* Add some more tests for the sqrt/sqrtf implementations.
The idea is to cross-check the software impl with the HW one whenever
possible.
* Fix a broken test, oops.
The code initializes twice `t` instead of `t1`, leaving the latter
uninitialized. The problem manifested itself by corrupting the LSBs of
the result in unpredictable ways.
Before this change every keypress in the search field causes a browser
history entry, which makes navigating back annoying.
On first keypress in the search field, a new history entry is created.
On subsequent keypresses, the most recent history entry is replaced.
Therefore a typical history after searching and navigating to an entry
might look like
1. documentation root
2. search page "print"
3. docs for `std.debug.print`
Co-authored-by: Žiga Željko <ziga.zeljko@gmail.com>
This is useful for build.zig files to check in some cases, for example
to adhere to the convention of installing config to /etc instead of
/usr/etc on linux when using the /usr prefix. Perhaps std.build will
handle such common cases eventually, but that is not yet the case.
This broke build scripts that wanted to refer to `exe_dir` or
`install_path`.
There has also been some pushback and discussion on this breaking
change. I think it should be re-evaluated.
This reverts commit a1a1929cf4ff979bdaba17f858d4ce8647e87a65.
We generally get away with atomic primitives, however a lock is required
around the refresh function since it traverses the Node graph, and we
need to be sure no references to Nodes remain after end() is called.
* Field global_cache_root was added to Builder struct along with
mandatory argument for build_runner.zig. Logic for using the custom
global cache was also added.
* The arguments --cache-dir and --global-cache-dir are no longer passed
directly through to build_runner.zig and are instead only passed through the
mandatory cache_root and global_cache_root arguments.
Let's a void any kind of compilation/LLVM errors for niche targets such
as AVR/MSP430 or ARM v6m. By not exporting any atomic builtin anymore
the user is free to provide their own implementation (that disable the
IRQs) or to provide the --single-threaded switch and forget about this.
The previous clone() definition was written for 32bit PPC targets,
trying to execute it on a PPC64 machine lead to countless crashes.
Use the clone() definition kindly lifted from Musl libc.
Comparisons with absolute epsilons are usually useful when comparing
numbers to zero, for non-zero numbers it's advised to switch to relative
epsilons instead to obtain meaningful results (check [1] for more
details).
The new API introduces approxEqAbs and approxEqRel, where the former
aliases and deprecated the old `approxEq`, allowing the user to pick the
right tool for the job.
The documentation is meant to guide the user in the choice of the
correct alternative.
[1] https://randomascii.wordpress.com/2012/02/25/comparing-floating-point-numbers-2012-edition/
