| *** tpb has joined #timvideos | 00:00 | |
| *** CarlFK has joined #timvideos | 00:10 | |
| *** CarlFK has quit IRC | 00:21 | |
| *** CarlFK has joined #timvideos | 00:30 | |
| *** iiie has quit IRC | 00:31 | |
| *** jahanzeb has quit IRC | 01:02 | |
| *** Jahanzeb has joined #timvideos | 01:03 | |
| *** mithro has quit IRC | 01:17 | |
| mithro-work | CarlFK, get anywhere? | 01:43 |
|---|---|---|
| CarlFK | no - I somehow went backwards. | 01:44 |
| CarlFK | hdmi ports behave as expected | 01:44 |
| CarlFK | power up, get atlys test pattern on monitor | 01:44 |
| CarlFK | upload firmware, get grid? test image on monitor | 01:45 |
| CarlFK | hook up rpi, hit the D? button, see the pi's video on the monitor | 01:45 |
| CarlFK | but I don't get an image on the /dev/video device | 01:45 |
| CarlFK | I do get "1024x768" and maybe 15fps | 01:46 |
| CarlFK | and now my dhcp server isn't working | 01:46 |
| *** Jahanzeb has quit IRC | 01:56 | |
| CarlFK | im running up and down stairts - server in the basment, just checked a tablet that is up stairs, dhcp worked.. but the dvswitch box .. even the pxe isn't making anything show up in the dhcp log | 01:57 |
| CarlFK | crap - i seem to have lost the ethernet port on one of my laptops | 02:14 |
| *** hyades has joined #timvideos | 04:16 | |
| *** mithro has joined #timvideos | 04:44 | |
| *** mithro has quit IRC | 05:44 | |
| *** mithro-work has quit IRC | 05:44 | |
| *** mithro-work has joined #timvideos | 05:57 | |
| *** mithro has joined #timvideos | 05:58 | |
| *** Jahanzeb has joined #timvideos | 10:25 | |
| Jahanzeb | Hi CarlFK | 10:28 |
| Jahanzeb | which firmware you are testing with | 10:28 |
| Jahanzeb | try https://github.com/jahanzeb/HDMI2USB_prebuild | 10:29 |
| tpb | Title: jahanzeb/HDMI2USB_prebuild ยท GitHub (at github.com) | 10:29 |
| Jahanzeb | and if its still not working can you also try one from archive so that i can see where might be problem | 10:30 |
| *** Jahanzeb has quit IRC | 10:38 | |
| *** bananadine has joined #timvideos | 10:54 | |
| mithro | bananadine, I don't think your GstNet "fix" is right | 10:56 |
| *** Jahanzeb has joined #timvideos | 10:56 | |
| bananadine | mithro, why | 10:56 |
| bananadine | importing works everywhere now | 11:00 |
| mithro | bananadine, if you remove those lines does it stop working? | 11:00 |
| bananadine | yes | 11:01 |
| bananadine | mithro that's why i put those lines, so importing GstNet works everywhere | 11:13 |
| mithro | bananadine, so you haven't found the real problem | 11:14 |
| mithro | what you have found is a random method for stopping the problem occuring | 11:14 |
| mithro | you still don't know *why* GstNet it not importable without those lines? | 11:14 |
| mithro | bananadine, understand the difference? | 11:16 |
| bananadine | that's the boot file | 11:16 |
| mithro | the boot.py file is nothing special | 11:16 |
| mithro | it's just a python file like anything else | 11:16 |
| bananadine | well i don't see it | 11:18 |
| bananadine | i can't look into it deeper that that | 11:18 |
| mithro | yes you can | 11:19 |
| mithro | why can boot.py import it | 11:19 |
| mithro | but not the other file? | 11:20 |
| bananadine | i assert it in boot to sys.modules | 11:21 |
| bananadine | and boot is one of the first modules that runs | 11:21 |
| bananadine | well it's not the assert... | 11:42 |
| bananadine | but still boot is one of the first modules | 11:42 |
| Sewar | What's the border between unittest and Selenium? | 12:09 |
| *** hyades has quit IRC | 12:17 | |
| mithro | Sewar, what do you mean? | 12:39 |
| mithro | A Selenium test is *not* a unittest | 12:39 |
| mithro | even if it uses the unittest.py module | 12:39 |
