Nested File-Import with Sub-Element Selector has unexpected result

[gu-stav]

/* main.less */

html[lang="ru"] {
  @import "languages/ru/clip.less";
}

/* clip.less */

.clip {
  &_sublabel { ... }
  &_module[data-module="touristinfo"] &_details { ... }
}

Result:

/* correct */
html[lang="ru"] .clip_sublabel { ... }

/* incorrect */
html[lang="ru"] .clip_module[data-module="touristinfo"] html[lang="ru"] .clip_details { ... }

/* expected */
html[lang="ru"] .clip_module[data-module="touristinfo"] .clip_details { ... }

OS: Ubuntu 14.10
Less-Version: less@1.7.5

[seven-phases-max]

Could you please provide more info on why you think it "fails"? In fact the result is expected (I guess the point of your confusion is the behaviour of & there so see the docs for more details, specifically notice the .grand { .parent { ... example).

[gu-stav]

The unexpected part is the difference between:

.some {
  &_thing &_within { ... }
}

which, I guess, renders to:

.some_thing .some_within { ... }

But when it's nested, as shown above

/* main.less */
.wrap {
  @import 'include.less'
}

/* include.less */
.some {
  &_thing &_within { ... }
}

compiles to

.wrap .some_thing .wrap .some_within { ... }

but I believe it should be

.wrap .some_thing .some_within { ... }

Otherwise it messes the nested single selector up. Don't you think so, @seven-phases-max? Otherwise it would not be possible to wrap a BEM-Block easy inside of another block (html.js/ html.no-js CSS-Class, e.g.) ...

EDIT: After thinking about this more, I understand your point - less respects the wrapping element around the element as parent. But don't you think this is confusing? Could it be thinkable to have the ability for passing in the context of an import like:

.wrap {
  @import 'some.less'
}

/* some.less */
// context == .wrap

.wrap {
  @import (no-context) 'some.less'
}

/* some.less */
// context == ''

[seven-phases-max]

but I believe it should be

Well, no, regardless of what we think personally the & keyword in Less "represents all parent selectors (not just the nearest ancestor)". This is stated in the docs section I linked above.

Otherwise it would not be possible to wrap a BEM-Block easy inside of another block

The & Less feature is not designed specifically for BEM or any other particular naming scheme.

---

So in summary: this is expected behaviour. (See also related #1075 feature request as well as duplicates of this one: #9, #127, #1158, #1575, #2388 etc.)

[seven-phases-max]

Closing as expected behaviour.