Developer forums (C::B DEVELOPMENT STRICTLY!) > CodeCompletion redesign

code completion fails for struct in a vector

(1/2) > >>

killerbot:
Consider the following code :

--- Code: ---#include <vector>

class Foo
{
public:
    void DoSomething();

private:
    struct Bar
    {
        int mMember1;
        int mMember2;

        Bar() : mMember1(), mMember2() {}
    };

    std::vector<Bar> mBars;
};


void Foo::DoSomething()
{
    Bar bar1;
    mBars.push_back(bar1);

    bar1.mMember1 = 0;

    mBars[0].mMember1 = 0;
}

int main()
{
    Foo foo1;
    foo1.DoSomething();
    return 0;
}

--- End code ---

Focus on the DoSomething() method.

1) if you type "bar1." --> completion works : OK
2) if you type "mBars[0]."  --> NO completion kicks in : NOT OK

hope this example can help for the improvement ;-)

Loaden:
This feature need blueshake help.
 :lol:

Jenna:
Just for the record: it works with clang (new toy  :lol: ):

--- Code: ---jens@debian-inspiron:/tmp/test$ clang++ -w -fsyntax-only -Xclang -code-completion-at=main.cpp:28:14 main.cpp
COMPLETION: Bar : Bar::
COMPLETION: mMember1 : [#int#]mMember1
COMPLETION: mMember2 : [#int#]mMember2
COMPLETION: operator= : [#struct Foo::Bar &#]operator=(<#struct Foo::Bar const &#>)
COMPLETION: ~Bar : [#void#]~Bar()
jens@debian-inspiron:/tmp/test$

--- End code ---

ollydbg:

--- Quote from: killerbot on October 27, 2010, 09:41:37 am ---2) if you type "mBars[0]."  --> NO completion kicks in : NOT OK

--- End quote ---

Current CC will do something like:

splite the statement

--- Code: ---mBars[0].

--- End code ---
to
two pieces

One is: mBars class
The next One is: empty string (everything)

Note:

--- Code: ---[0]
--- End code ---
is skipped. So, cc woks badly.

The way to solve the statement information is quite hard. codelite use the bison/yacc grammar to solve the statement grammar tree.
This is the way like gdb to parse user's statement like you enter some command (GDB use bison/yacc grammar either, you can see gdb's source of c-exp.y under gdb subfolder.

--- Code: ---p mBars[0].Value

--- End code ---

Implement this kind of expression solving is not easy. :D
like jens said, I suggest using Clang (or in the feature) :D it was more powerful and if we use the clang and PCH feature, I think codecompletion will run fast. (see some comments from eranif Clang command line support for codecompletion)


edit
from this page,
http://clang.llvm.org/features.html

clang is a hand-write recursive descent parser.


--- Quote ---A single unified parser for C, Objective C, C++, and Objective C++

Clang is the "C Language Family Front-end", which means we intend to support the most popular members of the C family. We are convinced that the right parsing technology for this class of languages is a hand-built recursive-descent parser. Because it is plain C++ code, recursive descent makes it very easy for new developers to understand the code, it easily supports ad-hoc rules and other strange hacks required by C/C++, and makes it straight-forward to implement excellent diagnostics and error recovery.

We believe that implementing C/C++/ObjC in a single unified parser makes the end result easier to maintain and evolve than maintaining a separate C and C++ parser which must be bugfixed and maintained independently of each other.
--- End quote ---

blueshake:
autually operator [] has been supported for a long time,you guys can write some codes to test it.
the vector doesn't work because of the class[vector]  inheritance is not built. I should do this a long ago.
but recently I am quite losing my time.sorry.

Navigation

[0] Message Index

[#] Next page