Date: Sat, 26 Mar 2005 13:52:11 -0600 From: "Scott T. Hildreth" <shild@sbcglobal.net> To: Randy Pratt <rpratt1950@earthlink.net> Cc: freebsd-gnome@freebsd.org Subject: Re: audio/streamtuner core dumping (ports/79201 fixes - needs committed) Message-ID: <1111866731.35069.88.camel@fbsd1.dyndns.org> In-Reply-To: <20050326134241.7662486c.rpratt1950@earthlink.net> References: <20050319073042.469d62b0.rpratt1950@earthlink.net> <20050324132833.16b8b79d.rpratt1950@earthlink.net> <1111851679.35069.74.camel@fbsd1.dyndns.org> <20050326134241.7662486c.rpratt1950@earthlink.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Worked perfectly, thanks Randy. Next time I'll try to pay attention to what I am reading. :-) On Sat, 2005-03-26 at 13:42 -0500, Randy Pratt wrote: > On Sat, 26 Mar 2005 09:41:19 -0600 > "Scott T. Hildreth" <shild@sbcglobal.net> wrote: > > > > > > portupgrade -fR streamtuner-\* -x perl-\* > > > > > > > > > I did this on 5.3, unfortunately still segfaults. > > The above was just a correction to my original post so that anyone > searching the archives wouldn't run into the same problem with > /etc/make.conf getting undefined by using just portupgrade -fR and > not excluding perl from the rebuild. > > > > > > Secondly and probably the most important is that the maintainer, > > > ( Jean-Yves Lefort <jylefort@brutele.be> ), submitted ports/79201 > > > which I can confirm fixes the core dumping on my 4.11-STABLE: > > > > > > http://docs.freebsd.org/cgi/mid.cgi?20050324152403.38E0A8649 > > > > > > Hopefully, someone will find some time to commit this before the > > > 5.4-R rollout. > > In order to fix audio/streamtuner, you can use the patch in ports/79201 > (above link) or wait until it has been comitted and update ports. > > If you want to try to make a temporary local patch just to get > streamtuner running, here is what I did (as root): > > cd /usr/ports/audio/streamtuner > mkdir files > cd files > > Create the file "patch-src::plugins::python::pst-main.c" in > in the files/ directory with the following contents: > > ===================================================================== > --- src/plugins/python/pst-main.c.orig Thu Mar 24 11:15:11 2005 > +++ src/plugins/python/pst-main.c Thu Mar 24 11:16:24 2005 > @@ -94,6 +94,7 @@ > { > gboolean status = FALSE; > PyObject *module; > + char *argv[] = { "" }; > > if (! check_api_version(err)) > return FALSE; > @@ -109,6 +110,7 @@ > } > > Py_Initialize(); > + PySys_SetArgv(G_N_ELEMENTS(argv), argv); > PyEval_InitThreads(); > > module = PyImport_AddModule("__main__"); > ===================================================================== > > Note that the "=======" is not part of the patch file and there are > no trailing blank lines. All the patch does is to add the two lines > with a "+" to the psg-main.c file. > > Then, audio/streamtuner is rebuilt with: > > portupgrade -f streamtuner-\* > > If all went well, then streamtuner should work as expected. > > I did not modify the audio/streamtuner/Makefile as the PR does. > The version number will not change using my approach. I do this > so that if there are any differences to the final commit, the > version bump will cause it to be rebuilt with the latest patch > from the commit. > > After doing something like this locally, I make sure that the local > patches won't affect future updates in this case by doing: > > cd /usr/ports/audio/streamtuner > mv files files-my-patch > > This way, CVSup and portupgrade will ignore the "files-my-patch" > directory for subsequent updates. Usually, I remove the > "files-my-patch" directory as soon as the commit has been made. > > If you don't feel confident in juggling patches then I'd suggest > waiting on ports/79201 to be committed. > > I hope this helps more than it confuses ;-) > > Best regards, > > Randy -- Scott T. Hildreth <shild@sbcglobal.net>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1111866731.35069.88.camel>