This is the mail archive of the
mailing list for the binutils project.
Re: Stdcall support in binutils
> Is it enough?
> * Log: Take into account --kill-at (-k) while generating .exp file.
Yes, but could you format it like all the other ChangeLog entries?
That way, I can just cut-n-paste it.
> > Also, the
> > documentation for -k in binutils.texi is wrong. Could you fix that
> > too?
> I'm afraid, not. For documentation my english is not fluent enough, sorry.
OK, I'll wait and see if someone else fixes it.
> I have a general question regarding stdcall names handling:
> Why so much special efforts are required to eliminate the names decoration
> with binutils, while MS tools allow to not think about it at all?
Normally, you shouldn't need to remove the decorations. You should
leave them on so that you don't have runtime errors because you called
the function the wrong way. So, people just haven't debugged that
option very much, because so few people really need to use it.
> Currently I'm using dllwrap to build a dll, and dlltool to build an import
> library. Could it be unified in future?
I added support to ld directly, to handle the common cases. If/when
-k becomed "common" enough, it will get unified. Until then, it waits
for people like you to find and fix the bugs.
> Is there any way to place read only data to .rdata section instead of spreading
> it over the whole .text? (That problem bothered me for years with djgpp).
That's a gcc issue (gcc.gnu.org). The answer is yes, but it was told