|
From: John K. S. <jo...@st...> - 2002-08-31 02:15:27
|
Hi Pachi - I'll have some more time to read through it all this weekend. We should be able to figure this out pretty soon. I'll let you know if I have any questions. sterling On Friday, August 30, 2002, at 06:59 AM, Pachi Sucunza wrote: > > Hi, John... > > It's me again :) > > I've been doing some more tests with mod_cplusplus and my simple > mod_HelloWorld. The results are just the same. Tried with Debian Woody, > hoping that some libraries might have been updated and would solve the > problem, but I was wrong :) > Also tried with apache 2.0.39, (and mod_cplusplus 1.0.2, in this case) > to avoid the API differences you mentioned. But again the same! > > Maybe it's a problem with apache, not with mod_cplusplus? But > everything > works OK if I don't load any c++ module... With gdb I found a more > exact > place where problems begin: it's in the fork() which apache uses to > become a daemon. Then, the parent exits (intentional and succesfully), > but the child just dies... > > Please let me know if you find anything (if you don't mind), for the > moment I need to go on working on my modules, so I'll make my tests > with > apache in single-process mode (although I read somewhere it's > discouraged...). > > Thank you very much for your help, again... > > Best regards, > > Pachi > > > On Tue, 2002-08-27 at 18:33, Pachi Sucunza wrote: >> >> Hi... >> >> Sorry, it's late suddenly (gdb is too hard for me :), I must go soon. >> Just a breif description of what I've found: >> >> running with -DONE_PROCESS isn't useful, as it works well (as with >> "httpd -X"). >> running it without it, in gdb (ddd actually), sent me to the place >> where >> it dies, it's rather surprising: >> After first initialization round (...), in main.c:617 (apache 2.0.40 >> from apache.org), when it enters ap_run_pre_config() --> ... --> in >> prefork_pre_config() (server/mpm/prefork/prefork.c), the call to >> apr_proc_detach() just makes the server end! (with return code 0 :) >> >> Maybe it's something about MPM's (although I tried several different, >> with the same result), about threads... ??? >> Any hint? >> >> Thanks again, sorry for being so brief this time... Tomorrow I'll tell >> you more... >> >> Bye! >> >> Pachi. >> > > |
