[01:48] *** bcarlyon|ubuntu_ has joined #kamaelia

[06:39] < Lawouach_> morning folks

[07:31] *** mhrd has joined #kamaelia

[07:32] *** mhrd forgot to mention - am away yesterday/today/tomorrow (project meetings abroad)

[07:32] *** mhrd shocked at the sudden deluge of commits :-)

[07:32] < mhrd> I'll loiter here a little while, but'll have to go in a bit

[07:59] *** bcarlyon|laptop has joined #kamaelia

[08:24] *** orphans has joined #kamaelia

[08:25] *** vmlemon_ has joined #kamaelia

[08:26] *** vmlemon_ has joined #kamaelia

[08:30] *** PJ_Coudert has joined #kamaelia

[08:53] *** Chong- has joined #kamaelia

[09:04] *** mhrd has parted #kamaelia

[09:06] *** Uraeus has joined #kamaelia

[09:12] < ms-afk> "mhrd shocked at the sudden deluge of commits"

[09:12] < ms-afk> -- caused by commits at the weekend, but no emails being sent by sourceforge. Looks like they had mailing list constipation

[09:14] *** bcarlyon|laptop has parted #kamaelia

[09:15] < orphans> there's a lovely image to start the day with :)

[09:30] *** bcarlyon|laptop has joined #kamaelia

[09:53] *** Davbo has joined #kamaelia

[10:13] *** Uraeus has joined #kamaelia

[10:40] *** bcarlyon|laptop has joined #kamaelia

[10:41] *** bcarlyon|laptop has parted #kamaelia

[10:55] < Davbo> ms-afk: I can't get the Logger to connect to another server anymore

[10:55] < Davbo> not sure if it's just a problem on my end

[10:55] < Davbo> other than that it seems fine for merge, what do you think. Shall I merge?

[10:56] < Davbo> I mailed the Mailing List about it but it hasn't shown up for some reason

[11:24] < ms-afk> What server are you trying to connect to>

[11:24] < ms-afk> ?

[11:24] < vmlemon> Hi

[11:28] < Davbo> Sorry ms-afk!

[11:28] < Davbo> i've got it connected to irc.quakenet.org but it didn't join the channel

[11:28] < vmlemon> Aaw, http://gizmodo.com/assets/images/gizmodo/2008/06/SnowLeopard3.jpg

[11:31] < Davbo> I thought it might be because IRCClient uses this format

[11:31] < Davbo> print "/user testingkbot irc.freenode.net 127.0.0.1 username"

[11:31] < Davbo> but the logger is doing

[11:31] < Davbo> self.send(("USER", self.name, self.name, self.name, self.name), "irc")

[11:31] < Davbo> I was just about to look into that for a solution

[11:53] < Davbo> Meh, works on EFNet

[11:54] < Davbo> Changed it so you can put host in the config file though

[11:55] < Davbo> maybe that quakenet server was unorthodox

[12:48] < Davbo> Could someone help me merge when they've got time please

[12:48] < Davbo> I don't want to break svn

[12:56] < ms-afk> If you've made changes, you ought to do a

[12:56] < ms-afk> svn diff -r< last version you didn't change>: HEAD .

[12:56] < ms-afk> summarise the changes you've made and post those changes, along with the diff line (with the right version number) to the list

[12:57] < ms-afk> Along with any observations on the rest of the code :)

[12:57] < ms-afk> It's worth bearing in mind though

[12:57] < ms-afk> That svn merge doesn't automatically check in, so you can only break your own repository

[12:57] < ms-afk> If you're worried about that,

[12:57] < ms-afk> cp -R /trunk /trunk.backup.just.in.case

[12:58] < Davbo> I was trying to do that ms-afk but I get a huge diff with things i didn't change

[12:58] < ms-afk> Where are you doing this:

[12:58] < ms-afk> svn diff -r< last version you didn't change>: HEAD .

[12:59] < ms-afk> If you're doing it on the branch, then it should be fine

[12:59] < ms-afk> (should)

[12:59] < ms-afk> (ie short)

[12:59] < ms-afk> ie really need

[12:59] < ms-afk> ~/kamaelia/branches/private_MPS_Scratch> pwd

[12:59] < ms-afk> /home/michael/kamaelia/branches/private_MPS_Scratch

[12:59] < Davbo> pj

[13:00] < Davbo> oh*

[13:00] < ms-afk> and that sort of detail (pastebin is your friend)

[13:01] < ms-afk> After all, if "I was trying to do < thing> but I get < other thing>", really need to be able to "see" what you're doing when you do < thing>

[13:01] < Davbo> [root@localhost private_MPS_KamaeliaLogger]# svn update

[13:01] < Davbo> At revision 4337.

[13:01] < Davbo> [root@localhost private_MPS_KamaeliaLogger]# svn diff -r4337:HEAD

[13:01] < ms-afk> and see results

[13:01] < Davbo> I'm totally doing this wrong aren't i

[13:01] < ms-afk> ta

[13:01] < ms-afk> not necessarily :-)

[13:01] < Davbo> it's empty

[13:01] < ms-afk> Yes, that is.

[13:01] < ms-afk> The reason is because "HEAD" is a special version

[13:02] < ms-afk> it means "the current last revision for this branch"

[13:02] < ms-afk> Since "At revision 4337."

