Re: raising maxclients in mod_perl test suites

2002-03-12 Thread Stas Bekman
Doug MacEachern wrote: > On Wed, 13 Mar 2002, Stas Bekman wrote: > > >>Doug, is there any special reason why we have only one server running >>for mod_perl tests? >> > > same reason as 1.x, to run all tests in the same process == better chance > of hitting bugs. 1.x also did it to test inter

Re: static readonly storage and ithreads

2002-03-12 Thread Stas Bekman
Doug MacEachern wrote: > On Wed, 13 Mar 2002, Stas Bekman wrote: > > >>Under threaded mpm, is it safe to use a static variable which is set >>once at the bootstrap time and then is used only to read from. >> > > 100% ok if the module is loaded at startup. there would be a small > race condit

Re: static readonly storage and ithreads

2002-03-12 Thread Doug MacEachern
On Wed, 13 Mar 2002, Stas Bekman wrote: > Under threaded mpm, is it safe to use a static variable which is set > once at the bootstrap time and then is used only to read from. 100% ok if the module is loaded at startup. there would be a small race condition if not loaded until request time. h

static readonly storage and ithreads

2002-03-12 Thread Stas Bekman
Under threaded mpm, is it safe to use a static variable which is set once at the bootstrap time and then is used only to read from. If it is, do I have to use the special threads function apr_thread_once_init to make sure that the initialization is run only once? e.g in Apache__Scoreboard.h:

Re: raising maxclients in mod_perl test suites

2002-03-12 Thread Doug MacEachern
On Wed, 13 Mar 2002, Stas Bekman wrote: > Doug, is there any special reason why we have only one server running > for mod_perl tests? same reason as 1.x, to run all tests in the same process == better chance of hitting bugs. 1.x also did it to test interpreter state, but we have the same_int

raising maxclients in mod_perl test suites

2002-03-12 Thread Stas Bekman
Doug, is there any special reason why we have only one server running for mod_perl tests? I need to raise it to at least two, since I need to test the fetch of the scoreboard image, which requires two parallel requests. diff -u -r1.4 TestRunPerl.pm --- Apache-Test/lib/Apache/TestRunPerl.pm

Re: auto-solib-add and threaded httpd don't play together

2002-03-12 Thread Doug MacEachern
i started seeing the problem when i moved to a new laptop/kernel. the original trick did not work for me. i ended up with the below. i use 'myrun' to start and run 'apshared' when the pre_config breakpoint is hit. haven't tried with a threaded mpm yet though. handle SIGPIPE nostop handle S

Re: TiPool: chicken and egg callback problem/question

2002-03-12 Thread Doug MacEachern
before you go any further, i meant to ask, have you looked at threads::shared yet? i don't really understand the problem you're having, but i fear there's going to be far worse. for example: what PerlInterpreter is associated with the item in the tipool? how can an item created in one inte

auto-solib-add and threaded httpd don't play together

2002-03-12 Thread Stas Bekman
Hi Daniel, Remember you've helped me at the mod_perl dev list with my gdb starting very slowly? http://marc.theaimsgroup.com/?l=apache-modperl-dev&m=100679267518055&w=2 Your auto-solib-add tip works great with preforked httpd. Now, I'm trying to debug mod_perl over threaded httpd (worker mpm) a