apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@rowe-clan.net>
Subject RE: cvs commit: apr/test testapp.c testapp.dsp testappnt.dsp aprtest.dsw
Date Fri, 31 May 2002 15:30:28 GMT

> > wrowe       2002/05/30 21:37:27
> >
> >   Modified:    test     aprtest.dsw
> >   Added:       test     testapp.c testapp.dsp testappnt.dsp
> >   Log:
> >     Some win32 experimental code I'm playing with.  Simply demonstrates
> > how  to configure an app for WinNT or all Win32 applications.  I expect
> > to  extend this simple code to actually set up and test different
> > 'signalling' APIs that are supported under Win32 (e.g. console, service, 9x
>hidden  console service, etc.)
>I'm really not sure that this belongs in the APR test directory.  This
>directory should only include programs that can be used to stress APR's
>API.  The goal of this directory is that if you can run every program in
>the directory, then your implementation of APR is correct, and will run
>all APR programs correctly.  Does this app do that, or is it Windows

Well, apr_app_initialize _is_ defined on Unix as well.  And yes - it tests
apr_app_initialize.  It could further add incorporate logic to go the extra
distance and test the API for Unicode application argument passing.  And
if args are passed as utf-8 and survive roundtrip, we are either running on
Unix or WinNT's implementation worked.  As far as Win9x, I'm not sure
how that will play out, but it should work [for user data, filenames are in
local code page convention on 9x.]

So it probably is fleshed out, but I wanted to commit that very ancient
work before I purged some copies of the cvs tree.  But I didn't add it to
the makefile list for a reason, just as I split the list in makefile for
non-portable unix tests.


View raw message