configure in qucs

[felix-salfelder]

• Romano Giannetti found a whitespace issue and proposed a fix.
• list pkg-config as a dependency.

[felix-salfelder]

On Wed, Jul 25, 2018 at 09:38:41AM -0700, in3otd wrote:

in3otd commented on this pull request.

@@ -36,6 +36,8 @@ Requirements

Qucs needs Qt(R) by Digia(R) http://www.digia.com.

not related but Qt development is now done by the Qt Group Plc, https://www.qt.io :grin:

yes, thanks.

[guitorri]

No need to update qucs-test.

[felix-salfelder]

On Wed, Jul 25, 2018 at 11:58:32AM -0700, Guilherme Brondani Torri wrote:

No need to update qucs-test.

thanks, fixed.

(still don't understand how this always happens since recently. i never touched any of this...)

[guitorri]

If you don' t git update submodule and git add -a you also pick up the old qucs-test hash.

The QtTest is needed if you run make check.

[felix-salfelder]

On Wed, Jul 25, 2018 at 07:35:42PM +0000, in3otd wrote:

 dnl Enable C++11 support
 CXXFLAGS="$CXXFLAGS -std=c++0x"

• PKG_CHECK_MODULES([QTest], [${qt_name}Test ])
• PKG_CHECK_MODULES([QTest], [${qt_name}Test])

by the way, any idea why this works with and without the space on my system and on Travis too?

don't know, perhaps it's actually a bug in an older pkg.m4. I wouldn't have put the space in, if it had not worked for me..

Trying to understand what PKG_CHECK_MODULES does, I saw that "this guy" and others discourage its usage; FWIW, I knowing nothing about it.

interesting article, I never gave it a thought. otoh, if PKG_CHECK_MODULES only induces trouble once every 10+ years, and only due to stupid whitespace, then why bother? it's just as easy to get a shell script wrong in a subtle way, YMMV.