[13:02] < ms-afk> that means you're doing

[13:02] < ms-afk> svn diff -r4337:4337

[13:02] < Davbo> oh.

[13:02] < ms-afk> which is asking for the diff between 2 versions which are the same :)

[13:02] < Davbo> i see lol

[13:02] < ms-afk> so the diff should be empty

[13:02] < Davbo> How do i see what revision it was before I changed anything?

[13:02] < ms-afk> hold on

[13:03] < ms-afk> ~/kamaelia/branches/private_MPS_KamaeliaLogger> svn update

[13:03] < ms-afk> U Apps/Kamaelia-Logger/App/KamaeliaLogger.py

[13:03] < ms-afk> U Apps/Kamaelia-Logger/Config/kamaelia_logger.conf.dist

[13:03] < ms-afk> U Apps/Kamaelia-Logger/Doc/README

[13:03] < ms-afk> Updated to revision 4337.

[13:03] < ms-afk> ~/kamaelia/branches/private_MPS_KamaeliaLogger> svn log --stop-on-copy

[13:03] *** ms-afk reads through that

[13:03] < ms-afk> To make clearer what I'm looking at:

[13:04] < ms-afk> ~/kamaelia/branches/private_MPS_KamaeliaLogger> svn log --stop-on-copy|grep ^r

[13:04] < Davbo> oh i see

[13:04] < ms-afk> r4335 | davbo | 2008-06-10 13:36:43 +0100 (Tue, 10 Jun 2008) | 1 line

[13:04] < ms-afk> r4334 | davbo | 2008-06-10 13:34:52 +0100 (Tue, 10 Jun 2008) | 1 line

[13:04] < ms-afk> r4333 | davbo | 2008-06-10 13:20:16 +0100 (Tue, 10 Jun 2008) | 1 line

[13:04] < ms-afk> r4332 | davbo | 2008-06-10 13:15:51 +0100 (Tue, 10 Jun 2008) | 1 line

[13:04] < ms-afk> r4313 | davbo | 2008-06-09 17:56:38 +0100 (Mon, 09 Jun 2008) | 1 line

[13:04] < ms-afk> r4312 | davbo | 2008-06-09 14:59:38 +0100 (Mon, 09 Jun 2008) | 1 line

[13:04] < ms-afk> r4255 | ms_ | 2008-06-08 21:38:24 +0100 (Sun, 08 Jun 2008) | 4 lines

[13:04] < Davbo> yeah

[13:04] < ms-afk> So the last version you didn't change was r4255

[13:04] < Davbo> yep

[13:04] < ms-afk> ~/kamaelia/branches/private_MPS_KamaeliaLogger> svn diff -r4255:HEAD

[13:04] < ms-afk> Is what you've been doing :-)

[13:05] < Davbo> thank you ms-afk

[13:05] < Davbo> I'll write this up on to the mailing list

[13:05] < ms-afk> It's *really* handy if someone says "what have you done to this branch" ;-) :-D

[13:05] < ms-afk> Please do write it up - I (like others) have many hidden assumptions

[13:06] < ms-afk> Lines that look like this:

[13:06] < ms-afk> @@ -156,6 +156,7 @@

[13:06] < ms-afk> Show specific changes, which can be really useful for summarising what's going on

[13:07] < ms-afk> eg:

[13:07] < ms-afk> +++ Apps/Kamaelia-Logger/App/KamaeliaLogger.py (revision 4337)

[13:07] < ms-afk> @@ -156,6 +156,7 @@

[13:07] < ms-afk> yield 1

[13:07] < ms-afk> while True:

[13:07] < ms-afk> + time.sleep(0.1) # better way to do this?

[13:07] < ms-afk> if self.currentDateString() != self.lastdatestring:

[13:07] < ms-afk> self.changeDate()

[13:07] < Davbo> Right

[13:07] < ms-afk> The question is what is that doing?

[13:07] < ms-afk> etc :)

[13:08] *** Trun has joined #kamaelia

[13:11] < Davbo> oh one question ms-afk

[13:11] < ms-afk> ?

[13:11] < Davbo> I thought it would be better for the user to see the default's the bot took, so i uncommented them in the config file, so those values are taken by default over the hardcoded ones.

[13:12] < Davbo> i thought a user might wonder why it connected to freenode if the config was all commented out

[13:13] < ms-afk> Trun: Your checkin messages should be proportional to the size of change

[13:13] < ms-afk> * Refactored KamTestCase (it does not inherit unittest.TestCase anymore but still delegates everything on it)

[13:13] < ms-afk> * KamMockObject is now "private" for the user

[13:13] < ms-afk> * More tests added, a couple of bugfixes

[13:13] < ms-afk> Doesn't really actually assist in following what you are working on

[13:14] < Davbo> you'll see what i mean with the diff ms-afk

[13:14] < ms-afk> Davbo: Ta :)

[13:14] < Davbo> :)

[13:14] < Trun> hmm... you mean by more doc in the messages or you mean by shorter commits?

