Duplicated nested mixin rule when @importing multiple times

[suisho]

This issue is like #49.
When import same file multiple times, I got result that has duplicated nested mixin rule (only nested mixin)
For example

_mixin.less

.mixin() {
  color:yellow;
  .mixin-inner{
    color:red;
  }
}

some-block.less

@import "_mixin";

.some_block{
  .mixin;
  color:red;
}

hoge-block.less

@import "_mixin";
@import "some-block.less";

.hoge_block{
  .mixin;
  color:blue;
}

and result
some-block.less

.some_block {
  color: yellow;
  color: red;
}
.some_block .mixin-inner {
  color: red;
}

hoge-block.less

.some_block {
  color: yellow;
  color: red;
}
.some_block .mixin-inner {
  color: red;
}
.some_block .mixin-inner {
  color: red;
}
.hoge_block {
  color: yellow;
  color: blue;
}
.hoge_block .mixin-inner {
  color: red;
}
.hoge_block .mixin-inner {
  color: red;
}

but expected result

.some_block {
  color: yellow;
  color: red;
}
.some_block .mixin-inner {
  color: red;
}
.hoge_block {
  color: yellow;
  color: blue;
}
.hoge_block .mixin-inner {
  color: red;
}

[lukeapage]

so use import-once or don't import twice?

[suisho]

I tried use import-once. But this issue cannot resolved.
And I can't care "don't import twice" when maintenance many less files.

[lukeapage]

then why did import-once not work ?

[Heilemann]

A more fundamental question is, what's the situation where you'd want to import something twice? It has never ever happened to me or anyone on our team. Why would you ever want to have the same thing defined twice? That doesn't make any sense, and causes major headaches when trying to debug. It was causing so many issues for us, because of the sheer number of .less files we're dealing with that we simply forked LESS and fixed it on our version so that @import IS @import-once (which, by the way, didn't actually work on trunk).

[suisho]

The discussion is different from my thinking.

I think this problem is failed filter duplicate rule
that committed by cb78933 and resolve issue #49.
And I expect resolve this problem by fixed this function.

If this result is feature, I come to terms this issue.

[lukeapage]

Can I close this and leave discussion to #212 ?

[Heilemann]

Yes.

[suisho]

@agatronic Issue of #212 is closed but this bug is not fixed in 1.3.1.

[lukeapage]

pull request #1010 has been pulled into 1.4.0

[suisho]

@agatronic hmm... I can't think that pull request not resolve this.
Why that clear this bug?

[lukeapage]

because the file will only be imported once and not twice, so the rules will not be duplicated...

why does it not clear your bug? What behaviour exactly is the problem or that isn't solved?

[suisho]

@agatronic
When I execute above description's example, output has duplicated rule.
I expected following result,

hoge-block.less

.some_block {
  color: red;
}
.some_block color:yellow .mixin-inner {
  color: red;
}
.hoge_block {
  color: blue;
}
.hoge_block color:yellow .mixin-inner {
  color: red;
}

I think that commit of cb78933 is not apply in nested rule (not ralete import once) and this issue isn't fixed.

[dmcass]

First, you're missing a semi-colon in the above example which is causing the weird selector output. After fixing that, your issue becomes more clear. Checked out 1.4.0 and tried it myself.

hoge-block.css

.some_block {
  color: yellow;
  color: red;
}
.some_block .mixin-inner {
  color: red;
}
.some_block .mixin-inner {
  color: red;
}
.hoge_block {
  color: yellow;
  color: blue;
}
.hoge_block .mixin-inner {
  color: red;
}
.hoge_block .mixin-inner {
  color: red;
}

Essentially it looks like nested imports need to be addressed with the import once functionality.

File A imports File B and File C
File B imports File C

File C still gets imported twice.

[cweekly]

FYI/PSA: Twitter-bootstrap 2.x users run into this when they @import bootstrap.less and responsive.less into a single .less file. (They're designed to be compiled separately, so each imports variables.less and mixins.less).
When they're compiled together, responsive.less's (re)import of variables.less and mixins.less becomes redundant, causing undesired side effects like what's described above. So it's an example of duplicate @import in the "real world", albeit relating to naive attempts to customize bootstrap without understanding LESS, vs a legitimate need to re-import. /$.02

[suisho]

Thank you for advice and pont mistake. I update issue

[lukeapage]

ok, I see there is a bug in import once, thanks @dmcass. I have re-labelled this as a bug.

[Heilemann]

Which again begs the question why there are two import methods, and why the
default import behaves in a way that makes no sense what so ever.

On Nov 10, 2012, at 6:42, Luke Page notifications@github.com wrote:

ok, I see there is a bug in import once, thanks
@dmcasshttps://github.com/dmcass.
I have re-labelled this as a bug.

—
Reply to this email directly or view it on
GitHubhttps://github.com//issues/884#issuecomment-10253207.

[lukeapage]

No, in the 1.4.0 branch and future release, the default behaviour has been changed to import-once

[Heilemann]

Fabulous!

Michael Heilemann

On Nov 10, 2012, at 8:10, Luke Page notifications@github.com wrote:

No, in the 1.4.0 branch and future release, the default behaviour has been
changed to import-once

—
Reply to this email directly or view it on
GitHubhttps://github.com//issues/884#issuecomment-10253677.

[kamranayub]

Has the duplicate issue/nested import issue (noted by @dmcass) been fixed as well with @import-once? I am just running into that now and was wondering if it was a LESS issue or the plugin I am using's issue. In order for the plugin (Web Workbench) to provide intellisense/code-completion, you have to import the files you reference. Obviously, this bug becomes rampant then, because the files you import also import the files they need to reference, causing duplicates.

Will the bug only be fixed in 1.4.x or the next 1.3.2 release as well?

[lukeapage]

not yet, its part of the major blockers for 1.3.2 (everything labelled paths)

[lukeapage]

I have committed a fix for import-once so that it uses the full path to determine duplicates, not whatever path is used in @import.. which may be relative.

[kamranayub]

@agatronic, thank you! Gotta say, this is the second issue I've brought up that you guys have fixed ASAP and I'm super thankful. Kudos and keep up the excellent work!