A lot of valid points however as i stated it is only a 2 line fix. The reason this fix is needed is because of the On Apr 12, 2013 8:53 AM, "Niphlod" <niph...@gmail.com> wrote:
> >> However why call nssm the "best" way to do this but still ship the >> winservice.py file? >> > Because things gets discussed and possibly overlooked. > There are a few open issues with the way web2py handles services and > absolutely no-one willing to take "responsibility" to test it more than the > current "once on a while". A full coverage in fact has never been tested > (and probably never worked). That being said, sometimes it works without > hiccups. The "batteries included" predicament stands as long as your > implementation works, all the times. > A very few users are using it (or a very few users just needed once and > forgot it), and when this pops up once in a while, a patch is threw in and > then the user disappears. > The moment he needs to update web2py and finds that his previous patch > didn't really solve the problem, there's another iteration on the matter. > "A few open issues?" WindowsXP scripts still work on VIsta/7/8 because it is nothing more than a simple registry entry. As far as maintaining the code, did you look at the two lines I proposed to add? from rewrite import load load() What is to maintain? the load() function in rewrite is the same one that gluon.main.py uses. If you would like to propose a better place to call the load() function in gluon.main.py such that it is always galled before launching an HttpServer then that would actually be useful and then we would not need to call load separately in the winservice.py file. > >> If nssm was truly *the* way to do this then I would think the book (which >> was just updated) would have removed the section about the winservice.py >> and only talk about nssm or at least make an effort to promote nssm over >> the native script. >> > I would have done it and proposed in the past, but I'm not the boss and > usually there are a few POVs to consider. > > The other issue with this is the fact that it is *another* thing to install rather than just being included. Makes it a lot harder to get web2py approved by corporate IT/Sys-addmins if you have to install a whole bunch of separate apps and each app has to be vetted. > This doesn't make a whole lot of sense from the standpoint of having >> "everything" in the box with web2py. >> > On the other end, doesn't make real sense to continue supporting something > that is clearly not tested and with a very little userbase (but with a lot > of people complaining for it the moment they realize they fall into the > "sometime doesn't" category). > > Clearly you are already supporting it by shipping a winservice.py file and someone just forgot to call the load() function to read the routes.py file. I don't understand what you are so against? > As far as support goes, Massimo is intent on supporting ie 7 for web2py's >> editor because some people still use it. So supporting new/different >> versions of windows is already being done not to mention the method for >> setting up a system service has hardly changed since XP (12+years) >> > same as before. My POV is more or less "if you have only IE7 as a browser, > use notepad" given that IE7 is currently being replaced in enterprises too > (and is very well below the 5% on the "browser's share"). > In any case, it's one thing to make a web page compatible with > IE6(7,8,9,10) and is another making web2py hooking up into > XP,Vista,Seven,8,2003,2008 (32 and 64 bits) service managers, with all the > options working ok, with source and binary distributions, for python2.5, > 2.6, 2.7 (someone with pypy necessities could come up as well) and have a > proof of it (i.e. not saying "this should work" but "this works"). > If you want it the way it is, i.e. "sometime works, sometime doesn't", by > all means continue to use it and send a patch every time it breaks, but > please take into consideration that ditching it and using nssm will save > you, angry users and developers a lot of headaches. > > -- > > Again, all the version of windows that are currently supported XP, Vista, 7, and 8 it makes no difference how you register a service. Even when 64 bit. Do yourself a favour and take a look at gluon.main.py and notice how out of place the call to load() is. That file is a bit of a hodgepodge, a global variable, a bunch of global functions, a class definition and a call to load(). It is clear that to make web2py more modular that the call to load() should not just be thrown in there. That being said, I do not have enough experience with the code base to propose a better place within the main.py file other than possibly just before the call to server.start in the start method of the HttpClass. I however am not certain this would always be called even when using different web-servers (e.g. calling anyserver.py and using rocket, cherrypy, gunicorn, tornado, etc.) This caused me to go a safer route and add this call to the winservice.py file instead. Are you saying this is wrong/bad/dumb? So far I can't tell what exactly it is you are against only that you are against fixing the issue. Please contribute something helpful so we can get the best solution possible available to the users, developers, and admins of web2py. Keeping in mind that part of web2pys goal is to lower the barrier of entry into developing rich applications. This is partially accomplished by including just about everything one needs in the web2py package and making it easy to deploy into production. Adding another external software requirement needlessly only raises the bar and prevents that many more people from discovering how awesome this framework can be. --- > You received this message because you are subscribed to a topic in the > Google Groups "web2py-users" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/web2py/oWmanNSKxdc/unsubscribe?hl=en. > To unsubscribe from this group and all its topics, send an email to > web2py+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to web2py+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