[13:15] < ms-afk> More commite more often which are smaller

[13:15] < Trun> ok, sorry for that

[13:15] < ms-afk> Let me put it this way

[13:15] < ms-afk> r4158 | nctrun | 2008-06-03 02:19:11 +0100 (Tue, 03 Jun 2008) | 7 lines

[13:15] < ms-afk> r4100 | nctrun | 2008-05-29 02:17:35 +0100 (Thu, 29 May 2008) | 6 lines

[13:15] < ms-afk> r4075 | nctrun | 2008-05-27 01:57:35 +0100 (Tue, 27 May 2008) | 5 lines

[13:15] < ms-afk> r4053 | nctrun | 2008-05-24 20:46:17 +0100 (Sat, 24 May 2008) | 3 lines

[13:16] < ms-afk> gah

[13:17] < ms-afk> r4338 | nctrun | 2008-06-10 14:12:16 +0100 (Tue, 10 Jun 2008) | 5 lines

[13:17] < ms-afk> r4191 | nctrun | 2008-06-05 15:33:40 +0100 (Thu, 05 Jun 2008) | 8 lines

[13:17] < ms-afk> r4185 | nctrun | 2008-06-05 03:17:25 +0100 (Thu, 05 Jun 2008) | 4 lines

[13:17] < ms-afk> r4158 | nctrun | 2008-06-03 02:19:11 +0100 (Tue, 03 Jun 2008) | 7 lines

[13:17] < ms-afk> r4100 | nctrun | 2008-05-29 02:17:35 +0100 (Thu, 29 May 2008) | 6 lines

[13:17] < ms-afk> The gaps between those checkins are very large

[13:17] < ms-afk> generally speaking

[13:17] < ms-afk> So between:

[13:17] < ms-afk> r4338 | nctrun | 2008-06-10 14:12:16 +0100 (Tue, 10 Jun 2008) | 5 lines

[13:17] < ms-afk> r4191 | nctrun | 2008-06-05 15:33:40 +0100 (Thu, 05 Jun 2008) |

[13:17] < ms-afk> You've got 1 line of summarising what you're doing per day of work

[13:18] < ms-afk> It's *impossible* to keep track of what you're doing if you only communicate 1 line of information for an entire day's work

[13:18] < ms-afk> (which is a huge shame because you're doing lots :) )

[13:18] < ms-afk> * More tests added, a couple of bugfixes

[13:19] < ms-afk> is actually quite bad as well - what tests? What bugs were fixed?

[13:19] < ms-afk> If you fix a bug, it should get checked in

[13:19] < ms-afk> Compare my checkin log:

[13:19] < ms-afk> http://cia.vc/stats/author/ms_

[13:20] < ms-afk> You can see quite clearly what's changing

[13:20] < Trun> yeah, I see

[13:21] < Trun> "If you fix a bug, it should get checked in" -> commit then as the bug is fixed, with explanation of the bug?

[13:21] < ms-afk> definitely

[13:21] < ms-afk> even a one line explanation :-)

[13:21] < Trun> ok :-)

[13:21] < ms-afk> svn ci filename.py -m "fixed missing comma, which destroys your filesystem if missing"

[13:22] *** ms-afk wonders

[13:23] *** ms-afk tries somethings

[13:27] < ms-afk> (sorry was interrupted)

[13:28] < ms-afk> OK, if there's a problem due to language, then this is works:

[13:28] < ms-afk> http://translate.google.com/translate?u=http%3A%2F%2Fcia.vc%2Fstats%2Fauthor%2Fms_&hl=en&ie=UTF8&sl=es&tl=en

[13:28] < ms-afk> which means that if you find it quicker, you can checkin in spanish.

[13:28] < ms-afk> BUT:

[13:28] < ms-afk> and its a huge but

[13:29] < ms-afk> If you DO do that, please ensure that you post that english language version to EITHER the commits mailing list

[13:29] < ms-afk> or to the google group

[13:29] < ms-afk> *before* you hit 20 versions

[13:29] < ms-afk> Preferably with just the check in messages !

[13:29] < ms-afk> and details

[13:30] < Trun> mmm... I think that I prefer to keep checking in in english, I'll place more efforts on writing the message

[13:31] < ms-afk> I'd prefer it too - mainly because I'd be utterly hopeless without that auto-translater

[13:31] < ms-afk> :)

[13:31] < Trun> yeah, I know that my english is not very good :-) but I'll try to place more effort on the checkins

[13:32] < ms-afk> Thank you :)

[13:33] < ms-afk> It's worth remembering that as well as the checkin emails, that due to the cia.vc site, it's possible for someone to see what you're doing more clearly

[13:33] < Davbo> my diff is up on the mailing list ms-afk

[13:33] < Davbo> (no rush or anything, just letting you know)

[13:33] < ms-afk> Davbo: ta

[13:35] < ms-afk> Trun: So, can you give me a bit more detail than just the 8 bullet points as to what you've been doing?

[13:35] < ms-afk> In case you're wondering why I ask, let me translate your checkin messages into "trying to track what's going on ese"

[13:35] < ms-afk> (inteneded to be slightly humourous translation)

[13:36] < Trun> sure, you mean now? (want to use #kamaelia-mentor ?)

