On Fri, 4 Oct 2002, Roth, Kevin P. wrote:

> If necessary, I'm willing to simply change that file every time I build a
> cygwin version of curl (I have to provide a customized source tarball
> anyways, so it's not much extra work).
>
> If you decide to leave "-no-undefined" out, I'm going to want to change my
> cygwin-specific README file. So I'd be sending a small update for that as
> well.
>
> Isn't there any way you could simply *define* each of the currently
> undefined things? And then reintroducing -no-undefined would be correct?

Now, I feel a bit silly since I can't remember exactly for who and why I did
this change. Is this a sign of me getting old? ;-)

Ok, let's do like this. We put it back and smile. Then, when or if we face
this reversed problem with someone who wants it removed, then we write up a
configure hack to detect that case. I'll add a comment to the Makefile.am
saying precisely this.

> Regarding a 7.10.1, as I said I'd rather wait until a second "version" of
> 7.10 is released, whether that be 7.10, 7.10a, 7.10.1, or whatever, before
> building a cygwin version.

In good old curl numbering, the next patch release will be named 7.10.1.

> It seems there have been a good number of build-time issues that have been
> fixed just recently... If you're hoping to avoid that, I can manually apply
> my needed changes, but I generally prefer to make as few changes like that
> as I can.

I have received and made a bunch of changes, sure, but none of them are so
vital or important that they must be released immediately to the world.
They're just parts of the continous stream of fixing going in.

I prefer to wait until I get a certain amount of fixes in, or very important
or big fixes, before I release the next one. This moment has not come yet.

-- 
 Daniel Stenberg -- curl related mails on curl related mailing lists please
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf