Re: tests fail because of g
Jonathan Vanasco wrote:
I have several helper classes that use g from pylons
You have a design problem then :) pylons.g is really a substitute for
module-level singletons so you can have them associated to an app's
instance and have several of these coexisting in the same process
without conflicts (eg: several blog instances each with it's own db
connection to a different database). This is actually the same problem
you would have if you depended on global state: More difficult to
isolate tests, more difficult to set them up, etc...
If you really want to go this route you can either mock it (using the
same pattern as has been suggested to mock pylons.session) or test the
whole stack so PylonsApp takes care of setting up and tearing down the
app context on every request.
app = paste.deploy.loadapp('config:mytestconfig.ini')
app = webtest.TestApp(app)
resp = app.get('/some/test/method')
assert resp.some_test_var == 'foo'
the controller action that responds to the above url can place variables
in environ['paste.testing_variables'] and they'd be bound to the test
response that TestApp.{get,post,put,etc...} returns:
def some_controller_method(self):*
if* *'paste.testing_variables'* *in* request.environ:
request.environ[*'paste.testing_variables'*][*'some_test_var'*]
= 'foo'
However, the right way to do it, IMHO, would be to separate the
functions/classes that depend on g in two. For example, say you needed a
pylons.g.db_connection inside a helper function bar:
def side_effect_free_bar(db_connection):
# do something with db_connection and refrain from accessing global
state
# this should be pretty easy to unit-test since it's output depends
solely on it's arguments.
# You can easily pass a db_connection you've explicitly prepared for
testing
def bar():
# This one you would test inside functional/integration tests, if at
all.
return side_effect_free_bar(pylons.g.db_connection)
HTH,
Alberto
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
pylons-discuss group.
To post to this group, send email to pylons-discuss@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/pylons-discuss?hl=en
-~--~~~~--~~--~--~---
Re: tests fail because of g
hm... in this instance, i use g for three things: 1- on startup, i pull constants out of the DB and stash them into g 2- misc form classes refer to g as the values to use for validation 3- templates use them to generate dropdowns i guess i could use some sort of factory function to pull g on each form init. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: best way to get user's IP address in pylons?
On Jul 7, 1:46 pm, Ian Bicking [EMAIL PROTECTED] wrote: Christopher Weimann wrote: REMOTE_ADDR is typically set by your webserver and can be trusted. X_FORWARDED_FOR is an HTTP header typically set by proxy servers or even the client and I would have to say it can NOT be trusted so should certainly not replace the reliable data in REMOTE_ADDR. This is in part the reason why it is handled in middleware. If you *do* have a proxy in your installation then you can (and should) trust X-Forwarded-For, and moving it to REMOTE_ADDR signifies the trustworthiness of the value. If you are not behind a proxy you should ignore the value. (Arguably, maybe you should even remove the header or reject the request, but nothing currently does that.) Correct. The middleware I use under modperl works like this: Port 80 proxy- Sets X-Forwarded-For to REMOTE_ADDR Sets X-Forwarded-For-Lan to LAN_SECRET Sets REMOTE_ADDR to the lan id / 127.0.0.1 ModPerl 'middleware' Requires X-Forwarded-For headers to have: REMOTE_ADDR of specifc lan ips X-Forwarded-For-Lan with correct lan secret If those qualifications are met, then the X-Forwarded-For becomes REMOTE_ADDR. If they do not, the request is rejected. This approach ensures that data IS reliable, and GREATLY simplifies development and clustering across load balancers. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: tests fail because of g
On Mon, Jul 7, 2008 at 1:33 PM, Wichert Akkerman [EMAIL PROTECTED] wrote: What's wrong with using paste.loadapp to properly set 'g' for every test that needs it? That way your test environment is the closest to the production environment -- which is the purpose of having tests in the first place. Unit tests want as little as possible of the full environment. And you want to be able to unit test code that uses things like g. Well, unit tests (as opposed to functional tests) should test the target routine as closely as possible, without the interference of something like 'g' which should be tested separately. After all, a value on 'g' is the same as the value standalone. -- Mike Orr [EMAIL PROTECTED] --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: tests fail because of g