[13:36] < ms-afk> Changed code in < some file> (but not how or why)

[13:37] < ms-afk> [ yes, now (after I show you how your check messages read) ]

[13:37] < Trun> [ ok ]

[13:37] < ms-afk> < code> is now "private" for the user (what does this actually mean in practice? How is it private ? Why is it private even ? )

[13:37] < ms-afk> More tests added, a couple of bugfixes -> "I wrote code, but you'll have to rummage to find out what"

[13:38] < ms-afk> "Added libraries to KamPlanet's "libs" folder to make it far easier to run the application" --> "I added more code, but not saying what"

[13:39] < ms-afk> "Added "libraries.py" to import these libraries, runTests, KamPlanet.py and runSingleTest use it" --> Um, that's just (almost literally) "I added some code. I'm not saying what it does though"

[13:39] < ms-afk> "Migrated from pmock to the "Python Mocker" module (http://labix.org/mocker) ** Reason: pmock is available in repositories (such as ubuntu or debian), but it will be incompatible with Python 2.6 & py3k.

[13:39] < ms-afk> Contacted the author and he confirmed that it's not going to be maintained anymore :-( ** Little class that added a feature to pmock removed since "Python Mocker" already supports it"

[13:39] < ms-afk> That's something that should be discussed on the list

[13:39] < ms-afk> (no machine I use uses ubuntu or debian)

[13:40] < ms-afk> (also it becomes a barrier to using your code, so needs discussion really)

[13:40] < ms-afk> "Adding more test all over KamPlanet (most units already have tests, although they still don't test every situation)"

[13:40] < ms-afk> --> "I wrote lots of code. "

[13:40] < ms-afk> "Changes in KamMockObject and KamTestCase, method renaming and more features" --> "I changed some things"

[13:41] < ms-afk> Now, that's meant lightheartedly

[13:41] < ms-afk> But

[13:41] < ms-afk> That means your checkin messages say this to me:

[13:43] < ms-afk> Changed code in < some file> (but not how or why)

[13:43] < ms-afk> < code> is now "private" for the user

[13:43] < ms-afk> I wrote code, but you'll have to rummage to find out what

[13:43] < ms-afk> I added more code, but not saying what"

[13:43] < ms-afk> I added some code. I'm not saying what it does

[13:43] < ms-afk> < discussion of something that should be on IRC or the mailing list - probably the latter>

[13:43] < ms-afk> "I wrote lots of code. "

[13:43] < ms-afk> "I changed some things"

[13:43] < ms-afk> Which is not really that easy to follow

[13:43] < ms-afk> Trun: You see ?

[13:43] < ms-afk> :)

[13:43] < Trun> yeah, definetely, I see

[13:43] < Trun> s/definetely/definitely/ :-)

[13:44] < ms-afk> It's a really nasty danger that people fall into with checkin messages - they don't state specifically what they've changed, merged, fixed, broken, etc

[13:45] < ms-afk> The other thing is did you get my email saturday?

[13:45] < Trun> yeah... it's definitely a bad habit I have :-) sorry for that

[13:45] < Trun> yeah

[13:46] < Trun> I understand it and accept it

[13:46] < ms-afk> OK, simply telling me that would've helped btw :)

[13:46] < ms-afk> I've currently got no idea if your code / approach/ideas will fit in with the current testing & logging frameworks that we use

[13:47] < ms-afk> or those that check nightly for problems, or rebuild docs etc

[13:47] < ms-afk> since you've not really discussed your ideas on the list

[13:49] < Trun> yeah, I guess I should have written about it some time ago, may I tell now [here]?

[13:49] < ms-afk> yes, also, for information:

[13:50] < ms-afk> ~/kamaelia/trunk/Tests/Python/Axon> for i in test*py; do echo $i; echo $i|sed -e "s/./=/g"; ./$i -v; done

[13:50] < ms-afk> That'll run the Axon test suite and document what works.

[13:51] < ms-afk> (go on)

[13:52] < ms-afk> (14:41:38) Trun: sure, you mean now? (want to use #kamaelia-mentor ?)

[13:52] < ms-afk> (14:42:23) ms-afk: [ yes, now (after I show you how your check messages read) ]

[13:52] < ms-afk> (14:42:43) Trun: [ ok ]

[13:52] < ms-afk> (14:54:26) Trun: yeah, I guess I should have written about it some time ago, may I tell now [here]?

[13:52] < ms-afk> (14:55:15) ms-afk: yes

[13:52] < ms-afk> (14:57:16) ms-afk: (go on)

[13:52] < ms-afk> Please don't ask for permission to talk about what your project is

[13:53] < ms-afk> If you don't get an answer posting the same thing to the list is appreciated

[13:53] < ms-afk> OK

[13:53] < Trun> ok, I go

[13:54] < Trun> well, on the one comment about debian/ubuntu in the checkin, I just wanted to mean that it might not be so easily found installed. Ubuntu and debian were just "examples"

[13:55] < ms-afk> Anyway, I'll be quiet and listen

[13:56] < Trun> well, I don't know pretty well how to start talking about it

[13:56] < Trun> the main class in the testing framework is the KamTestCase class

[13:57] < ms-afk> "well, I don't know pretty well how to start talking about it" - talk high level then

[13:57] < Trun> that I have changed several times and I still don't think it has an "stable api" or sth like that

