Thursday, September 1 2005
From LDTP
Sep 01 16:29:46 nags hey guys meeting time
Sep 01 16:29:52 nags krish_: ping
Sep 01 16:29:55 nags mnagashree: ping
Sep 01 16:29:59 nags shaheed: ping
Sep 01 16:30:12 * rkrish_ (~rkrishnan@202.144.95.244) has joined #ldtp
Sep 01 16:30:12 shaheed nags: i am in
Sep 01 16:30:21 nags shaheed: oka :)
Sep 01 16:30:28 nags rkrish_: ping
Sep 01 16:30:47 * shaheed gives channel operator status to rkrish_
Sep 01 16:30:49 rkrish_ nags: I am in
Sep 01 16:30:51 rkrish_ nags: pong
Sep 01 16:31:03 nags rkrish_: okay :)
Sep 01 16:31:21 mnagashree nags, as of now, I have no updates regarding file-roller,
Sep 01 16:31:29 nags let us start :)
Sep 01 16:31:32 nags mnagashree: oops
Sep 01 16:31:33 mnagashree will find some time and finish that :)
Sep 01 16:31:40 nags mnagashree: sure, thanks
Sep 01 16:32:08 nags Today our agenda
Sep 01 16:32:11 nags - Progress of prem's new LDTP search algorithm
Sep 01 16:32:11 nags - New component functions yet to be written in pyldtp
Sep 01 16:32:11 nags - Plans for integrating libstatgrab to get the CPU / Memory utilization
Sep 01 16:32:11 nags - Shaheed's LDTP editor implementation details
Sep 01 16:32:45 rkrish_ nags: also we can ask kaveri to update her plan
Sep 01 16:33:13 nags I'm interacting closely with prem for CPU / Memory utilization using libstatgrab... we are facing few issues
Sep 01 16:33:25 nags I mean bug in libstatgrab
Sep 01 16:33:36 nags filed bug in their bugzilla by prem
Sep 01 16:33:41 rkrish_ nags: what is libstatgrab ?
Sep 01 16:33:58 prem For gathering system performance related statistics like cpu utilisation,memory utilisation etc., , libstatgrab is really useful. The good thing is that it is compatabile with many platforms and also python binding is provided which eases our integration efforts .
Sep 01 16:34:17 nags me shaheed, rkrish_ discussed a new proto type to generate XML
Sep 01 16:34:17 rkrish_ prem: cool
Sep 01 16:34:29 prem http://www.i-scream.org/libstatgrab
Sep 01 16:34:43 nags with which we can decide which script needs to be executed in this run...
Sep 01 16:35:03 nags shaheed, will update more about his impl
Sep 01 16:35:21 shaheed i am implementing a lightweight editor for ldtp using PyGtk. Now the basic framework is ready it looks similar to gedit :).
Sep 01 16:35:37 shaheed the idea behind implementing a seperate editor is that we can add functionality as per our project needs.
Sep 01 16:35:46 nags shaheed: soon we should publish your good and great work :)
Sep 01 16:35:46 rkrish_ shaheed: Can you post a screenshot somewhere for us to see ?
Sep 01 16:35:50 prem this is a really good functionality since it allows us to have more control over the execution of the scripts
Sep 01 16:35:53 * Shingo-- (~Shingonoi@201.36.98.165) has joined #ldtp
Sep 01 16:36:02 shaheed rkrish_: nags: sure
Sep 01 16:36:12 shaheed one such thing is grouping scripts for execution and generating an XML file based on the grouped scripts.
Sep 01 16:36:17 shaheed pyldtp will read that XML file and excute the scripts based on the inputs given by the XML file.
Sep 01 16:36:28 nags shaheed: cool :)
Sep 01 16:36:33 shaheed currently i am implementing the interface part for grouping of scripts, once it is done i will be doing XML file generation.
Sep 01 16:36:41 rkrish_ shaheed: Who generates the XML ?
Sep 01 16:37:11 shaheed rkrish_: in editor it self i am planing to provide that functionality
Sep 01 16:37:26 rkrish_ shaheed: What will the XML file contain ?
Sep 01 16:37:48 nags rkrish_: shaheed has to provide one toolbar with which generate option should be there
Sep 01 16:37:53 shaheed rkrish_: it will contain the script files and corresponding data file
Sep 01 16:37:58 nags on clicking that xml has to be generated
Sep 01 16:38:26 nags shaheed: maybe you can provide a sample xml here ?
Sep 01 16:38:26 rkrish_ shaheed: nags ok
Sep 01 16:39:10 prem nags: should the dat files also be in XML format?
Sep 01 16:39:27 rkrish_ prem: that's what I understood initially
Sep 01 16:39:39 rkrish_ prem: basically, data in xml format
Sep 01 16:39:41 nags prem: yes
Sep 01 16:39:55 rkrish_ nags: Then why 2 XML files ?
