing
is probably still not right.
Closes PR 13654
Adds `pkg/pnr-db', `raco pkg index-copy', and `raco pkg index-show'.
Includes tools to build a database of modules that are supplied by
packages, which will be useful for a tool to recommend package
installs when a module is not found.
Also, document `pkg/lib' and add extra helper functions
for getting package information from a package name resolver.
The validator was not as smart as the compiler in determining
that a `let' expression could be relied on to produce a
constant-shaped function (without side effect or error) in the case
that a right-hand side expression is a call to a known structure
constructor or predicate.
Closes PR 13679
Merge to v5.3.4
Swap order of argument for `environment-variables-get'
and `environment-variables-set!', so that the environment
variables come first --- which follows the usual order.
This change means that the parameter isn't used to get
the default environment variables, but that seems ok; the
convenient interface is `getenv' and `putenv'.
On Windows, case-normalized environment variable names.
Also, change the implementation to use an immutable hash
internally.
By default, a sandbox gets a fresh environment variable set,
which means that it does not affect environment variables
outside the sandbox (which means that sandboxed code cannot
set the Racket process's OS-level environment variables).
Closes PR 13667
The `current-environment-variables' parameter determines the current
mutable "environment variable set". If that set is the initial one for
a Racket process, then using the set corresponds to working with OS
environment variables. Otherwise, it's really just a hash table that
is packaged up as OS environment variables if a subprocess is created.
The new environment-variable interface works in terms of bytes, instead
of assuming that environment variable names and values fit in a string
encoding.
The string-based `getenv' and `putenv' are still available as
convenience wrappers. The checking on environment-variable names
for those wrappers is a little tighter, preventing any attempt to use a
name that contains "=".
The code duplication was there only to support
constructing the name for the optimized contract;
instead we actually just built the name as we go
(the old version actually built the old contract
and then used that to get the name)
also:
- racket/contract/base now requires basic-opters.rkt
so all of the opters are registered when racket/contract/base
is loaded, not just the non-basic ones
- fix the ordering of the names of subcontracts in or/c
- make opt-contracts print a more meaningful name
* Some racketisms.
* Use explicit `in-list' etc in for loops.
* Remove some redundant requires from `net/dns'.
* Move all tests to `tests/net', including a new `tests/net/ip'. In the
future there's a plan to have things like stripped zos etc for
distribution, but we're not there yet, and the net collection is
already organized nicely so this also makes it more uniform.
* Include the dns tests in the main test file.