[13:57] < ms-afk> then go into detail

[13:57] < Trun> ok

[13:58] < Trun> the idea is to have a class (KamTestCase) that "runs kamaelia" in each test, and stops it at the end

[13:58] < ms-afk> Not surprised that there isn't a stable API yet - I sort of expect that for the first 1/2 of GSOC you'll be doing exploratory work, and the second half doing stuff that actually tightens it up

[13:58] < ms-afk> k

[13:59] < Trun> the main problem is that if a test actually fails, it can go into an endless loop

[13:59] < Trun> (a test that is testing a Axon.Component.component )

[13:59] < ms-afk> Correct

[13:59] < Trun> while this can happen in other common tests, this is specially easy in kamaelia

[14:00] *** ms-afk nods

[14:00] < ms-afk> 2 suggestions then

[14:00] < ms-afk> 1) use my branch (private_MPS_Scratch) - that has a formal version of what you're trying to do here:

[14:00] < ms-afk> self._graph.__class__.schedulerClass.run = None

[14:00] < ms-afk> self._graph.__class__.schedulerClass()

[14:01] < ms-afk> 2) Use LikeFile (again perhaps the one in my /Sketches/MPS area)

[14:01] < ms-afk> to run systems

[14:01] < ms-afk> since then you're still left in control of the system

[14:01] < ms-afk> (That is actually a crutch, but it may help initially)

[14:02] < Trun> yeah, I used that first approach (right now it is the clearThread method at KamTestCase)

[14:02] < ms-afk> "self._graph.__class__.schedulerClass.run = None" is copied and pasted from your code

[14:03] < Trun> oh

[14:03] < ms-afk> In https://kamaelia.svn.sourceforge.net/svnroot/kamaelia/trunk/Sketches/MPS/pprocess/MultiPipeline.py

[14:03] < ms-afk> I do this:

[14:03] < ms-afk> background(zap=True).start()

[14:04] < ms-afk> Which starts a background scheduler & zaps the current scheduler state

[14:04] < Trun> oh, I see

[14:04] < ms-afk> (this is necesarry for reasons irrelevant here)

[14:04] < ms-afk> But "zap=True" relies on code in my branch

[14:04] < ms-afk> (it'll get merged soon enough)

[14:05] < ms-afk> Anyway, please continue

[14:06] < Trun> anyway, if you run a background scheduler, the problem of knowing if the system has finished or not still continues, right?

[14:06] < Trun> i mena

[14:06] < ms-afk> Depends if that is a test you're making

[14:06] < ms-afk> But it is then something you can test

[14:06] < Trun> I mean, right now I assert that "if in N steps the system hasn't finished, this is a bug"

[14:06] < ms-afk> and if it takes too long, you can exit the test

[14:07] < ms-afk> Since the scheduler runs in the background you can also query it's current run queues etc

[14:07] < Trun> but I don't know what should be "N", or how long something should take to know if it is a bug

[14:08] < ms-afk> This isn't necessarily threadsafe, BUT the snapshot could be useful

[14:08] < Trun> yeah, yeah, that code in LikeFile definitely is really useful for the testing framework (thanks!)

[14:08] < ms-afk> I think rather than "N" it should be time based

[14:09] < Trun> why time based?

[14:09] < Trun> I assumed that it would be based on "steps" to ensure that it will run the same expectations in any machine

[14:09] < ms-afk> You'll want to use likefile from here: /home/zathras/kamaelia/trunk/Sketches/MPS/Experiments/Likefile2/likefile

[14:10] < ms-afk> OK, time based or "steps" based

[14:10] < ms-afk> You also can't assume that "steps" will be the same on any machine either

[14:10] < ms-afk> especially if the component under test is actually process based - then the number of steps taken will vary with the number of CPUs

[14:11] < ms-afk> time based in terms of "if a test that should take nano-seconds or microsends takes 10 seconds, it's broken"

[14:12] < Trun> ok, for example there, it's normal to assume the user to provide that number (10 seconds), right?

[14:12] < ms-afk> they will know their code

[14:12] < Trun> ok

[14:12] < ms-afk> https://kamaelia.svn.sourceforge.net/svnroot/kamaelia/trunk/Sketches/MPS/GSOC08/MultiPipeTest.py

[14:12] < ms-afk> contains lots of "expect" style tests as well

[14:13] < ms-afk> which is a form of test I don't know if you're planning on automating & making re-runnable

[14:13] < Trun> I see

[14:14] < ms-afk> https://kamaelia.svn.sourceforge.net/svnroot/kamaelia/trunk/Sketches/MPS/Experiments/Likefile2/likefile/TestLikeFile.py - is again, effectively a manual test harness.

