1005004 - Make default address localhost:2828

Status: RESOLVED INCOMPLETE

(Remote Protocol :: Marionette, defect, P4)

Description

[Zac C (:zac)]

It would be great to just assume a default --address of localhost:2828 so that it does not need to be passed in at the command line.

However if the user does pass it in then use the one they have requested.

Comment 1

[Florin Strugariu [:Bebe]]

+1  on this having a smaller simple command  line to run the tests would be very helpful

Comment 2

[Dave Hunt [:davehunt] [he/him] ⌚BST]

I suspect that adding --address indicates the intention to connect to a remote instance, whereas excluding it requires us to launch a Gecko instance or emulator. It's certainly worth experimenting with this, or alternatively allowing command line options to be stored in a config file so they don't need to be retyped in full each time. Would this change be made in gaiatest or upstream in the Marionette test runner? I think both are valid options.

Comment 3

[Zac C (:zac)]

I put it into this component because it's easier for us to get a result or trial it without having to do a whole marionette release but I'm not actually sure of the technical solution yet, it's just an idea in my head.
If it's not achievable here then we can move it to Marionette.

Comment 4

[Zac C (:zac)]

I see we can just put 
        self.set_defaults(address='localhost:2828')
 into mixins/gaia.py

however I haven't checked this with desktopb2g variations :)

If it works I think it would be better put into MarionetteTestRunner.

Comment 5

[Zac C (:zac)]

Malini, is this something we could/should do on the TestRunner side?

Comment 6

[Malini Das [:mdas] - Away, not checking bugmail]

(In reply to Zac C (:zac) from comment #5)
> Malini, is this something we could/should do on the TestRunner side?

Yup, we can do this in the runner. davehunt is right; we use the combination of --address and other options to determine the test environment, but we can change the client to follow this behaviour:

* if no options are used, assume localhost:2828
* if no --address is used, and any option other than --emulator is selected, assume localhost:2828
* if --address is used, and any option other than --emulator is selected, use the given address
* if --emulator is used, and no --address is passed in, then we retain our current behaviour: launch emulator and forward port for you
* if --emulator and --address is used, then we retain our current behaviour: assume there is an emulator running and use the given address to communicate with marionette


I think this behaviour will cover all cases, and retains expected behaviour in tbpl/other people's scripts

Comment 7

[Zac C (:zac)]

Thanks Malini, I'll change the component to Marionette.

Comment 8

[David Burns :automatedtester]

http://dxr.mozilla.org/mozilla-central/source/testing/marionette/client/marionette/runner/base.py needs to be updated to handle details mentioned in comment 6.

Setting mdas as mentor but if she is still on PTO when someone picks it up will mentor

Comment 9

[Rohith PR]

I would like to work on this bug. I am unable to find the code though. Please help.

Comment 10

[Jonathan Griffin (:jgriffin)]

This code lives in an hg repo, http://hg.mozilla.org/mozilla-central.  You'll have to clone this, modify the file from comment #8, then submit a patch according to https://developer.mozilla.org/en-US/docs/Mercurial_FAQ#How_can_I_generate_a_patch_for_somebody_else_to_check-in_for_me.3F.

Comment 11

[Rohith PR]

I still can't find it. Searching on that page (http://hg.mozilla.org/mozilla-central) returns a blank.

Comment 12

[Dave Hunt [:davehunt] [he/him] ⌚BST]

Perhaps there was an issue with the service when you tried. Here's a guide to getting hold of the source code: https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/Source_Code/Mercurial

Comment 13

[Rohith PR]

I give up! Everything is going over my head. Maybe in a few years!

Comment 14

[Dave Hunt [:davehunt] [he/him] ⌚BST]

Sorry that you're having difficulty getting started Rohith. I'd be happy to help you with getting the repository cloned. The code is stored in a Mercurial repository, which means you will need to have a Mercurial client installed on your machine. Mozilla provide documentation on Mercurial [1], but it can be a little overwhelming.

What operating system are you using, and do you have Python installed? Knowing these, I should be able to help guide you through installing Mercurial and then cloning the repository. If you are more comfortable with Git repositories then we do have a Git mirror of our mozilla-central repository. Let me know if you'd like more details on this.

[1] https://developer.mozilla.org/en-US/docs/Installing_Mercurial

Comment 15

[Rohith PR]

I am using Windows 8 and have Python 3.3.2 installed!

