[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Question about testing a library.

From: Schrader, Glenn
Subject: RE: Question about testing a library.
Date: Wed, 18 Jun 2008 16:01:00 -0400

> -----Original Message-----
> From: Ralf Wildenhues [mailto:address@hidden
> Sent: Wednesday, June 18, 2008 3:50 PM
> To: Schrader, Glenn
> Cc: address@hidden
> Subject: Re: Question about testing a library.
> * Schrader, Glenn wrote on Wed, Jun 18, 2008 at 09:10:13PM CEST:
> > I believe that I found the answer to my poorly asked question. My
> > confusion started when I noticed that target lists that begin with
> > check_ are special in that they are only built when 'make check' is
> > executed. This is reasonable since you probably want to do a complete
> > build before building any of the tests. I am using the target
> > 'installcheck' to verify that my installed libraries are working by
> > building test programs against the files installed at the install
> > prefix. It doesn't make any sense to make installcheck before making
> > install since there won't be anything at the install prefix to allow a
> > build to succeed.
> Right.  But automake doesn't ensure that you have typed 'make install'
> beforehand, either.  I think you should just tell your users, better
> document it somewhere, that they have to use 'make install' before using
> 'make installcheck'.

This is exactly what doesn't work. All programs other than the check_ programs 
are unconditionally built by the default 'all' target. There doesn't seem to be 
a clean way to defer building my tests until the user explicitly does a 'make 


> Note that automake cannot add a dependency: it is certainly possible
> that installation has to be done by a privileged user, while
> installcheck should typically be usable for nonprivileged users.
> Cheers,
> Ralf

reply via email to

[Prev in Thread] Current Thread [Next in Thread]