This commit has no behavior changes.
In particular, this does not fix any of the wrong uses of the previous
default parameter (which used to be 'false', meaning "only replace the
first occurence in the string"). It simply replaces the default uses by
String::replace(..., ReplaceMode::FirstOnly), leaving them incorrect.
The POSIX documentation for `endgrent` only mentions that it "closes
the group database", not that it clears the backing storage for return
values. This means that applications might make use of the returned
values even after closing the group database itself. This includes our
own implementations for `getgrnam` and `getgrgid`.
The specification also states that "the storage areas might be
overwritten by a subsequent call to `getgrgid`, `getgrnam`, or
`getgrent`". This implies that `getgrgid` and `getgrnam` aren't meant
to have their own static storage and instead rely on the storage of
`getgrent`.
The main thing that is missing is validating certain pieces of data
against XML productions in well-formed mode, but nothing uses
well-formed mode right now.
Required by Closure Library for sanitising HTML.
e687b3d8ab/closure/goog/html/sanitizer/safedomtreeprocessor.js (L117)
Previously we would just tightly pack the different libraries' TLS
segments together, but that is incorrect, as they might require some
kind of minimum alignment for their TLS base address.
We now plumb the required TLS segment alignment down to the TLS block
linear allocator and align the base address down to the appropriate
alignment.
glibc before 2.28 defines major() and minor() macros from sys/types.h.
This triggers a Lagom warning for old distros that use versions older
than that, such as Ubuntu 18.04. This fixes a break in the
compiler-explorer Lagom build, which is based off 18.04 docker
containers.
We had a really naive and simplistic implementation, which lead to
various issues where the optimiser incorrectly rewrote the regex to use
atomic groups; this commit fixes that.
This a simple RAII helper for the BorderRadiusCornerClipper it
samples under the corners on construction, then blits them back
on exiting the scope. This encapsulates a fairly common pattern.
Note: With this change the border-radius is clipped if ethier the
overflow-x or overflow-y is hidden (it is a little unclear what
happens if just one is set, but it seems like most browsers
treat one set + border-radius the same as if overflow: hidden
was set).
JsonArray.h does not #include the definition of JsonValue::serialize, as
it lives in JsonObject.h. The macOS Clang target handles symbol
visibility slightly differently (I couldn't figure out how exactly), so
no visible instantiation ended up being created for the function,
causing a link failure.
This aligns it with the spec again, it was clarified that the additional
range check before ArrayCreate is intentional:
https://github.com/tc39/proposal-change-array-by-copy/issues/94
Also cast the final variable to an u64 instead of size_t after we have
determined that it is safe to do so, as that's what Array::create()
takes now.
This doesn't matter per se as the value is immediately validated to be
in the 0 to 2^32 - 1 range, but it avoids having to cast a number that
potentially doesn't fit into a size_t into one at the call site. More
often than not, array-like lengths are only validated to be <= 2^52 - 1,
i.e. MAX_SAFE_INTEGER.
This is fully backwards compatible with existing code as a size_t always
fits into an u64, but an u64 might not always fit into a size_t.
