This is the mail archive of the cygwin@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: binutils auto-import bug


Dmitry Timoshkov wrote:

> "Charles Wilson" <cwilson@ece.gatech.edu> wrote:
> 
> 
>>>The reason for the quotation marks is: Paul's patch doesn't actually fix 
>>>the bug.  It just detects the condition that *causes* the bug, prints a 
>>>diagnostic message, and exits with error.
>>>
>>Oh, one other thing: this is *meant* as an interim step towards actually 
>>*fixing* the bug.  But, because of certain idiosyncracies of the windows 
>>dynamic loader, detection-and-prevention, as in the current patch, MAY 
>>be the best we can do.
>>
>>It isn't yet clear.
>>
> 
> Did you try to use different compiler/linker? For instance from MIcrosoft
> or Borland? How do they behave in your tests?


Ummm...you're missing the point. This is about how the binutils linker 
links, when the "auto-import" feature is used. NOT about how any other 
linker would work in some hypothetical situation.  As far as the DLLs 
themselves are concerned, if they are produced by binutils-ld, it is 
theoretically possible to create a .lib-style import library for the 
DLL, and then use MS-link.exe to link to them, if the client code uses 
__declspec(dllimport).

MS nor Borland linkers HAVE an "auto-import" feature.  So how could I 
test it?

--Chuck

Auto





--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]