| mithro | unittests have no external dependencies | 12:39 |
| Sewar | When do I use unittest or Selenium? | 12:39 |
| mithro | bananadine, so that means something happens between boot.py and when gstnet is imported later that causes it to fail | 12:40 |
| Sewar | and not overlap, testing same thing in both ... | 12:40 |
| Sewar | "Selenium is especially useful for testing javascript. unittest is for python" from #django | 12:41 |
| *** hyades has joined #timvideos | 12:49 | |
| mithro-work | Selenium is what I'd call an "integration test" | 12:49 |
| mithro-work | so it's quite confusing | 12:50 |
| mithro-work | so there is the unittest.py module | 12:50 |
| mithro-work | and the concept of a unittest | 12:50 |
| mithro-work | and these two things should not be confused | 12:50 |
| mithro-work | because unittest.py is used for all types of tests, unit tests, integration tests, etc | 12:51 |
| hyades | mithro-work, even i'm listening :) | 12:51 |
| mithro-work | Unit tests are all about testing very small "units" of code | 12:52 |
| mithro-work | they should be fast | 12:52 |
| mithro-work | they should not have external dependencies | 12:52 |
| hyades | mithro-work, is such a thing a unittest? | 12:52 |
| hyades | http://pastebin.com/7JAmzhRp | 12:52 |
| tpb | Title: [Python] from server import BaseServer import pytest class TestClass(object): - Pastebin.com (at pastebin.com) | 12:52 |
| mithro-work | In Python "real unittests" are often best done via the doctest module | 12:53 |
| mithro-work | hyades, yes it looks that way - I'm not sure it's really that useful | 12:53 |
| mithro-work | IE ports values should be things like | 12:54 |
| mithro-work | ports = [-1, 1e6, None, '200', {}] | 12:54 |
| mithro-work | testing a couple of valid inputs makes sense though | 12:54 |
| mithro-work | Sewar, poke? | 12:55 |
| hyades | so you mean like if i put a port say 80, which will be in use almost always, it should throw an exception | 12:55 |
| mithro-work | hyades, nope - 80 is a perfectly valid port number | 12:55 |
| Sewar | that makes sense | 12:55 |
| mithro-work | hyades, what happens if you give it -1? | 12:55 |
| mithro-work | Sewar, unittests generally look at a single function | 12:55 |
| mithro-work | or single class | 12:55 |
| hyades | it passes this port onto gst-switch-srv | 12:55 |
| mithro-work | hyades, does -1 make sense? what about -3? | 12:56 |
| hyades | idk | 12:56 |
| hyades | im getting the point | 12:56 |
| mithro-work | hyades, "I don't know" is something you don't want to be in the position of | 12:56 |
| mithro-work | :) | 12:56 |
| mithro-work | tests help you make sure that you *do* know what will happen | 12:57 |
| mithro-work | unittests have a major problem | 12:57 |
| mithro-work | if doesn't matter if all the functions work in isolation | 12:57 |
| mithro-work | if they fail when put together | 12:57 |
| mithro-work | your still screwed | 12:57 |
| hyades | so should the code have exception handlers which do not allow such ports to be entered | 12:57 |
| hyades | or something else? | 12:58 |
| mithro-work | hyades, possibly | 12:58 |
| hyades | ok | 12:58 |
| Sewar | ok, so Selenium should be used to run multiple operations at one session | 12:58 |
| mithro-work | so integration tests are about looking at system as a whole | 12:58 |
| mithro-work | making sure that a person can login | 12:58 |
| mithro-work | would be an integration test | 12:58 |
| mithro-work | the closer your integration test is to how a real user would operate the better | 12:59 |
| mithro-work | hence why selenium makes a lot of sense | 12:59 |
| mithro-work | because it doesn't matter if your system works if you send user/password as arguments | 12:59 |
