<br><br><div><span class="gmail_quote">2005/9/17, Peter Selinger <<a href="mailto:selinger@mathstat.dal.ca">selinger@mathstat.dal.ca</a>>:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Arnaud Quette wrote:<br>><br>> 2005/9/16, Peter Selinger <<a href="mailto:selinger@mathstat.dal.ca">selinger@mathstat.dal.ca</a>>:<br>> ><br>> > * I committed into CVS the gendb patch that I sent to the mailing list
<br>> > on September 12. Arnaud, I hope you approve of this patch; it is<br>> > more reliable than the previous method in tracking the correct<br>> > dependencies.<br>><br>><br>> no, I reject this one as is.
<br>> it's gcc specific to gcc, so has to be tied some way to autodetecting if<br>> we're running gcc...<br><br>The logical answer to this is: dependencies don't need to be generated<br>by the user at compile time. They need to be generated by the
<br>maintainer at release time (and occasionally by developers in CVS).<br>Assuming that the maintainer has access to "gcc", there is no<br>portability cost to generating the dependencies this way.</blockquote><div>
<br>
it wasn't acceptable if being user linked (ie called by make clean).<br>
it's more acceptable as a maintainer feature . But I still want some test<br>
to make a clear error msg if the compiler is not gcc. Thanks to complete it.<br>
<br>
</div>Arnaud</div>