Sep 01 16:40:13 nags rkrish_: data xml file is a seperate one
Sep 01 16:40:28 prem then what kind of changes should we make in the scripts for it to take input from a xml file?
Sep 01 16:40:34 nags rkrish_: and the one thats going to be generated by shaheed editor is a different one
Sep 01 16:40:50 rkrish_ nags: ok, I was wondering how the data XML file gets generated ?
Sep 01 16:41:13 nags rkrish_: as of now manually
Sep 01 16:41:14 rkrish_ nags: because It would be difficult for anyone to write test cases in XML format
Sep 01 16:41:25 nags rkrish_: maybe we can ask shaheed to implement in his editor ?
Sep 01 16:41:27 nags shaheed: ?
Sep 01 16:41:36 nags rkrish_: not test cases
Sep 01 16:41:42 nags rkrish_: just data
Sep 01 16:41:46 nags rkrish_: yes you are right
Sep 01 16:41:47 rkrish_ nags: ok
Sep 01 16:41:56 shaheed nags: data XML files??
Sep 01 16:41:57 nags rkrish_: we can have some UI kind of...
Sep 01 16:42:09 nags shaheed: that's a different UI
Sep 01 16:42:16 shaheed nags: ok
Sep 01 16:42:21 prem nags: I guess even giving the necessary data in XML manually is a painful job
Sep 01 16:42:22 nags prem: about new component function that needs to be written, updates ?
Sep 01 16:42:27 shaheed nags: may be we can discus later
Sep 01 16:42:36 nags prem: right
Sep 01 16:42:40 nags shaheed: sure :)
Sep 01 16:42:56 prem Regarding new component functions, as discussed with nags, I am going to try and test many if not all possible widgets in Evolution from python prompt and identify the bugs or new features with respect to component functions.
Sep 01 16:43:09 prem Once this is done, I guess this would help in developing complete automation scripts for evolution without much TODO comments in them ... ;)
Sep 01 16:43:23 nags prem: cool :)
Sep 01 16:43:32 rkrish_ prem: that looks fantastic !
Sep 01 16:43:46 prem nags: rkrish_ : thanks
Sep 01 16:43:47 nags prem: about libstatgrab integration with LDTP ?
Sep 01 16:43:53 rkrish_ prem: I was thinking of how LDTP would manage the threaded message display in Evolution
Sep 01 16:44:18 nags rkrish_: its a tree according to LDTP
Sep 01 16:44:30 nags rkrish_: so there won't be much diff, I guess
Sep 01 16:44:47 rkrish_ nags: ok, basically LDTP must be able to catch a bug
Sep 01 16:45:12 rkrish_ nags: say if a message inside a thread is actually not displayed properly or some such.
Sep 01 16:45:14 nags rkrish_, prem: we should consider this as one of the automation area
Sep 01 16:45:24 nags rkrish_: ya...
Sep 01 16:45:32 prem nags: sure
Sep 01 16:45:41 prem nags: As i told earlier libstatgrab eases our job of getting performance related statistics
Sep 01 16:45:43 nags prem: about libstatgrab integration with LDTP ?
Sep 01 16:45:58 nags prem: okay...
Sep 01 16:45:59 prem nags:
Sep 01 16:45:59 prem The only issue that I am not clear with that is how and when should these statistics be given to the user ... should it be logged periodically or given only when the user rquests for ...
Sep 01 16:46:00 rkrish_ kaveri: what do you think about this ?
Sep 01 16:46:29 rkrish_ prem: I think, it should be left to the test case
Sep 01 16:46:38 nags prem: as of now we can give it to the user when required
Sep 01 16:47:04 nags prem: as rkrish_ suggested, we can take that as an arg and based on that we can act upon
Sep 01 16:47:13 prem nags: ok
Sep 01 16:47:20 nags prem: maybe we can discuss further in detail about this in offline :)
Sep 01 16:47:25 prem nags: sure
Sep 01 16:47:26 rkrish_ prem: for example, there could be a case where we do some bulk loading of contacts on evo and then during that time, call this function
Sep 01 16:47:46 rkrish_ prem: So, I think, it is best left to the tester's judgement
Sep 01 16:47:55 prem rkrish_: ok
Sep 01 16:48:04 nags prem: any updates for us about your new LDTP algo ?
Sep 01 16:48:14 nags prem: when it will be ready ?
Sep 01 16:48:24 prem Implementation of the new algorithm is complete except for some issues with respect to remapping.
Sep 01 16:48:29 prem But I havent tested the implementation fully, which I will do in the next week.
Sep 01 16:48:34 nags prem: okay
Sep 01 16:48:40 nags prem: fine...
Sep 01 16:48:42 rkrish_ prem: which algo is this ?
