Keil Logo

How to prevent Error: L6200E: Symbol multiply defined?

Next Thread | Thread List | Previous Thread Start a Thread | Settings

Details Message
Read-Only
Author
Alistair Lowe
Posted
12-Oct-2012 11:49 GMT
Toolset
None
New! How to prevent Error: L6200E: Symbol multiply defined?

Hi guys,

I'm receiving the Error: 'L6200E: Symbol multiply defined'.

I understand one workaround is to include externs, however this is unideal. With most compilers/linkers something such as:

#ifndef bla
#define bla

// Content

#endif

Would prevent double inclusion, however it doesn't appear to work with Keil.

Any ideas on a similar solution?

Many thanks

Read-Only
Author
Per Westermark
Posted
12-Oct-2012 12:24 GMT
Toolset
None
New! RE: How to prevent Error: L6200E: Symbol multiply defined?

"With most compilers"? You seen any C compiler that does not support the use of inclusion guards in header files?

You have given too little information, but you definitely seem to do something wrong.

Read-Only
Author
l kampot
Posted
12-Oct-2012 12:31 GMT
Toolset
None
New! RE: How to prevent Error: L6200E: Symbol multiply defined?

quoter@ Would prevent double inclusion, however it doesn't appear to work with Keil.

that does work on keil.

wothca probably got is function in two .c file with the same name.

Read-Only
Author
Andrew Neil
Posted
12-Oct-2012 12:44 GMT
Toolset
None
New! RE: function in two .c file with the same name

That would also do it.

And note that it's not just functions - anything defined at file scope is global unless you specifically declare it as static

Read-Only
Author
Andrew Neil
Posted
12-Oct-2012 12:42 GMT
Toolset
None
New! This has (almost certainly) nothing to do with "double inclusion"!

"I understand one workaround is to include externs"

Pardon?

extern is not a "workaround"! It is the correct way to give one file access to symbols from another - ie, to symbols which are external to it!

"however this is unideal"

How so?

"With most compilers/linkers something such as ... Would prevent double inclusion"

So-called "include guards" like this have nothing to do with Linkers.

"it doesn't appear to work with Keil"

It does work with Keil! But, of course, if your problem is not multiple-inclusion, then it won't solve it!

Include Guards prevent multiple inclusion of the header in the same translation unit - but do nothing to prevent multiple definitions due to including the same header in multiple translation units!

The solutions is to not have definitions in your headers.


See: http://c-faq.com/decl/decldef.html

Read-Only
Author
Alistair Lowe
Posted
12-Oct-2012 16:03 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

Sorry about this guys,

I'm coming at this as a person who uses C++ primarily for any project large enough to be in multiple files, where naturally I'd know straight away if I'd defined something in my header and got into the bad habit of doing it without realising.

I've just rewritten my code in C++ and now it's looking much cleaner.

Cheers

Read-Only
Author
Hans-Bernhard Broeker
Posted
12-Oct-2012 20:43 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

Sorry to rain on your parade, but this

I've just rewritten my code in C++ and now it's looking much cleaner.

clearly indicates you haven't understood the problem. This has nothing to do with C vs. C++, so moving to C++ can only have brushed the problem under the carpet, or made it a nicer-looking problem --- but it rather certainly didn't solve anything.

Your code had a plain and simple coding mistake which the compiler did tell you about right there in the original error message: multiple definitions. Get out your textbooks and look up the distinction between definitions and declarations, and you'll understand what actually did go wrong.

Read-Only
Author
Bob Dole
Posted
12-Oct-2012 23:19 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

While you may be correct, you are also making assumptions as to what he means by "cleaned up." It could be that when he rewrote his code, it caused him to review what it was written as and he addressed a number of these issues.

Read-Only
Author
Alistair Lowe
Posted
14-Oct-2012 09:17 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

I'm quite aware of what definitions mean and there was never any doubt in my mind as to the difference between this and a declaration, however having never received the error 'L6200E: Symbol multiply defined' before it wasn't immediately obvious (I'm used to symbol errors being related to library linking).