Previously Mike Orr wrote: On Mon, Jul 7, 2008 at 1:33 PM, Wichert Akkerman [EMAIL PROTECTED] wrote: What's wrong with using paste.loadapp to properly set 'g' for every test that needs it? That way your test environment is the closest to the production environment -- which is the purpose of having tests in the first place. Unit tests want as little as possible of the full environment. And you want to be able to unit test code that uses things like g. Well, unit tests (as opposed to functional tests) should test the target routine as closely as possible, without the interference of something like 'g' which should be tested separately. After all, a value on 'g' is the same as the value standalone. You might want to unittest a method that uses g in some way. More typically you may want to unit test something that stuffs data in c. Wichert. -- Wichert Akkerman [EMAIL PROTECTED]It is simple to make things. http://www.wiggy.net/ It is hard to make things simple. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: tests fail because of g
On Mon, Jul 7, 2008 at 1:48 PM, Wichert Akkerman [EMAIL PROTECTED] wrote: Previously Mike Orr wrote: On Mon, Jul 7, 2008 at 1:33 PM, Wichert Akkerman [EMAIL PROTECTED] wrote: What's wrong with using paste.loadapp to properly set 'g' for every test that needs it? That way your test environment is the closest to the production environment -- which is the purpose of having tests in the first place. Unit tests want as little as possible of the full environment. And you want to be able to unit test code that uses things like g. Well, unit tests (as opposed to functional tests) should test the target routine as closely as possible, without the interference of something like 'g' which should be tested separately. After all, a value on 'g' is the same as the value standalone. You might want to unittest a method that uses g in some way. More typically you may want to unit test something that stuffs data in c. True. I'm just saying if you can code to avoid 'g' and 'c' in your lib routines, so much the better. I have never found a use for 'g', except in one case to put a structure of static data that's parsed on app startup (or read from a cache pickle). -- Mike Orr [EMAIL PROTECTED] --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: tests fail because of g
On Mon, Jul 7, 2008 at 2:10 PM, Jonathan Vanasco [EMAIL PROTECTED] wrote: On Jul 7, 4:53 pm, Mike Orr [EMAIL PROTECTED] wrote: True. I'm just saying if you can code to avoid 'g' and 'c' in your lib routines, so much the better. I have never found a use for 'g', except in one case to put a structure of static data that's parsed on app startup (or read from a cache pickle). That is *exactly* what I use g for -- stashing 'static' data ( the names and ids of 'options' that the database fkeys on . example: query= select * from _gender for row in rows: g.gender.id[ row.id ]= row.name g.gender.name[ row.name ]= row.id i also have some other misc. config data -- like amazon s3 urls, etc - which are initialized on startup based out of config values. i'm a little less-than-enthusiastic by the built in testing with Pylons. not having g,c,session is annoying. You're not initializing the application completely and you're wondering why it isn't working? Don't you see the contradiction there? i've also noticed that for every test it does a full app initialization - which means ( in my case ) querying the database for 50 tables worth of data. Unit tests always do that: setUp creates a full test environment, which is discarded after each test. Generally that's considered a good thing even if it's slow. Python apps have been testing this way since long before Pylons was created. If you have to read 50 tables worth of data, it's more a case that you have an unusual application structure than that Python's testing is screwed up. You may have to test this application by pre-reading the data at module level and putting it in place. (And if it's place is in pylons.g, maybe have some lightweight g-initialization routine that copies the variables from their test module locations. There's a module_setup function that may persist between tests, although I'm not sure. -- Mike Orr [EMAIL PROTECTED] --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
supervisor questions
Hi, I'm looking at http://wiki.pylonshq.com/display/pylonscookbook/Monitor+Pylons+application+with+supervisord. I have a few questions: 1. Why does it use a manually created server.py instead of using paster like normal? 2. Does anyone have an rc script to start supervisor under Ubuntu? 3. Is anyone doing any fancy log rotation or are the defaults reasonable? Thanks, -jj -- It's a walled garden, but the flowers sure are lovely! http://jjinux.blogspot.com/ --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: A new SQLAlchemy migration toolkit - miruku 0.1a3 has been released
