Re: undefined reference to `__chkstk_ms'

Dear Keith,

I really appreciate your answer.

Here are some notes:
- I will try to raise the issue with my team leader

- the web site www.mingw.org nor the installation process does not mention
trademark as far as I can see.

- the link to http://www.mingw.org/wiki/MinGW_licensing listed in
http://www.mingw.org/wiki/FAQ#toc11 should be probably changed to
http://www.mingw.org/license

- can GCC compiler provided by MinGW generate 64-bit binaries?
A brief test tells me:

  C:\temp>gcc --version
  g++ (MinGW.org GCC-8.2.0-3) 8.2.0
  ...
  C:\temp>g++ -m64 -c -o foo.o foo.c
  cc1plus.exe: sorry, unimplemented: 64-bit mode not compiled in

This may be the reason why I got my hands on mingw-w64.
--
	bye
		rastos


On 27/05/2019 17.04, Keith Marshall wrote:
> On 27/05/19 09:05, Rastislav Stanik wrote:
>> I've hit a problem when compiling my project with
>> gcc --version
>> gcc (x86_64-win32-sjlj-rev0, Built by MinGW-W64 project) 8.1.0
> 
> First, let's make it completely clear: this compiler originates from a
> third party project which illegally uses our registered trademark,
> without our authorization; it is NOT a MinGW product, and is not
> supported here.
> 
>> I'm trying to build a 32-bit (using -m32) shared library that uses
>> PCRE (Perl Compatible Regular Expressions library) compiled by myself
>> using the same compiler. At linking stage I get a report about
>> missing symbol __chkstk_ms:
>>
>> {MinGW-root}/bin/../lib/gcc/x86_64-w64-mingw32/8.1.0/../../../../x86_64-w64-mingw32/lib/../lib32/libmingw32.a(lib32_libmingw32_a-pseudo-reloc.o):pseudo-reloc.c:(.text+0x18d):
>> undefined reference to `__chkstk_ms'
>>
>> Reference to __chkstk_ms() is added by compiler when some function
>> declares local variables exceeding page size (~4k).
> 
> Amongst other reasons.
> 
>> In my case the function that exceeds the limit is in PRCE, so I
>> cannot simply work around it in my code. The function __chkstk_ms()
>> is provided by libgcc.a located in directory
>>
>>   {MinGW-root}/lib/gcc/x86_64-w64-mingw32/8.1.0/32/
>>
>> So if I add linking of that libgcc.a then the linking stage succeeds.
>> But I also have to ensure that libgcc.a is in link search path.
>>
>> Questions:
>> - why the compiler does not add libgcc.a by itself when it added
>> __chkstk_ms() by itself?
> 
> Our distribution does exactly that; irrespective of whether or not
> __chkstk_ms() is referenced, -lgcc is specified by default, and the
> appropriate path is also specified.
> 
>> - if I add -lgcc, why do I also have to add path with -L? Shouldn't
>> the directory be in the default linker search path? - is there a
>> solution that will keep working after next release of MinGW/GCC when
>> the path to libgcc changes?
> 
> You aren't using a genuine MinGW product, so who knows?  A possible
> explanation could be that your build system is trying to invoke ld
> directly, rather than letting gcc do the linking ... you don't show the
> actual link command.  If you don't indulge in that misguided shortcut,
> then this problem should not arise, with our GCC-8.2.0
> 
> 
> _______________________________________________
> MinGW-Users mailing list
> [email protected]
> 
> This list observes the Etiquette found at
> http://www.mingw.org/Mailing_Lists.
> We ask that you be polite and do the same.  Disregard for the list etiquette may cause your account to be moderated.
> 
> _______________________________________________
> You may change your MinGW Account Options or unsubscribe at:
> https://lists.osdn.me/mailman/listinfo/mingw-users
> Also: mailto:[email protected]?subject=unsubscribe
> 

_______________________________________________
MinGW-Users mailing list
[email protected]

This list observes the Etiquette found at
http://www.mingw.org/Mailing_Lists.
We ask that you be polite and do the same.  Disregard for the list etiquette may cause your account to be moderated.

_______________________________________________
You may change your MinGW Account Options or unsubscribe at:
https://lists.osdn.me/mailman/listinfo/mingw-users
Also: mailto:[email protected]?subject=unsubscribe