[14:14] < ms-afk> which is another type of problem to automate

[14:15] < Trun> I see

[14:16] *** Trun takes note :-)

[14:16] < ms-afk> You've also not (yet :-) ) related what you're doing to the tests in /trunk/Tests/Python/Axon /trunk/Tests/Python/Kamaelia , /trunk/Code/Python/Kamaelia/Test/

[14:17] < Trun> I read some of them on Friday (I didn't notice that the comments were later used in the doc until you said on saturday, though)

[14:18] < ms-afk> It's also probably worth noting that some tests have been broken by changes in the core of Axon itself

[14:18] < ms-afk> rather than the code being broken

[14:18] < ms-afk> for example:

[14:18] < ms-afk> https://kamaelia.svn.sourceforge.net/svnroot/kamaelia/trunk/Code/Python/Kamaelia/Test/Protocol/test_framing.py

[14:18] < ms-afk> now fails because it does things like:

[14:19] < ms-afk> for chunk in chunks:

[14:19] < ms-afk> system._deliver(chunk, "inbox")

[14:19] < ms-afk> and expects similarly

[14:19] < ms-afk> chunk = system._collect("outbox")

[14:19] < ms-afk> to exist

[14:19] < ms-afk> But it fails because it uses private methods to Axon which no longer actually exist

[14:20] < ms-afk> That particular test suite is perhaps the most useful here actually

[14:20] < ms-afk> partly because it shows how the scheduler can be run manually for a limited number of steps

[14:21] < ms-afk> which is similar to backgrounding

[14:21] < ms-afk> OK, I think I'm going to suggest you post to the mailing list

[14:21] < ms-afk> since you're not actually saying what you're doing

[14:22] < Trun> oh

[14:22] < Trun> you want an e-mail with what I'm doing exactly?

[14:25] < ms-afk> I can't mentor you if you don't communicate what you're doing

[14:25] < ms-afk> If you're not going to communicate what you're doing, no one will use your test suite

[14:25] < ms-afk> please start making it clear what you're up to

[14:25] < ms-afk> Start with a high level view and get more detailed from there

[14:25] < ms-afk> work from what you already have, but also where you want to go

[14:25] < Trun> I was reading what you had pointed out (I was recalling having used the _deliver method, but I see that it's in my misc folder, not in the test framework itself)

[14:25] < Trun> ok

[14:26] < Trun> currently the KamTestCase is the class that the user would use

[14:26] < Trun> you can (now) override the setUp and tearDown methods, write tests as if it was a unittest.TestCase instance

[14:27] < Trun> and there is a classmethod that provides a dynamically generated class that subclasses unittest.TestCase

[14:27] < Trun> this way, unittest test runners can run these tests

[14:27] < Trun> def suite():

[14:27] < Trun> return KamTestCase.TestSuite((

[14:27] < Trun> KamTestCase.makeSuite(FeedParserTestCase.getTestCase()),

[14:27] < Trun> KamTestCase.makeSuite(FeedParserFactoryTestCase.getTestCase()),

[14:27] < Trun> ))

[14:28] < Trun> (where KamTestCase.TestSuite = unittest.TestSuite and KamTestCase.makeSuite = unittest.TestSuite )

[14:28] < Trun> right now, the user that writes a test method, has two auxiliar attributes:

[14:28] < Trun> messageAdder and messageStorer, which are kamaelia components that send and receive messages to/from the component under test

[14:29] < Trun> these two components are created whenever the user calls a KamTestCase method called "initializeSystem"

[14:30] < Trun> in this method, the user says "I want to test this component", and KamTestCase will generate the messageAdder and messageStorer that will have the required inboxes/outboxes

[14:31] < Trun> ok, so what the user is expected to do is: setting up the environment through the setUp method (as usually)

[14:31] < Trun> call at some time the initializeSystem method

[14:31] < Trun> provide messages to be added to the component inboxes (you can also add "yields" between message and message)

[14:32] < Trun> and run the system (currently during a certain number of "steps")

[14:33] < Trun> if you expect it to finish, you call the self.assertStopping() method, if you expect it to not finish, you call the self.assertNotStopping() method

[14:33] < Trun> after that, you retrieve the messages stored in the messageStorer component, and check that the results are the expected ones

[14:34] < Trun> in order to avoid "problematic" components (those accessing disk/network/etc.) in tests, I have started a "KamMockObject"

[14:34] < Trun> you create it through an inherited KamTestCase method: self.createMock(whatever)

[14:35] < Trun> being whatever a component

[14:35] < Trun> the returned component will have the same public inboxes and outboxes that "whatever" has

[14:36] < Trun> and you can add messages that it will send (again, you can force a number of yields between message and message)

[14:36] < Trun> and check what messages it has stored

[14:36] < Trun> the user is not expected to create instances of this class directly, but through the self.createMock due to some "optimizations"

[14:37] < Trun> the idea is that the KamTestCase method that runs the system during "steps"/time can stop as soon as there is no more components running

[14:37] < Trun> and the KamMockObject objects shouldn't count

[14:38] < Trun> and I think that that's pretty all the infrastructure itself as it is right now

[14:40] < Trun> the rest of the code is just KamPlanet and testing KamPlanet to test the testing framework

[14:40] < Trun> for instance, until last commit, KamTestCase was actually a unittest.TestCase, so it could be run directly by the unittest runners

[14:41] < Trun> but I didn't find a way to allow the user to override the setUp and tearDown methods without having problems (KamTestCase needs to have its own initialize and finish method)

[14:43] < Trun> that's why there is now an "ugly" _AuxiliarTestCase class in the KamTestCase module, and a couple of new methods that dynamically generate the missing classes

[14:43] < Trun> (i.e. a class per KamTestCase, and all the assert* and fail* methods of unittest.TestCase are also added to KamTestCase)

[14:44] < Trun> do you want me to continue to a lower-level detail, or do you prefer information about KamPlanet?