Coding in C++ there's a natural line between definitions only going in class header definitions and declarations in the accompanying cpp, generally with slightly more obvious error messages if you break from this. Likewise, I rarely find myself having to use the keyword 'extern' in an OO environment and am a bit rusty as to its use.

So what I meant by cleaned up with a C++ rewrite was that I simply followed standard C++ coding practise and it naturally solved the issue (after forum users generously highlighted the cause for me) and gave me the OO code I was wanting at the same time.

And thanks Bob for not jumping to conclusions.

Read-Only
Author
Dominic Fandrey
Posted
14-Oct-2012 11:02 GMT
Toolset
None
New! RE: No need for extern in C++

That's just because you rarely use globals in C++, the same problem would have occurred in C++ if you had tried the same thing.

Instead you hand around pointers and references. The advantage is that pointers are just addresses, so declaring a class is sufficient, you don't need to know any internals to be able to create a pointer to it. References are just a different syntax for constant pointers (not pointers to constants), so the same applies to them.

Read-Only
Author
Hans-Bernhard Broeker
Posted
14-Oct-2012 13:11 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

I'm quite aware of what definitions mean and there was never any doubt in my mind as to the difference between this and a declaration,
[...]
Coding in C++ there's a natural line between definitions only going in class header definitions and declarations in the accompanying cpp,

You've just proved yourself wrong in the space of two paragraphs. You got the very difference that you claim to understand fully exactly bass-ackwards in the second part of that quote.

So regardless whether I might have jumped to it or not, you've just proved my conclusion was correct.

Read-Only
Author
Alistair Lowe
Posted
14-Oct-2012 13:42 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

Potato Potato, it's called dyslexia, doesn't mean I don't know the difference and rather, the fact that I've appropriately treated the two separately proves the opposite, that's proof, transiently interchanging two words proves nothing on the other-hand, though thank you for pointing out the error so that it doesn't confuse other readers.

Read-Only
Author
Hans-Bernhard Broeker
Posted
14-Oct-2012 15:11 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

doesn't mean I don't know the difference

I'm calling male bovine excrement on that one.

the fact that I've appropriately treated the two separately proves the opposite,

And the fact your started this thread, asking the question you did proves you wrong. If you actually knew which is which, there's just no way the original error message you got from the linker could have been anything but totally clear to you, so you wouldn't have had to ask here to begin with.

And frankly, if you really have dyslexia at that level you're laying claim to, I have to point out you've picked just about the worst profession you possibly could have.

Read-Only
Author
Alistair Lowe
Posted
14-Oct-2012 15:45 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

Reason for starting this thread: Error message was new (having never mixed the two before in actuality, familiar only with the notion of symbols in library linking).

Reason for resolution: Understood the problem at hand and resolved (only possible with understanding) - that's called proof fyi.

Reason for profession: Spell-check, documentation and creative license kind of renders dyslexia mute, thankfully dyslexia and logical thinking aren't related, though, the logical reasoning that dyslexia would some how prevent me from doing this profession leaves me somewhat bemused.

Reason for your unprofessional/unfriendly attitude on dedicated product support forums: I don't care, however, you're an annoyance and a distraction so go away.

Read-Only
Author
J Rakhomen
Posted
14-Oct-2012 18:17 GMT
Toolset
None
New! One of those annoyances best ignored

Reason for your unprofessional/unfriendly attitude on dedicated product support forums: I don't care, however, you're an annoyance and a distraction so go away.

Unfortunately he sticks around like a particularly bad smell. Must be something relevant to his use of the term male bovine excrement.

Read-Only
Author
Bob Dole
Posted
15-Oct-2012 16:43 GMT
Toolset
None
New! RE: This has (almost certainly) nothing to do with "double inclusion"!

This forum is filled with jerks. Just ignore them. Glad you were able to resolve your issues!

Next Thread | Thread List | Previous Thread Start a Thread | Settings

  Arm logo
Important information

This site uses cookies to store information on your computer. By continuing to use our site, you consent to our cookies.

Change Settings

Privacy Policy Update

Arm’s Privacy Policy has been updated. By continuing to use our site, you consent to Arm’s Privacy Policy. Please review our Privacy Policy to learn more about our collection, use and transfers
of your data.