This reverts commit 772a0eb68ac95b7e24508580499b49872fdb541f, reversing
changes made to 0bb178bbb2451238a326c6e916ecf38fbc34cab1.
This needs a rebase against master branch - it has build-breaking merge
conflicts. I also added a "changes requested" review on the original
pull request.
* Scan from line start when finding tag in tokenizer
This resolves a crash that can occur for invalid bytes like carriage
returns that are valid characters when not parsed from within literals.
There are potentially other edge cases this could resolve as well, as
the calling code for this function didn't account for any potential
'pending_invalid_tokens' that could be queued up by the tokenizer from
within another state.
* Fix carriage return crash in multiline string
Follow the guidance of #38:
> However CR directly before NL is interpreted as only a newline and not part of the multiline string. zig fmt will delete the CR.
Zig fmt already had code for deleting carriage returns, but would still
crash - now it no longer does so. Carriage returns encountered before
line-feeds are now appropriately removed on program compilation as well.
* Only accept carriage returns before line feeds
Previous commit was much less strict about this, this more closely
matches the desired spec of only allow CR characters in a CRLF pair, but
not otherwise.
* Fix CR being rejected when used as whitespace
Missed this comment from ziglang/zig-spec#83:
> CR used as whitespace, whether directly preceding NL or stray, is still unambiguously whitespace. It is accepted by the grammar and replaced by the canonical whitespace by zig fmt.
* Add tests for carriage return handling
This also makes another breaking change to for loops: in order to
capture a pointer of an element, one must take the address of array
values. This simplifies a lot of things, and makes more sense than how
it was before semantically.
It is still legal to use a for loop on an array value if the
corresponding element capture is byval instead of byref.
RunStep is supposed to auto-detect whether the intend is for
side-effects or for producing an output file. The auto-detection logic
was incorrect, and this commit fixes it.
I tested this manually locally. Automated testing will require a more
significant investment in the test harness, which I will work on in a
future enhancement.
closes#14666
This makes it so that when there is a tree of std.Build objects, only
one zig-cache is used (the top-level application) instead of polluting
package directories with zig-cache folders.
* Use std.Build.Cache.Directory instead of a string for storing the
cache roots and build roots.
* Set up a std.Build.Cache in build_runner.zig and use it in
std.Build.RunStep for avoiding redundant work.
Now that this API is used by the build system, these debug logs are
problematic because build scripts run in debug mode, making these logs
noisy output.
This provides file path as a command line argument to the command being
run, and returns a FileSource which can be used as inputs to other APIs
throughout the build system.
Unfortunately, it is implemented by pooping a ton of temporary files
into zig-cache/tmp for the time being. I think one of the very next
improvements to the build system should be moving the compiler's cache
system to the standard library and using it in the build system. I had a
look at the dependencies and it is already pretty untangled.
Deprecate CompileStep.run. The problem with this function is that it
does the RunStep with the same build.zig context as the CompileStep, but
this is not desirable when running an executable that is provided by a
dependency package. Instead, users should use `b.addRunArtifact`.
This has the additional benefit of conforming to the existing naming
conventions.
Additionally, support enum literals in config header options values.
Ascon has been selected as new standard for lightweight cryptography
in the NIST Lightweight Cryptography competition.
Ascon won over Gimli and Xoodoo.
The permutation is unlikely to change. However, NIST may tweak
the constructions (XOF, hash, authenticated encryption) before
standardizing them. For that reason, implementations of those
are better maintained outside the standard library for now.
In fact, we already had an Ascon implementation in Zig:
`std.crypto.aead.isap` is based on it. While the implementation was
here, there was no public API to access it directly.
So:
- The Ascon permutation is now available as `std.crypto.core.Ascon`,
with everything needed to use it in AEADs and other Ascon-based
constructions
- The ISAP implementation now uses std.crypto.core.Ascon instead of
keeping a private copy
- The default CSPRNG replaces Xoodoo with Ascon. And instead of an
ad-hoc construction, it's using the XOFa mode of the NIST submission.
This reverts commit abc9530a88d24350481d9264edcde300f293929a.
This patch implies that the idiomatic Zig way of handling anytype
parameter is to write a bunch of boilerplate instead of directly
accessing type information and relying on the compiler to be useful.
I don't want it to be this way.
It is the compiler's job to make useful error messages when the wrong
field of a type info result is accessed, and it is the zig programmer's
job to understand what it means when a compile error points at the field
access of `@typeInfo` (along with the relevant callsites).
One thing that might be useful would be having the compiler be aware of
module boundaries and highlighting the boundaries of them. The first
reference note after crossing a module boundary is likely the most
interesting one.
Breaking API change to std.Build.addConfigHeader. It now uses an options
struct.
Introduce std.Build.CompileStep.installConfigHeader which also accepts
an options struct. This is used to add a generated config file into the
set of installed header files for a particular compilation artifact.
std.Build.ConfigHeaderStep now additionally supports a "blank" style
where a header is generated from scratch. It no longer exposes
`output_dir`. Instead it exposes a FileSource via `output_file`.
It now additionally accepts an `include_path` option which affects the
include path of CompileStep when using the `#include` directive, as well
as affecting the default installation subdirectory for header
installation purposes.
The hash used for the directory to store the generated config file now
includes the contents of the generated file. This fixes possible race
conditions when generating multiple header files simultaneously.
The values hash table is now an array hash map, to preserve order for
the "blank" use case.
I also took the opportunity to remove output_dir from TranslateCStep and
WriteFileStep. This is technically a breaking change, but it was always
naughty to access these fields.