Comment 16

[Dave Hunt [:davehunt] [he/him] ⌚BST]

Okay, I've never used Mercurial on Windows 8, but let's see if this works! :) Download and run the MozillaBuild package available at https://ftp.mozilla.org/pub/mozilla.org/mozilla/libraries/win32/MozillaBuildSetup-Latest.exe

You should then have Mercurial available. Open a command prompt and type |hg --version| to see the version of Mercurial installed. It should be 2.9.1. If that command doesn't work you may need to change directory to where the MozillaBuild unpacked itself. Look for a 'MozillaBuid' folder or similar.

Then, you should be able to clone the repository using the command |hg clone http://hg.mozilla.org/mozilla-central <path>| replacing <path> with the location you want to store the repository. Once this has completed (it may take a while, depending on your connection) you should find the files you need to modify in <path>\testing\marionette\client\marionette\runner.

Let us know how you get on with this. It may also be worthwhile installing Python 2.7 as many of our Python frameworks are running on boxes with Python 2.7, so we need to ensure we maintain compatibility.

Comment 17

[David Burns :automatedtester]

Assigning to Rohith as you are getting setup

Comment 18

[Rohith PR]

This is what has happened:

>hg clone http://hg.mozilla.org/mozilla-central "C:\..."
requesting all changes
adding changesets
adding manifestos
adding file changes
transaction abort!
rollback completed
abort: An existing connection was forcibly closed by the remote host

I've used about 700MB of data and not a single file in my comp to show for it!

Comment 19

[Sunil Kumar]

Hi. I'm new for contributing to open source projects. Please help me in getting started. I want to work on this bug.

I use Archlinux distribution of Linux, I have python 3.4.0 installed on my system and I'm more comfortable using Git than Mercurial. Can you give me the path to Git mirror of mozilla-central repository so that i can work on this bug.

Thanks,
Sunil

Comment 20

[Sunil Kumar]

Also, how do i check the working of the application ?

Comment 21

[Sunil Kumar]

Okay. I think I have found the project on github.
https://github.com/mozilla/gecko-dev/blob/master/testing/marionette/client/marionette/runner/base.py
this is the one right ?

Now tell him how do i check the working of the application. I mean, i want to know the detailed steps to get to the bug.

Comment 22

[Malini Das [:mdas] - Away, not checking bugmail]

(In reply to Sunil Kumar from comment #19)
> Hi. I'm new for contributing to open source projects. Please help me in
> getting started. I want to work on this bug.
> 
> I use Archlinux distribution of Linux, I have python 3.4.0 installed on my
> system and I'm more comfortable using Git than Mercurial. Can you give me
> the path to Git mirror of mozilla-central repository so that i can work on
> this bug.
> 
> Thanks,
> Sunil

Hi Sunil, thanks for your interest!

Rohith currently is assigned this bug, but please feel free to look through unassigned good first bugs here: http://www.joshmatthews.net/bugsahoy/?automation=1&py=1&unowned=1

Comment 23

[Malini Das [:mdas] - Away, not checking bugmail]

(In reply to Rohith PR from comment #18)
> This is what has happened:
> 
> >hg clone http://hg.mozilla.org/mozilla-central "C:\..."
> requesting all changes
> adding changesets
> adding manifestos
> adding file changes
> transaction abort!
> rollback completed
> abort: An existing connection was forcibly closed by the remote host
> 
> I've used about 700MB of data and not a single file in my comp to show for
> it!

Unfortunately, I don't have any experience running mercurial in windows, but according to stackoverflow, it might be a mercurial configuration error. Can you check if your %USERPROFILE%\Mercurial.ini file and see if you have a section titled "[http_proxy]"? If so, can you delete that section and retry? Here's an example of what I mean:

if you have a mercurial.ini filie that looks like

[ui]
username = somename

[http_proxy]
host = http://localhost:8000/
user = myname

[diff]
git=true

Then you should remove the [http_proxy] section so it looks like this:

[ui]
username = somename

[diff]
git=true

Comment 24

[Rohith PR]

I quit. :( Please Assign it to Sunil Kumar.

Comment 25

[Malini Das [:mdas] - Away, not checking bugmail]

(In reply to Rohith PR from comment #24)
> I quit. :( Please Assign it to Sunil Kumar.

I'm sorry to hear that, Rohith! :(

Sunil, are you still interested in working on this?

Comment 26

[Malini Das [:mdas] - Away, not checking bugmail]

This isn't really a good first bug. You can't test the emulator related changes locally as a contributor. You have to push to try to test, which is very frustrating.

Comment 27

[David Burns :automatedtester]

A lot of the emulator code is about to go and the kwargs already handle his.