Cryptic chromium-style error

[Paul Miller]

This code:

struct Foo {

  std::string s1; 

  std::string s2; 

  std::string s3; 

  std::string s4; 

};

generates a compile error:

../../components/variations/seed_response.h:22:1: error: [chromium-style] Complex class/struct needs an explicit out-of-line constructor.

struct Foo {

^

However, each of these:

struct Foo {

  std::string s1; 

  std::string s2; 

  std::string s3; 

};

typedef struct {

  std::string s1; 

  std::string s2; 

  std::string s3; 

  std::string s4;

} Foo;

struct Foo {

  int i1;

  int i2;

  int i3;

  int i4;

};

compiles successfully.

Questions: what is a "complex class/struct", and why must we write constructors for them? I don't see anything about this in the style guide. And why does only the first example trigger the error?

[Bo Liu]

(from right address..)

std::string is not a POD (plain old data) type, which means the constructor and destructor for your struct actually contains code. My understanding is if you don't write constructors/destructors, then the default generated ones are considered inline. If the struct is used widely, then that causes code bloat, due to many copies of inlined constructor/destructor.

The clang plugin is probably not meant to be bulletproof, just catching the most common use case.

--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev
---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CADsaz-kgBW1_656f%2BvJHfJu5W1SZRqVN6sfoFXiQK-bB-fz2pw%40mail.gmail.com.

[Peter Boström]

Allegedly this didn't trigger for 3 non-POD std::strings. Is that threshold arbitrary or somehow based on the complexity of the inlined constructors?

The "defaulted default" constructor (C++ is good at naming things) is useful here as it arguably doesn't really make you define a constructor, just move it.

In .h:

struct Foo {

  Foo();

  ..
}

In .cc:

Foo::Foo() = default;

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAPCJOUpSS4L6o3ymUgTs%3D-GdiajZf6NnwUood2awbYYMAq8aSw%40mail.gmail.com.

[Daniel Cheng]

The threshhold is somewhat arbitrary. It's actually specifically set to not trigger for up to 3 non-trivial members: https://cs.chromium.org/chromium/src/tools/clang/plugins/FindBadConstructsConsumer.cpp?rcl=4ea283bffe496b27dd41051c0348ff2600e0bab0&l=399

Daniel

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAGFX3sGqNXJ%3DCCX56w0v%2BKiewYBy9EztwKrVinD_ZF4NqwRVRQ%40mail.gmail.com.

[John Rummell]

https://www.chromium.org/developers/coding-style/chromium-style-checker-errors documents this (as I've run into it before).

"Internally, the plugin determines whether a class is complex using a point system, adding 9 points for each templated base class, 10 for each templated member variable, 3 for each non-POD type member variable, and (only for the constructor score) 1 for each integral data type. If a class's score is greater than or equal to 10, it will trip these errors."

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAF3XrKpuE_EGScdP2w2i952FB7auq3pV0kEGsLTFEz1vgKrXWg%40mail.gmail.com.

[Peter Kasting]

Beyond what others noted, there's a little bit related to this at https://www.chromium.org/developers/coding-style/cpp-dos-and-donts#TOC-Stop-inlining-constructors-and-destructors , as well as the sections above and below.

PK

[Paul Miller]

+1 to a better error message ^

Is the typedef thing just a bug in the plugin, then?

On Thu, Jun 7, 2018 at 2:42 PM Alexei Svitkine <asvi...@google.com> wrote:

Could the error from the clang plugin be made to point to the docs?