Java has a WeakHashMap, js has a WeakMap. It'd be nice if Kotlin/native also provided an equivalent so that a WeakHashMap can be added to the stdlib.

There are a few flags that currently don't have proper error checking. The ones I found are -gc, -scheduler and -size. Using the wrong value will lead to no or unexpected errors.

I think the best way to handle it is to add a new method to compileopts.Config that checks whether there are any faulty flags, which is then called from the main function.

The CUDA datetime tests (numba.cuda.tests.cudapy.test_datetime) only test with datetime[D] - testing with other units would be a good idea.

I also can't find any reference to datetimes being supported on CUDA in the documentation - some explicit mention that they are supported would be helpful.

Per the BufferedInputStream documentation, the skip(n) method corresponds to the general contract in InputStream. Which is:

"If n is negative, the skip method for class InputStream always returns 0, and no bytes are skipped. Subclasses may handle the negative value differently. "

From this I'd expect scala-native's BufferedInputStream.skip to return 0. Instead an `IllegalArgumentEx

According to https://www.weld.rs/grizzly/ the command to get started is $ pip install grizzly however this doesn't seem to be correct. The grizzly package on PyPi is some sort of USB driver.

I believe the documentation should be: pip install pygrizzly but that is failing for me, too

None of the classes hyperlinks lead to the class doxygen class documentation.

PR #342 revamps the checking of bounds declarations. It contains some examples of checking. @lenary noted that none of the examples have unsequenced assignments, where the order in which assignments happen is not specified. This means that pending sets are always empty in the examples. We should add an example where there are some unsequenced assignments.

While the test_suite presented in #589 does work, it is still pretty simple and can be improved and enhanced.

• Add tests for gnutils and coreutils.
• Add necessary utilities so more complex programs can be compiled from sources.
• Integrate CMake, so one could write something like make validate and the subset of test that is deemed necessary (for example everything with min t

from https://github.com/root-project/cling/blob/master/tools/Jupyter/README.md :

export PATH=/cling-install-prefix/bin:$PATH
cd /cling-install-prefix/share/cling/Jupyter/kernel

maybe should be replaced with

export PATH=$PATH:/cling-install-prefix/bin
cd /cling-install-prefix/share/cling/Jupyter/kernel

because cling is required but not clang.
The clang binary mus

Yes, I'm that bored :)

Some of the quoted text in the TeX documentation uses "" rather than ``''. I've done a pass at changing this and will submit a PR.

Having read the Faust documentation, it is unclear what is the best way to get host tempo (BPM, speed etc) into a plugin DSP.

I am referring to DSP code that is compiled with the faust2/plugin/ scripts.

Can this be made clearer in the documentation?

As for Display the special operations generated by the compiler #224 discussion, the documentation will improve if we had a section describing the transformation of each featre and his limitations.

In particular for the special operation, to explain that only the operations that are really generated would appear in the transformation. This is due to the fact that the tool s based on the AST.

Package system managers' build commands, such as makepkg, git clone a new copy of this repo in a clean temporal path, checkout a specific version, proceed with the build, and remove the temporal dir. I'm using the prepare() step, which is executed just before the build, to copy subdir testsuite out of the temporal path; so it can be executed later, after the path is removed.

The procedu

• incremental CTU (on the fly CTU with skip files) feature usage configuration, limitations when ready
• web UI usage
  • how to compare the previous (marked by tag, use the date filter) results of two runs (same run?)
• more simple and advanced use case examples
  • CI job for a quality review (check for high severity results and fail the job if any)
  • usage of the upcoming f

If you grab the header here:

It will scale down to the center of the monitor, and the cursor stays where you grabbed it, but AvalonStudio won't be at the cursor:

Expected

LLVM will soon support adding per-function attributes about available C library functions. We should add support for that, to solve issues like a memcpy implementation being replaced by a memcpy call: https://forum.dlang.org/thread/clptjumxigcozfcyhzzx@forum.dlang.org
Probably ldc.attributes.llvmAttrwill already work for this, but I think it'd be nice to have a more userfriendly attribute.
Som

Would be nice to see padding, in class members.

For field1 padding is:
offset of field2 - offset field1 - sizeof field1
or
size of class - offset of field1- size of field1

Priority: Low

The standard accelerate test suite, used by all the backends, can be quite slow. Several of the tests are significantly slower than the others, for example segmented folds and scans, which I believe is because the reference implementations are very inefficient. Writing some more efficient reference implementations (e.g. using Data.Vector.Unboxed) should help speed things up.

Refactor host -> guest context switch assembly codes that have been duplicated when we introduced AVX support.