| mithro-work | if the fields don't exist in your webpage your not going to be able to send them | 12:59 |
| mithro-work | Sewar, make more sense? | 13:00 |
| Sewar | indeed | 13:00 |
| mithro-work | The problem with integration tests is that there is pretty much an infinite number of ways to use any normal app | 13:00 |
| mithro-work | so you break it down into "work flows" | 13:00 |
| mithro-work | IE "login into the system" | 13:00 |
| mithro-work | or "create a new entry in the system" | 13:01 |
| mithro-work | to do the second you probably need to be logged in - but testing that you *can* login shouldn't really be part of that test | 13:01 |
| Sewar | yeah, I already do the same in unittests, for setUp() objects | 13:03 |
| Sewar | setUp()'s objects (eg. EDID) | 13:03 |
| mithro-work | so on the note of testing | 13:03 |
| mithro-work | if your tests are flaky they they shouldn't exist - flaky tests are *worse* then no tests | 13:04 |
| mithro-work | flaky tests are ones which pass most of the time but sometimes randomly fail | 13:04 |
| Sewar | lol | 13:04 |
| Sewar | that makes sense | 13:05 |
| mithro-work | because | 13:05 |
| mithro-work | a) they give you false confidence | 13:05 |
| mithro-work | b) you never know if your code is broken or the test is broken | 13:05 |
| mithro-work | so there is a trade off between creating tests and making sure they are not flaky | 13:06 |
| mithro-work | depending on external services is one way to make your tests flaky | 13:06 |
| mithro-work | for example, if your login flow needs to talk to facebook | 13:06 |
| mithro-work | it's going to be flaky because occasionally facebook will be done | 13:06 |
| mithro-work | s/done/down/ | 13:06 |
| Sewar | yeah | 13:07 |
| mithro-work | make sure your test suites don't require any internet access | 13:07 |
| mithro-work | because needing network access is one way to make your tests flaky | 13:07 |
| Sewar | noted | 13:09 |
| mithro-work | and it is really easy to accidently do | 13:09 |
| mithro-work | I know from experience :) | 13:09 |
| mithro-work | hyades, does that help at all? | 13:10 |
| hyades | yeah | 13:10 |
| mithro-work | It's really hard to test error cases in integration tests | 13:13 |
| mithro-work | but really easy in unittests | 13:13 |
| Sewar | unittests to make sure every part works as expected, and integration tests to make sure things are still coupled together | 13:14 |
| * Sewar goes to sleep | 13:15 | |
| bananadine | mithro-work, GstNet.NetTimeProvider.new(clock, None, port) doesn't work, i need to put an address in place of None | 13:17 |
| bananadine | what should i put? | 13:17 |
| mithro-work | bananadine, I have no idea | 13:17 |
| mithro-work | <mithro> bananadine, so that means something happens between boot.py and when gstnet is imported later that causes it to fail | 13:18 |
| bananadine | oh i needed to go out for something so i didn't notice you wrote | 13:18 |
| bananadine | indeed | 13:19 |
| bananadine | ah | 13:21 |
| bananadine | flumotion-manager, flumotion-worker and flumotion-admin | 13:22 |
| bananadine | all call boot | 13:22 |
| *** hyades has quit IRC | 13:26 | |
| mithro-work | bananadine, that is not the point | 13:26 |
| *** hyades has joined #timvideos | 13:27 | |
| bananadine | mithro-work, i don't know what could it possibly be | 13:28 |
| mithro-work | bananadine, neither do I, but you have a range to work with | 13:28 |
| mithro-work | keep dividing that in half | 13:28 |
| mithro-work | until you workout where the issue is | 13:28 |
| bananadine | i thought the issue was that the flumotion manager and worker use boot.py and there was no GstNet in boot.py and that caused the problem | 13:31 |