Sep 01 16:48:50 nags prem: so we can have this in production by next week ?
Sep 01 16:49:01 nags rkrish_: change in appmap with child index
Sep 01 16:49:02 * rkrish_ is loosing track
Sep 01 16:49:02 prem rkrish_: Its about getting the gui handle
Sep 01 16:49:16 nags and get the gui handle of child directly
Sep 01 16:49:20 rkrish_ prem: oh, yeah
Sep 01 16:49:25 nags instead of traversing the whole loop
Sep 01 16:49:27 prem nags: Hopefully we should be able to have that in production ;)
Sep 01 16:49:28 rkrish_ prem: that is to improve the performance, right ?
Sep 01 16:49:36 prem rkrish_: exactly
Sep 01 16:49:37 nags prem: cool :)
Sep 01 16:49:47 nags rkrish_: right
Sep 01 16:50:23 nags thats all from my side
Sep 01 16:50:27 nags any other updates guys ?
Sep 01 16:50:29 rkrish_ prem: I have a question about the cpu stuff
Sep 01 16:50:44 nags rkrish_: yes please
Sep 01 16:50:47 prem rkrish_: sure
Sep 01 16:50:58 rkrish_ prem: Does the utility give you the CPU utilization for the given time period ?
Sep 01 16:51:07 nags rkrish_: yes
Sep 01 16:51:24 rkrish_ nags: Yeah, but we need to observe the CPU util for some time right ?
Sep 01 16:51:33 nags rkrish_: I mean only the max utilization
Sep 01 16:51:48 nags rkrish_: ya the coder can use it where ever required
Sep 01 16:51:51 prem rkrish_: It actually starts monitoring the statistics once it is initilised and we can stop them whenever we dont want it
Sep 01 16:51:53 rkrish_ nags: right, but how would you measure it over a peroid of time ?
Sep 01 16:52:17 rkrish_ prem: But will it run in parallel to other scripts ?
Sep 01 16:52:18 nags rkrish_: a seperate thread will keep an eye of whats happening
Sep 01 16:52:22 prem rkrish_: yes
Sep 01 16:52:25 rkrish_ nags: cool...
Sep 01 16:52:32 rkrish_ nags: wise men think alike
Sep 01 16:52:36 prem I have implemented it as a seperate thread
Sep 01 16:52:40 nags rkrish_: :)
Sep 01 16:52:59 rkrish_ prem: Have you already implemented it ?
Sep 01 16:53:05 nags shaheed: any updates ?
Sep 01 16:53:14 nags rkrish_: yes saw the initial demo yesterday
Sep 01 16:53:20 shaheed nags: thats all from my side
Sep 01 16:53:23 rkrish_ nags: cool
Sep 01 16:53:30 nags rkrish_: still few minor issues with libstatgrab
Sep 01 16:53:32 prem rkrish_: almost complete ... but I was not sure about when i should give the values to the user
Sep 01 16:53:38 nags rkrish_: which should get fixed soon
Sep 01 16:53:48 rkrish_ prem: Will it be a separate function to call ?
Sep 01 16:53:54 nags rkrish_: yes
Sep 01 16:53:59 rkrish_ nags: ok
Sep 01 16:54:02 prem rkrish_: yes ...
Sep 01 16:54:13 rkrish_ prem: cool
Sep 01 16:54:16 nags prem: any other updates ?
Sep 01 16:54:20 nags prem: evolution scripts ?
Sep 01 16:54:22 prem nags: I am done
Sep 01 16:54:45 nags prem: okay...
Sep 01 16:54:51 nags rkrish_: any updates ?
Sep 01 16:55:04 rkrish_ nags: none at this point
Sep 01 16:55:09 nags rkrish_: okay
Sep 01 16:55:33 nags haip_away and varadhan: helped us for evolution testing to proceed using LDTP scripts
Sep 01 16:55:37 rkrish_ nags: you may want to call Kaveri and find out what's going on - on their side .
Sep 01 16:55:52 rkrish_ haip_away: Thank you very much
Sep 01 16:55:52 nags haip_away and varadhan: I take this opp to thank both of them
Sep 01 16:56:03 rkrish_ varadhan: great work
Sep 01 16:56:06 nags rkrish_: I had some discussion with her...
Sep 01 16:57:01 nags thanks guys
Sep 01 16:57:10 nags see you next in community meeting...
Sep 01 16:57:12 nags bye for now
Sep 01 16:57:29 rkrish_ nags: Yes, we are all busy with GNOME 2.12 release
Sep 01 16:57:40 rkrish_ nags: So, hopefully they will be available after the release.
Sep 01 16:58:00 rkrish_ nags: ok
Sep 01 16:58:02 rkrish_ nags: bye
Sep 01 16:58:12 nags rkrish_: ya
