Issues in the section 12.1 of the specification

[DartBot]

This issue was originally filed by ms...@unipro.ru

---

1. "Compiling an import directive of the form #import(s) is equivalent to compiling the directive #import(s, prefix: “”);."
   At the same time the spec states that "It is a compile-time error if the value of s2 [prefix] is not a valid identifier."

So the directive #import(s, prefix: “”) is wrong since "" is not a valid identifier.

1. "It is a compile-time error to import two or more libraries that define the same name."
   It is not clear what is the "name" "defined" by a library.
   Conflicting top-level names are prohibited earlier in this chapter, so in this case this rule looks like a duplicate of the previous rule:

"It is a compile-time error if a name N is introduced into the library scope of a library A, and either: ... "

1. The spec does not describe a multiple import of a library with different prefixes. Does this produce any errors/warnings? Can names be accessed by any prefix or only by the last one?
   This question is inspired by the issue Importing core library with a prefix does not remove the names from the global namespace #939.
2. In this section it's not explicitly stated that it is a compile-time error if an imported library does not have a library directive. There is the library production in the section 12:
   libraryDefinition:
     scriptTag? libraryName import* include* resource* topLevelDefinition*

;
So the program composed according to this production could be imported fine since it must contain a libraryName.
But the script definition in section 12.3 contains only optional libraryName and it is not clear is it fine or not to import the script without libraryName.
The only hint is given as a non-normative text: "Named scripts are composable: they can be used as libraries by other scripts and libraries."

[sethladd]

Added Area-Language, Triaged labels.

[gbracha]

Set owner to @gbracha.
Added Accepted label.

[DartBot]

This comment was originally written by rodion...@unipro.ru

---

I assume it means library name.
In that case, there isn't a compile-time error, when you import two different library files, with different prefixes, that declare the same library name in their headers.

[DartBot]

This comment was originally written by rodion...@unipro.ru

---

the above comment was about point 2, which is unchanged in 0.08 draft

[anders-sandholm]

Added this to the M1 milestone.

[gbracha]

The 0.9 draft addresses these points.

1. The empty string is now explicitly allowed.
2. The duplication has been eliminated.
3. Due to the introduction of a separate import scope, the issue is clear.

1. We now require a library.

   ---

   Added Done label.

[DartBot]

This comment was originally written by rodion...@unipro.ru

---

there's still ambiguity in the spec:
Any name N dened by I is in scope in L, unless either:
-� a declaration with the name N exists in L. OR
�- a prex combinator with initial argument N is used in L.

Seems to imply that it's ok to have name conflicts between imported namespace and script's own namespace and in case of identical names one just shadows the other.
Doesn't explicitly say anywhere it's a compile-time error, yet it is in the latest VM.

[gbracha]

There is no ambiguity in the spec. It odes not seem to imply that imports may conflict with the library and the library takes precedence - it just says so. The implementation is behind and teher is a bug filed against it.

[DartBot]

This comment was originally written by rodion...@unipro.ru

---

Got it.

Another question regarding point 2 now:
the wording was changed but I don't think it's become any clearer
"It is a compile-time error to import two or more namespaces that define
the same name." - could you please explain what exactly does this mean if it's not a duplicate of the earlier assertion? :)

[DartBot]

This comment was originally written by rodion...@unipro.ru

---

Also, the requirement that URI must be a compile-time constant and not involve string interpolation have been removed. Please clarify the situation.
Latest VM build seems to require a string constant but, like you said, it is not necessarily a reliable witness.

[DartBot]

This comment was originally written by rodion...@unipro.ru

---

Is the following a typo?
"It is a compile-time error if any of the elements of the first argument of a use of a show combinator is not a valid identifier."

last assertion of the subchapter 12.2

[DartBot]

This comment was originally written by rodion...@unipro.ru

---

Regarding spec v0.10

Is one of these statements in 12.2 redundant or am I reading it wrong?

"It is a compile-time error if a name N is introduced into the import names-
pace of a library A, and another import also introduces N into the import
namespace of A."

"It is a compile-time error to import into a library two or more namespaces
that define the same name."