gnuclient_setup suggestion: gnuserv-start via after-init-hook (was: [Adding gnuclient_setup, removing Netinstall and InstallShield from website])

"Vin Shelton" <acs(a)alumni.princeton.edu&gt; writes: > Dear Adrian, > > On 10/13/07, Adrian Aichner <adrian(a)xemacs.org&gt; wrote: > >> > Then my extensive customizations get loaded and the font-locking >> > changes. >> > >> > And the way the scratch buffer gets initialized seems to be different >> > in some respect. >> >> I guess initial-scratch-message must be initialized too late or not >> get inserted for some reason. > > Actually, there's a race condition that I have been unable to track > down. When gnuclient starts a fresh XEmacs, it currently waits 3 > seconds before sending XEmacs a message because XEmacs can get > confused if the gnuclient command arrives too early on stdin from > gnuserv. I think your setup requires a longer wait. > > If I knew what the actual race condition is, I would track it down and > squish it. Setting a wait timeout was the easier (if less robust) > path to take. I could: > a) do nothing > b) increase the timeout > c) make the timeout configurable via a command line parameter > d) find the real bug. I haven't had much luck with this approach. > > Let me know what you think. In any case I'll release a new setup kit > replacing the reference to the testing directory - please let me know > your preference around the race condition mentioned above. Hi Vin, I'm happy to settle for (a) for now. I'll see if I can help shed some light on (d) as time goes by.

But tomorrow I'll climb a mountain instead! Have a good weekend! Adrian > > Regards, > Vin