Re: libltdl exports no symbols (cygwin)

[Charles Wilson]

Ralf Wildenhues wrote:
> Hello Charles,
>
> Apologies for the huge delay.

> > This patch is against CVS on branch-1-5.  I'll follow up with a similar 
> > patch for HEAD.

Sorry that *I* dropped the ball on a forward port to HEAD.

> I applied this patch now, with a tiny change: instead of
>   #  /* some comment */
>
> use
>      /* some comment */
>
> as I guess that's less likely to disturb a preprocessor.  Below is what
> I applied.

Looks OK to me.

> I am a bit wary that it may cause regressions in users' packages, as
> they may be using LT_SCOPE.  But I see that with the other change, this
> is needed.

User-package use of internal libtool macros (e.g. LT_SCOPE) is outside 
libtool's control.

> I assume the patch below which is a straightforward port to HEAD, is
> needed there as well?  

Yes.

> I've applied it but not tested it yet on w32.
> (The readtext mode was already fixed in HEAD.)

Right.

> I'm still wondering what to put in NEWS for this and the depending
> change.  We should mention this somehow.  Suggestions appreciated.

"Fix regression in libltdl symbol exports on Cygwin. Side effect: 
LT_GLOBAL_DATA and LT_SCOPE are now explicitly defined as 
declspec(dllexport), bypassing auto-export logic on Cygwin.  This tracks 
existing behavior on MinGW."

> A small nit: that's a mailing list address, right?  I don't think that's
> appropriate here, please correct me if wrong.  For now I used the entry
> as below.

I try to avoid hardcoding my actual email address (or alias) on archived 
mailing lists. Plus, I'm more likely to see/read a libtool-related 
message from an unrecognized source if it appears on a mailing list that 
I monitor; direct messages from unrecognized sources are likely to get 
caught in my heavy spam filtering.  But do as you will (just refrain 
from re-posting the "updated" ChangeLog header, next time).

--
Chuck