On Thu, Jul 3, 2008 at 6:59 PM, Olli Wang [EMAIL PROTECTED] wrote: On Jul 4, 9:42 am, Shannon -jj Behrens [EMAIL PROTECTED] wrote: How is it possible to not require an upgrade script? What happens if I'm changing an email field into two fields, one for the username and one for the server? -jj -- It's a walled garden, but the flowers sure are lovely!http://jjinux.blogspot.com/ What miruku does is finding the difference between original schema and modified schema, then upgrade the original schema to the modified schema (add/drop columns/tables, etc). In your question, it seems you just need to add a new email column which may be located in the same or different table. It doesn't matter, just change the model schema code by add a new email column and run miruku upgrade command. Then you will have two email field. However, currently miruku only support add/drop tables/columns/index. So there is no way to copy the email value to the new field automatically. But I think it's a good idea to work it our in the future. :) I think it's an 80/20 thing. You got 80% of what everyone needs by writing 20% of the code, and, as usual, I'm obsessed with the remaining 20% ;) Happy Hacking! -jj -- It's a walled garden, but the flowers sure are lovely! http://jjinux.blogspot.com/ --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: supervisor questions
Shannon -jj Behrens wrote: Hi, I'm looking at http://wiki.pylonshq.com/display/pylonscookbook/Monitor+Pylons+application+with+supervisord. Huh... I'm getting a 404...? I have a few questions: 1. Why does it use a manually created server.py instead of using paster like normal? I use paster serve (no --reload, --monitor, --daemon, any of which would mess it up). 2. Does anyone have an rc script to start supervisor under Ubuntu? Huh, actually no. It did occur to me that supervisor should ship with such a thing. Incidentally we use a per-site supervisord. 3. Is anyone doing any fancy log rotation or are the defaults reasonable? The defaults have seemed reasonable to me (though there's no cleanup or compression AFAIK, just rotation). -- Ian Bicking : [EMAIL PROTECTED] : http://blog.ianbicking.org --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: supervisor questions
Ian Bicking wrote: 1. Why does it use a manually created server.py instead of using paster like normal? I use paster serve (no --reload, --monitor, --daemon, any of which would mess it up). Incidentally, one thing I don't think is all that well explained (at least in the supervisor docs), is that shell scripts must end with: exec paster serve production.ini So, if you want a shell script (e.g., to setup the environment) you have to be sure to use exec at the end; this makes paster replace the shell process. -- Ian Bicking : [EMAIL PROTECTED] : http://blog.ianbicking.org --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: distribute Pylons app with bytecode files only?
Hi Phlee, it is true that it is possible to decompile bytecode python but it is also true that it is possible to decompile or disassemble any language. That in itself does not make it worthless for two reasons: The first is that it is illegal (if you say so in the license). How easy it is to open the front door of your house? that does not mean anybody does it. People have sued and have won trials in similar circumstances. VB and C# sell because they claim compilation protection but it is not better than in Python. In the specific case of web2py. web2py does one more step before bytecode compiling. It translates all view hierarchy for every action into a single python file (resolves all extended and included templates into one view, and then tuns the html into python code), the it bytecode compiles it. It manly does this to serve pages faster without parsing and string substitution at runtime. This also means that even if they were to decompile it, they would not get exactly the original code and what they'd get would not be too easy to turn into a working web2py application. I am here temporarily so if you have questions we have a google list for web2py. Massimo P.S. Sorry for intruding but I thought this was relevant. Pylons rocks! --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: best way to send email w/o blocking?
On Jul 6, 12:35 am, kevin [EMAIL PROTECTED] wrote: What are your recommendations for sending email without blocking? Because I can imagine several ways, but maybe there's some niceness I'm not aware of. I actually spawn off a new thread that does some verification and then sends the email and exits. It's a bit heavyweight, I admit, but it allows you to take as much time as you want without interfering with the user experience. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
Re: Debugging failed tests with pylons globals
On Mon, Jul 7, 2008 at 10:04 PM, Justin Tulloss [EMAIL PROTECTED] wrote: I have to admit that despite my digging through the source, I do not have a good understanding of the StackedObjectProxy class and am constantly doing battle with it. I don't think there's anybody who doesn't do battle with StackedObjectProxy. -- Mike Orr [EMAIL PROTECTED] --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-discuss@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en -~--~~~~--~~--~--~---
