[mythtv] Update PVR 350 patch

[Brad Allen]

From: Isaac Richards <ijr at po.cwru.edu>
Subject: Re: [mythtv] Update PVR 350 patch
Date: Wed, 14 Jan 2004 13:30:19 -0500

> On Wednesday 14 January 2004 01:07 pm, David Engel wrote:
> > On Wed, Jan 14, 2004 at 10:55:53AM -0500, Carl wrote:

> > > Prescient observation, except the rant showed up on the ivtv
> > > list... :-)
> >
> > I saw it.  Perhaps myth should print the message and then just
> > exit to make it more obvious.

That would work.

> Then people will just write in reporting a segfault and still not
> bother to read the error message =)

Oh please.  I would have read it *then*.

You know, I read that message that Dave Engel wrote, thinking to
myself, "now why would anybody be so blathering stupid as to do THAT?
It clearly says to get CVS latest from both", so I made sure to get
latest CVS from IVTV and Myth and compile it all.  Only thing was, CVS
wasn't as fast as me, for some reason.

So I didn't read the failure message he wrote about here.

However, there is a lot of resistance to reading the output of
mythfrontend:

1.  I have seen it complain and babel about a lot of things, and almost
    never anything I can do anything about or understand.  Mostly it
    just is whiny and babbling.  I suppose there are occasional
    exceptions, but since I got latest CVS (or thought I did --- one
    hour early) I didn't even think to check.

2.  I fixed this right now:  I have it in a login script that just
    left the output in .xsession-errors.  It was late and I didn't
    feel like figuring out where those file descriptors ended up.

3.  It may well have crashed the computer before I could go back
    to read what it had hidden underneath its own OSD for the times
    I ran it on a command line.

Due to #3 and #2 above, I think it should automatically behave like
most other big motormouth programs which diagnose off their main
screen (XFree86 is an example) and put all its diagnosis output in a
file, and I think it should also do it to stderr, at the same time.
Also, I like the diagnosis standard XFree86 uses: it flags the
beginning of each line with its meaning, with II for informational and
EE for error.


I can't help but wonder why D.E. put that dumb check in there:  it
would have worked for me if he hadn't, since I did have latest CVS,
with the only change being that blathering 0.1.9 version # coming
after my compile.