[14:44] *** ms-afk notices that the lot is going in the channel after I specifically said to please post it to the list

[14:45] < ms-afk> (15:26:50) ms-afk: OK, I think I'm going to suggest you post to the mailing list

[14:45] < Trun> oh, sorry

[14:45] < ms-afk> I'd actually moved on to do other work, ignoring this window

[14:46] < Trun> after that you talked about communicating what I was doing I thought you meant "now", sorry for that :-S

[14:46] < ms-afk> After all, the time you can "guarantee" me being there is from 2pm til 3pm

[14:46] < ms-afk> not "all afternoon"

[14:46] < ms-afk> email *IS* communicating

[14:46] < ms-afk> sigh

[14:46] < Trun> yeah, sorry for that :-( I'll write the e-mail

[14:49] *** Uraeus_ has joined #kamaelia

[14:51] < Davbo> don't worry about it Trun just refactor it into an email

[14:53] < ms-afk> ++ Davbo

[14:53] *** j_baker has joined #kamaelia

[15:04] *** Davbo has joined #kamaelia

[15:08] < Davbo> let me know if there's anything else you'd like me to review ms-afk

[15:08] < Davbo> private_MPS_Tools2AppsConsolidation looks like it isn't claimed

[15:08] < ms-afk> That'd be a good one :)

[15:08] < Davbo> I'll do that tomorrow if you'd like

[15:08] < ms-afk> Cool

[15:08] < Davbo> i'll add to wiki

[15:09] < ms-afk> Try not to allow these reviews to get in the way of your project work (goes for everyone :-) ), but they are very much appreciated :)

[15:09] < ms-afk> And in many respects probably better experience ;)

[15:10] < Davbo> well closer to release i might not be able to help out as much as now

[15:10] < Davbo> so i'd like to get some done now

[15:10] < Davbo> :-)

[15:11] < Lawouach_> I got spam today, one was title: "Update your penis"

[15:12] < Lawouach_> They know my geek side.

[15:12] < Davbo> hah Lawouach_

[15:44] < Davbo> ms-afk: any idea's why i'm getting this error?

[15:44] < Davbo> TypeError: __init__() got an unexpected keyword argument 'zap'

[15:45] < Davbo> I can paste the rest for you if you like

[15:47] < Davbo> http://davbo.pastebin.com/m5c1dcda0

[16:02] < ms-afk> Oh

[16:02] < ms-afk> that

[16:03] < ms-afk> That's because you're not using the branch private_MPS_Scratch which contains that extra keyword

[16:03] < ms-afk> (specifically not using *axon* from that branch)

[16:06] < Davbo> Weird ms-afk, i've installed it and still get it

[16:07] < Davbo> (i had already installed it before asking, assuming that was the problem)

[16:14] *** j_baker has joined #kamaelia

[16:16] < ms-afk> really ?

[16:16] < ms-afk> can you check in your current code version

[16:16] < ms-afk> ?

[16:17] < Davbo> Yeah Sketches/DK/MagnaDoodle.py is popping the error

[16:18] < ms-afk> works here

[16:18] *** ms-afk did this:

[16:18] < ms-afk> ~>

[16:18] < ms-afk> ~> cd ~/kamaelia/branches/private_MPS_Scratch/Axon/

[16:18] < ms-afk> ~/kamaelia/branches/private_MPS_Scratch/Axon> sudo python setup.py --quiet install --force

[16:20] < Davbo> works

[16:20] < Davbo> thanks ms-afk

[16:20] < Davbo> :-)

[16:21] < Davbo> I was just doing regular install didn't know I had to force

[16:21] < j_baker> Oooooohhhh.... OS X 10.6 will have ZFS.

[16:21] < Davbo> I guess that means it overwrites it all?

[16:23] < Davbo> Heh j_baker :-) Yesterday I was "ooooo! the iPhone is going to be affordable in this country"

[16:23] < j_baker> Here too. $199 isn't bad.

[16:24] < Davbo> It was about $700 over here before + contract ended up being near $1800 ;/

[16:24] < Davbo> too much for a phone

[16:24] < j_baker> They charge you MORE for a contract?

[16:25] < Davbo> well that's for the entire duration of the contract

[16:25] < j_baker> Ah, I see. That's still expensive though.

[16:25] < Davbo> O2 are doing a Pay as you Go deal for the iPhone this time round

[16:25] < Davbo> not announced prices yet

[16:25] < j_baker> Really? Now THAT is cool.

[16:28] < Davbo> I guess AT&T don't do that?

[16:28] < j_baker> I don't believe so. I haven't checked though.

[16:30] < j_baker> I'm still happy with my Ming though

[16:30] < j_baker> http://www.motorola.com.hk/eng/motomobile/a/a1200/

[16:31] < Davbo> ah, thats a cool phone

[16:31] < j_baker> Yeah. Only problem is that it's GPRS only. No Edge or 3G. But I hardly ever use the internet on my phone anyway

[16:46] *** j_baker just spent 20 mins trying to figure out why the internet wasn't working on my desktop only to figure out the ethernet cord was unplugged.

