From: BlaisorBlade <bla...@ya...> - 2004-07-01 19:26:44
|
Alle 20:12, gioved=EC 1 luglio 2004, Martin Maney ha scritto: > > It is the 10-flock one. > > And it goes back to 24-2 at least? Strange that I've never seen it. No, you must *APPLY* that patch to solve the problem. > Do you know offhand what would make the #if USE_FCNTL_LOCK be false, as > seems to be what triggers it? I'd enjoy knowing whether I've been > lucky all this time or if there's a good reason for it. I think that USE_FCNTL_LOCK is always undefined, just provided for who want= s=20 it. And I think it is not applied. > Hmmm... can't find any occurence of USE_FCNTL_LOCK in either the 24-um2 > or 26-1 (configured) trees here. os_lock_file contains no conditional > code, only the fcntl-using version. Is this a bug introduced by some > other patch, not Jeff's patch set? No, the Jeff's patch have the bug (actually it's in the host), and that pat= ch=20 fixes it. =2D-=20 Paolo Giarrusso, aka Blaisorblade Linux registered user n. 292729 |