From: Martin K. <maj...@gm...> - 2002-09-30 18:16:33
|
I wouldn't mind using another system, because I'm almost certain that Dat= a::Config is the bastard that kick me in the teeth with this one. Conver= sion shouldn't be too difficult with release 1.5 of PlainConfig, as that = one implements a write method. Just use Data::Config to read the old conf= ig file and write it back with PlainConfig. BTW. I think it's pretty funny too see that cpan is more up to date than = the site of the author: http://search.cpan.org/author/CORLISS/ParsePlainC= onfig-1.5/PlainConfig.pm Martin JT Smith wrote: > It may work. I'm not sure. >=20 > I'm wondering if it would help to switch to another config utility. I'v= e=20 > been giving serious thought to switching to the PlainConfig system. It = > has support for hashes and arrays without the complexity of XML. (If we= =20 > use an XML based configuration script then people have to intstall expa= t=20 > and a bunch of perl modules, which adds to the complexity of the instal= l.) >=20 > http://www.digitalmages.com/perl/PlainConfig/pod/PlainConfig.html >=20 > The only problem is that if we do that we'll have to distribute a confi= g=20 > converter for people to use. Not a big thing, but still extra work. Wha= t=20 > do you think of this idea? >=20 >=20 >=20 > On Mon, 30 Sep 2002 17:16:49 +0200 (MEST) > Martin Kamerbeek <maj...@gm...> wrote: >=20 >> It is already on only 100. And if this is the problem it's still prett= y >> weird that another server that I know of uses MaxRequestsPerChild=3D0 = >> (which means >> unlimited), and that beast never shows this error. The worst thing is = >> that >> there appears to be no pattern in the error showing up. >> >> But all hope has not yet gone! :) I'm gonna try to let mod_perl handle= WG >> _except_ for Data::Config. The latter I'll handle with the slow stuff.= =20 >> You >> think this might work? >> =20 >> >>> I think I have an idea of what's going on here. I don't know if you'r= e >>> aware of this, but mod_perl creates dirty threads in apache fairly=20 >>> easily and regularly. >>> To prevent this set MaxRequestsPerChild to something low like 300 (or= =20 >>> less). I'm told >>> by the Apache folks that this won't be a problem in mod_perl 2. >>> >>> >>> On Mon, 30 Sep 2002 16:15:05 +0200 (MEST) >>> Martin Kamerbeek <maj...@gm...> wrote: >>> >Hi guys, >>> > >>> >A few mothns ago I reported a bug about WebGUI not loading the >>> WebGUI.conf >>> >config ops correctly, or better said: not loading these things at=20 >>> all. At >>> >first I thought it was a problem with a crappy compiled mod_perl=20 >>> provided >>> with >>> >RedHat 7.1. >>> > >>> >To avoid this I installed Debian 3.0, and all was well. Until now...= I >>> just >>> >had my second error on this configuration. My apache error log conta= ins >>> the >>> >following, which is also shown in my browser window: >>> > >>> >[Mon Sep 30 17:04:27 2002] [error] [Mon Sep 30 17:04:27 2002] null: = >>> Can't >>> >connect( HASH(0x8e638b8)), no database driver specified and=20 >>> DBI_DSN env >>> var >>> >not set at ../lib/WebGUI/Session.pm line 216 >>> > >>> >I backtracked this to see that the >>> $session{settings}{dsn}/{dbuser}/{dbpass} >>> >vars are empty. I can only figure that the other configops in >>> WebGUI.conf >>> >(like logfile) are also not loaded. > >>> >>>From my earlier experiences with this form of frustrating distress= , I >>> know >>> >that Data::Config warned me about empty substitutions, From what I >>> remember >>> >they were in the read_source method. So I suspect Data::Config in >>> combination >>> >with my mod_perl for this. I've just put the use warnings; prama in >>> >Data::Config to get a more extensive report of what's going on in=20 >>> there. >>> > >>> >The problem is that the error appears to show up at random. At least= , I >>> >can't reproduce it. The two things I do know are: >>> > >>> > 1) it only shows up after apache is running for some time >>> > 2) when I run WG without mod_perl all works fine (but horribly = >>> slow, >>> >ofcourse) > >>> >Probably something goes wrong with mod_perls caching scheme or so. >>> > >>> >I'm not exactly a mod_perl expert, but some of you undoubtly are. Do= =20 >>> you >>> >have any clue to what's causing this bug? >>> > >>> >Martin >>> > >>> >-- >Werden Sie mit uns zum "OnlineStar 2002"! Jetzt GMX w=E4hlen - >>> >und tolle Preise absahnen! http://www.onlinestar.de >>> > >>> > >>> > >>> >------------------------------------------------------- >>> >This sf.net email is sponsored by:ThinkGeek >>> >Welcome to geek heaven. >>> >http://thinkgeek.com/sf >>> >_______________________________________________ >>> >Pbwebgui-development mailing list >>> >Pbw...@li... >>> >https://lists.sourceforge.net/lists/listinfo/pbwebgui-development >>> >>> >>> >>> JT ~ Plain Black >>> >>> Create like a god, command like a king, work like a slave. >>> >>> ------------------------------------------------------- >>> This sf.net email is sponsored by:ThinkGeek >>> Welcome to geek heaven. >>> http://thinkgeek.com/sf >>> _______________________________________________ >>> Pbwebgui-development mailing list >>> Pbw...@li... >>> https://lists.sourceforge.net/lists/listinfo/pbwebgui-development >>> >> >> --=20 >> Werden Sie mit uns zum "OnlineStar 2002"! Jetzt GMX w=E4hlen - >> und tolle Preise absahnen! http://www.onlinestar.de >> >> >> >> ------------------------------------------------------- >> This sf.net email is sponsored by:ThinkGeek >> Welcome to geek heaven. >> http://thinkgeek.com/sf >> _______________________________________________ >> Pbwebgui-development mailing list >> Pbw...@li... >> https://lists.sourceforge.net/lists/listinfo/pbwebgui-development >=20 >=20 >=20 >=20 > JT ~ Plain Black >=20 > Create like a god, command like a king, work like a slave. >=20 > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > Welcome to geek heaven. > http://thinkgeek.com/sf > _______________________________________________ > Pbwebgui-development mailing list > Pbw...@li... > https://lists.sourceforge.net/lists/listinfo/pbwebgui-development >=20 >=20 |