[16:57] < Davbo> Imagine how long it would have took if you weren't so tech savy though :)

[16:57] < j_baker> lol... good point.

[16:57] < Davbo> you'd have took it to the Geek Squad ;-)

[16:58] < Davbo> or whatever they're called

[16:58] < j_baker> Thank god I know how to fix computers. I'm not particularly fond of dealing with technicians/salespeople who over charge you.

[17:04] *** Uraeus_ is now known as Uraeus

[17:45] *** vmlemon_ has joined #kamaelia

[17:46] < vmlemon_> Hi

[17:53] < ms-afk> hi

[18:01] < Lawouach> ms-afk: http://rafb.net/p/dLF2e392.html

[18:01] < Lawouach> the patch to remove trailing print statements in HTTP components

[18:01] < Lawouach> I can just commit it if you will

[18:01] *** vmlemon_ wonders if any tools for mining source code locally exist...

[18:02] < Lawouach> what do you mean by mining? what would you want to see?

[18:02] < ms-afk> Lawouach: (looking)

[18:03] < ms-afk> Ah, I see :)

[18:03] < ms-afk> If theyre already on a checkout of /trunk feel free to commit :)

[18:03] < vmlemon_> Essentially a keyword-based search engine of some sort that gets pointed to a set of directories and archive files

[18:04] < Lawouach> ms-afk: I've remove thf irst one because I had introduced it I think

[18:04] < Lawouach> The second one was already there

[18:04] < Lawouach> so I just commented out

[18:04] < vmlemon_> I'm sure something for general text or even HTML handling would be useful, though

[18:04] < ms-afk> I must admit, my preference is to comment out debugging info, rather than delete it

[18:04] < ms-afk> Until I'm sure it all works as intended for absolute certain that is :)

[18:05] < Lawouach> I can revert the first one locally and just comment it

[18:05] < ms-afk> Up to you - they're trivial - just talking personal preference

[18:05] < ms-afk> it's only 2 lines - hardly worth discussion :)

[18:06] < Lawouach> indeed

[18:06] < Lawouach> I'll comment both and commit then

[18:06] < ms-afk> ta :)

[18:06] *** j_baker has joined #kamaelia

[18:07] < vmlemon_> http://gonzui.sourceforge.net/ sounds like it might be useful for what I'm after

[18:08] < ms-afk> woo, it supports BF

[18:09] < j_baker> Hmmm... apparently the Minimal handler is REALLY vulnerable to DOS attacks.

[18:10] < ms-afk> ?

[18:10] *** bcarlyon|laptop has joined #kamaelia

[18:10] < vmlemon_> Except it doesn't appear to support searching zip/tar archives

[18:10] < ms-afk> j_baker: why do you say that?

[18:10] < j_baker> Well, I tried doing a little stress testing on your new server code.

[18:10] *** bcarlyon|laptop has parted #kamaelia

[18:11] < j_baker> And it gave me a too many files open error about 34% of the way through.

[18:11] < ms-afk> So that's a bug. It can be fixed :)

[18:12] < j_baker> And fairly easily I'm assuming. I would imagine that all we'd have to do is change the intelligent file reader to a max speed file reader.

[18:12] < j_baker> Either that or just open the file directly.

[18:15] < ms-afk> I don't actually know - I suspect that the problem is likely to be caused by files not actually being closed after being read to be honest

[18:15] < j_baker> That could also be the case. I don't recall seeing that code anywhere.

[18:16] < j_baker> Hey btw, were you aware that Kamaelia.Internet.TimeOutCSA doesn't have any documentation, ms-afk?

[18:16] < ms-afk> yeah, I suck

[18:16] < ms-afk> :)

[18:17] < j_baker> :)

[18:17] < ms-afk> It was written because I *really* needed it for K.Grey, but you're right it needs doc'ing

[18:17] < ms-afk> noting that as a blocker for merge is fine fwiw

[18:19] < ms-afk> j_baker: IFR doesn't actually close the file

[18:19] < ms-afk> That'll almost certainly be a problem

[18:19] < ms-afk> (It may not be the only one)

[18:19] < j_baker> Yeah. You're correct.

[18:24] *** ms-afk goes afk. Hopefully back later

[18:24] < ms-afk> (many thanks for taking this level of care BTW - it's very much appreciated!)

[18:25] < ms-afk> biab

[18:25] *** ms-afk has parted #kamaelia

[18:30] *** vmlemon has joined #kamaelia

[18:36] < vmlemon_> kamaeliabot: dance

[18:36] Reply: does the macarena

[19:43] *** j_baker_ has joined #kamaelia

[20:02] *** j_baker_ is now known as j_baker

[20:08] *** vmlemon has joined #kamaelia

[20:17] *** vmlemon_ has joined #kamaelia

[20:17] *** bcarlyon|laptop has joined #kamaelia

[20:24] *** vmlemon_ has joined #kamaelia

[21:37] *** Uraeus has joined #kamaelia

[21:48] < vmlemon_> Hah, http://cache.gizmodo.com/assets/resources/2008/04/psystarofficial.png

[22:27] *** vmlemon__ has joined #kamaelia

[22:27] *** vmlemon__ is now known as vmlemon_

[23:06] *** j_baker has joined #kamaelia