Skip to Content.
Sympa Menu

freetds - RE: Error compiling

freetds AT lists.ibiblio.org

Subject: FreeTDS Development Group

List archive

Chronological Thread  
  • From: "Lowden, James K" <LowdenJK AT bernstein.com>
  • To: "'TDS Development Group'" <freetds AT franklin.metalab.unc.edu>
  • Subject: RE: Error compiling
  • Date: Thu, 11 Jul 2002 15:19:26 -0400


> From: Eric Deutsch [mailto:edeutsch AT systemsbiology.org]
> Sent: July 11, 2002 12:21 PM
>
> I'm not sure I understand all the implications of this
> tds_willconvert.h, but if its current implementation prevents
> a simple:
>
> cvs checkout
> ./configure
> make
> make install
>
> I would like to suggest that it be in CVS, even if autogenerated.

That won't work today. I decided to leave cvs a little broken, rather than
adding a file that might not belong there.

My thought was, there are two kinds of developers:

A. CVS users, who need the GNU autotools to generate the configure
script
B. Tarball users, who just run ./configure

The question is, should Type A users ;) also be required to have Perl,
because that's what generates tds_willconvert.h. Or, is it better just to
put big warnings on tds_willconvert.h that say, "Please Don't Edit Me, edit
convert.c intead" (in the appropriate Billy Goat Gruff voice).

If the answer is, No Perl Needed (for anyone), I'll check in the file and
hope no one touches it with vi. As long as everyone plays nice, that's the
easier route. It's really really easy to edit the "truth table" in
convert.c and run

As a practical matter, I don't see the file changing very much. Right now,
it holds what's *supposed* to be convertible, according to Sybase. Once
Bill is done with TDS convert, Sybase's answer and ours will be the same, so
there won't be an ongoing need to tweak the file with changing yes/no
answers.

Hope that makes the implications clearer. :)

--jkl




Archive powered by MHonArc 2.6.24.

Top of Page