| hyades | mithro-work, i will be needing custom exception handlers | 13:47 |
| mithro-work | bananadine, no | 13:47 |
| mithro-work | hyades, what do you mean? | 13:47 |
| hyades | suppose if i put port -2, then some exception should be raised | 13:48 |
| hyades | something like ValueError | 13:48 |
| mithro-work | hyades, correct | 13:49 |
| bananadine | mithro-work, i can't do that now | 14:05 |
| bananadine | i want to fix this | 14:05 |
| bananadine | WARN [ 9285] "producer-video" feedcomponent Jul 15 15:12:57 Setup failed: failure <type 'exceptions.TypeError'> at flumotion/component/producers/videotest/videotest.py:53: get_pipeline_string(): 'tuple' object does not support item assignment (flumotion/component/component.py:586) | 14:05 |
| bananadine | here's the difference between the old gst and the new | 14:14 |
| bananadine | http://pastebin.com/EdgNMb0a | 14:14 |
| tpb | Title: gst In [3]: gst.structure_from_string('aha') Out[3]: (at pastebin.com) | 14:14 |
| bananadine | https://github.com/bananadine/flumotion/blob/master/flumotion/component/producers/videotest/videotest.py#L53 | 14:16 |
| tpb | <http://ln-s.net/-Uh7> (at github.com) | 14:16 |
| bananadine | and the error points here | 14:16 |
| bananadine | mithro-work? | 14:31 |
| bananadine | mithro-work ?* | 14:32 |
| hyades | mithro-work, testing everything will be huge! woah | 14:37 |
| *** mithro has quit IRC | 14:39 | |
| *** Jahanzeb1 has joined #timvideos | 15:12 | |
| *** Jahanzeb has quit IRC | 15:23 | |
| *** Jahanzeb1 has left #timvideos | 15:26 | |
| *** Jahanzeb1 has joined #timvideos | 15:26 | |
| *** Jahanzeb1 has quit IRC | 15:26 | |
| *** hyades has quit IRC | 15:27 | |
| *** Jahanzeb has joined #timvideos | 15:27 | |
| *** hyades has joined #timvideos | 15:27 | |
| *** bananadine has quit IRC | 16:11 | |
| *** Jahanzeb has left #timvideos | 16:58 | |
| *** Sewar has quit IRC | 17:07 | |
| *** micolous_ has quit IRC | 17:07 | |
| *** hyades has quit IRC | 17:09 | |
| *** hyades has joined #timvideos | 17:09 | |
| *** Sewar has joined #timvideos | 17:09 | |
| *** micolous_ has joined #timvideos | 17:09 | |
| *** hyades has quit IRC | 17:43 | |
| *** hyades has joined #timvideos | 17:44 | |
| *** iiie has joined #timvideos | 21:10 | |
| *** hyades has quit IRC | 21:21 | |
| *** Sewar has quit IRC | 22:42 | |
| *** micolous_ has quit IRC | 22:42 | |
| *** mithro-work has quit IRC | 22:42 | |
| *** parx has quit IRC | 22:42 | |
| *** micolous_ has joined #timvideos | 22:44 | |
| *** Sewar has joined #timvideos | 22:44 | |
| *** mithro-work has joined #timvideos | 22:44 | |
| *** parx has joined #timvideos | 22:44 | |
| CarlFK | magic. atlys is working again | 22:52 |
| CarlFK | so is the laptop eathernet port that went out to lunch | 22:53 |
| iiie | welcome back ethernet port! how was lunch? | 22:57 |
| CarlFK | lol | 22:58 |
| CarlFK | you ever do pxe boot? | 22:58 |
| CarlFK | like you know how the nic eprom has a tcp/ip stack, dhcp client... | 22:58 |
| CarlFK | and I wasn't trying to flash it | 22:58 |
| iiie | not successfully | 22:58 |
| CarlFK | yeah.. so it has all that. most do. | 22:59 |
| iiie | ah, so setting is off and it can't both use and boot? | 22:59 |
| CarlFK | it was still on | 22:59 |
| CarlFK | anyway... pxe didn't work, boot into ubunutu, ifconfig showed a port, link light was on, but dhcp requests didn't show up in the dhcp server log | 23:01 |
| CarlFK | power cycled.. different cards, power cycles the switch.. | 23:02 |
| CarlFK | turned it off, went to lunch, turned it back on, now it works. | 23:02 |
| iiie | it was tired, and like you said, it needed a lunch | 23:15 |
| CarlFK | apparently | 23:19 |
Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!