I too am *very much* interested in this feature change, and I hope it will be slipped into the next release (Deer Park) relatively easy.
Sadly, there doesn't apear to be any other alternative to getting socket state events back up to a XUL/JavaScript extension without turning off this coalesce feature by recompiling the source. This obviously makes it impossible to distribute a firefox/mozilla extension to the average user that requires this option to be turned off. Longterm, making instrumented timings available for state changes at the socket level would be a wonderful addition. But for the immediate short term, Boris' & Leroy's suggestion of adding a simple boolean pref is an ideal solution, since that would allow us to temporarily toggle the coalesce option on & off in the extension's javascript code. Until this is available as a pref, my project is dead in the water. Perhaps we could have Darin Fisher weigh in and let us know if we can squeek a prefs change into the next release for this. LeRoyK wrote: > Boris Zbarsky wrote: > > LeRoyK wrote: > > > >> This change would slow down every nsHttpTransaction:Init() call by the > >> time it takes to check a preference. > > > > > > Actually, if you cached the pref value (on nsHttpHandler, probably, and > > made sure the pref is observed by the handler by adding things in > > nsHttpHandler::Init as needed), there wouldn't have to be a pref lookup > > here; just a quick boolean comparison. > > > >> I should be able to make the changes to the source myself, but being a > >> total newbee, I am not sure of how to get the approvals needed to make > >> it into the code base. > > > > > > File a bug, post a patch, request review from the module owner by > > setting the review flag on the attachment, request super-review once you > > have review, then ask someone (probably one of the reviewers) to check > > it in. The process is described at > > http://www.mozilla.org/hacking/life-cycle.html and the module owners are > > listed at http://www.mozilla.org/owners.html (for this change you'd want > > Darin Fisher). > > > > Let me know if you run into problems, > > -Boris > > Thanks Boris, > > I saw where nsHttpHandler monitored most of the network.http.* prefs. > > I will solicit Darin's opinion about whether he thinks this change is a > good thing to do; and if he has a suggestion about how to best link the > preference from nsHttpHandler to nsHttpTransaction::init > > Thanks again, > > LeRoy _______________________________________________ Mozilla-netlib mailing list Mozilla-netlib@mozilla.org http://mail.mozilla.org/listinfo/mozilla-netlib
