Should common LESS elements be combined in the CSS output?

[jt000]

Take the following example (repro'd with 2.5.3):

.nav {
  min-height: 20px;

  > a {
    color: blue;
    text-decoration: none;
  }
}

.nav {
  display: block;

  > a {
    padding-top: 3px;
  }
}

.nav > a {
  padding-bottom: 2px;
  color: green;
}

.nav-sm {
  .nav;

  > a {
    color: red;
  }
}

The above LESS creates the following CSS

.nav {
  min-height: 20px;
}
.nav > a {
  color: blue;
  text-decoration: none;
}
.nav {
  display: block;
}
.nav > a {
  padding-top: 3px;
}
.nav > a {
  padding-bottom: 2px;
  color: green;
}
.nav-sm {
  min-height: 20px;
  display: block;
}
.nav-sm > a {
  color: blue;
  text-decoration: none;
}
.nav-sm > a {
  padding-top: 3px;
}
.nav-sm > a {
  color: red;
}

From this example, the .nav, the .nav > a, and .nav-sm > a are created multiple times which could lead to inheritance\ordering differences on different browsers. Should they be combined into a single .nav, .nav > a, and .nav-sm > a elements to be more explicit in the value & decrease the overall size of the CSS?

[seven-phases-max]

See #930. In short: no, Less does not analyze outcoming CSS (except some trivial stuff) and thus cannot perform any optimization of that kind for free. Considering you can get these optimizations with clean-css, this feature is most likely to stay beyond Less scope.

are created multiple times which could lead to inheritance\ordering differences on different browsers.

No, no matter how many times .nav-sm > a appears in the output, inheritance/ordering remains the same in all conformant browsers.

[seven-phases-max]

Merging to #930.