home *** CD-ROM | disk | FTP | other *** search
- x-gateway: rodan.UU.NET from bug-lucid-emacs to alt.lucid-emacs.bug; Tue, 5 Jan 1993 17:53:10 EST
- Date: Tue, 5 Jan 1993 14:50:50 PST
- Message-ID: <9301052250.AA19147@thalidomide.lucid>
- X-Windows: Sometimes you fill a vacuum and it still sucks.
- From: jwz@lucid.com (Jamie Zawinski)
- Sender: jwz%thalidomide@lucid.com
- Subject: Re: Shouldn't the audio part be moved out of lemacs?
- References: <9301052055.AA18807@thalidomide.lucid>
- <Pine.3.05.9301051642.D17104-c100000@trillian>
- Newsgroups: alt.lucid-emacs.bug
- Path: sparky!uunet!wendy-fate.uu.net!bug-lucid-emacs
- Lines: 58
-
- In message <Pine.3.05.9301051642.D17104-c100000@trillian> Maximilian Ott wrote:
- >
- >> but the system doesn't provide such a program: you'd have to write the
- >> audio-handling code yourself, which is the hard part. So why not just link
- >> it in to emacs directly and avoid the overhead of having to fork a process?
- >
- > You have to read in the sound data at the beginning anyway. So you could
- > start the audio client by handing over a list of sound files and later
- > use symbolic names. The client then could preload the sound data.
-
- You're still just talking about scenery. The same work needs to be done; the
- same code needs to be written, since it doesn't exist yet. Whether this code
- is linked directly into the emacs process or is running in some other process
- is largely inconsequential to the amount of work that needs to be done.
-
- Making it so that the code can be in an external process is fraught with
- problems. Doing it that way might build some false sense of generality, so
- that one could imagine that emacs was theoretically compatible with a number
- of different audio-server protocols, none of which exist yet, but this hardly
- seems worthwhile.
-
- > The big problem I have with the current setup is that on a sun a program
- > needs exclusive access to the audio port.
-
- But this is a misfeature of the /dev/audio device driver, not of Emacs, right?
-
- > I'm working on shared environments where I use the built-in audio port
- > for audio-conferencing. During the time another program is using the
- > audio port emacs is "dumb". (In fact I get a few error messages.)
-
- I don't think it's dumb. Given that Sun's audio device can't be shared, what
- would you have emacs do when it wants to play a sound and the kernel won't let
- it? Right now it prints a message in the minibuffer. That message could be
- suppressed, but so what: the problem still exists, and it's not a problem that
- emacs can solve. You want a better device driver.
-
- > Don't get me wrong, it would be preferable to link in a library which
- > connects to an audio server ala "rplayd". But there is no standard, yet.
- > Forking a separate process gives you enough flexibility to do whatever you
- > want without needing to relink emacs everytime.
-
- The flexibility that *who* wants? I think that the only people that this
- would benefit are those who are implementing new audio protocols, and the only
- real benefit to them is that they don't have to relink emacs to debug their
- code. Maybe I'm misunderstanding what it is you're trying to accomplish?
-
- > However, the issue of sound and X is another argument for separating the
- > sound so we can experiment with sound servers until we get that capability
- > in X. (As I said before "rplay[d]" is a good first step)
-
- The emacs sound code and the emacs X code don't have anything to do with
- each other. They never have.
-
- The MIT people have said in no uncertain terms that audio protocols will
- never be a part of the X standard. They're of the opinion that windows and
- sounds are orthogonal.
-
- -- Jamie
-