* Fix result not being able to use in quote do
This fixes the annoying issue of not be able to use result inside a
quote do block. It works by a simple trick. The quote do mechanic is
based on dynamically creating a template and immediately calling it with
the arguments found within the quote do block. Since this is called in
the scope of the macro the result variable is shadowed. This trick works
by changing all occurences of result (which shouldn't cause any issues
as result isn't used for anything else for the same reason) to another
name and then passing in an IdentNode with result as a named parameter
with that name.
Note that currently this just replaces it with a fixed named variable
"res" which should be changed to a non-colliding, dynamically created
name.
* Fix hard coded parameter "res" to anonymous symbol
This fixes the hard coded parameter "res" to be an anonymous symbol
instead so it won't collide with other parts of the argument list.
* Add test case for result in quote do block
A simple test case based on GitHub issue #7323 on how you can't put
result in a quote do block. This test verifies that it actually works
correctly now.
* Add test for explicit capturing of result
* Rebased against devel
* Early evaluation of mIs
The `evalIs` implementation was just a broken copy of `isOpImpl` so
let's just avoid it alltogether: `mIs` nodes are either resolved during
the semantic phase or bust.
* Remove dead code and tidy it up
* Don't consider tyAnd/tyNot/tyOr/tyAnything as generic
`containsGenericType` was too shallow and didn't check all the branches.
The resulting half-processed nodes are often simplified by the constant
folding pass but when that's not possible we get a nasty error during
codegen.
Fixes#8693
* Move the blame onto the semFold pass
Slightly better evaluation of `is` forms.
* fix#7405 and #8195
* control pushInfoContext in semExprNoType,genStmts via a new hintExtendedContext; make NotesVerbosity computation more DRY
* addressed comments
* call `semExpr` in `semIs` if node is `strLit`, fixes#8129
In case the second son of the node in `semIs` is of kind `strLit`, we
now call `semExpr` to set the `typ` field of that node.
Also removes the `t2 != tyTypeDesc` check in `isOpImpl`, since the
kind of `n[2]` is already assertet with the `internalAssert`.
* reintroduce check for `t2.kind != tyTypeDesc` to fix test case
The `internalAssert` in the `isOpImpl` doesn't check
`n.sons[2].typ.kind` as I previously read, but rather
`n.sons[2].kind`. Therefore the check for `tyTypeDesc` here is
useful. Otherwise the last test case in `isopr.nim` fails.
Also removes the flag `efDetermineType` from the call to `semExpr`.
* late instantiation for the generic procs' default param values
* automatic mixin behaviour in concepts
Other fixes:
* don't render the automatically inserted default params in calls
* better rendering of tyFromExpr
This also makes the first baby steps towards a sound treatment of
higher-order kinds (type type int).
Adds test cases showcasing the new features.
* Also fixes breakage after the rebase
Fixes issue #7997, which was caused by an export of a `release` proc
in `locks`. Thus the `release` in `defined(release)` of the `ifDebug`
template, was of kind `nkSym` instead of `nkIdent`.
We fix this by getting the `PIdent` of the argument to `defined` using
`considerQuotedIdent`.
This has the nice property of also checking for a valid identifier for
us. E.g. `defined(123)` would fail with
```
Error: in expression 'defined(123)': identifier expected, but found
'123'
```
The `localError` is removed, due to a clear distinction between
`declared` and `defined` now.
