Vesuvius July 16, 2013 IRC about GSoC

Listed here because official logging didn't capture this session. Time zone is EDT.


Status #sahana-meetingX snehaX 

(no topic set)

[11:09] == glennp [9c287802@gateway/web/freenode/ip.] has joined #sahana-meeting

[11:09] <akanduru> I am back

[11:10] <glennp> hi, finally here.  It's gonna be that kind of day.

[11:10] <ramindu> As I was saying, bonne huit, mes amis

[11:10] == kaushikaruk [~AndChat32@] has joined #sahana-meeting

[11:11] <glennp> Hi, ramindu.  Why don't you and kaushikaruk start off today's chat?

[11:13] <glennp> Also, roshan and mayank

[11:13] <ramindu> kaushikaruk, I took a look at your code

[11:13] <ramindu> Working pretty well

[11:13] <ramindu> So I'd say on track

[11:14] <mayank_jain> Hi glennp roshan .

[11:14] <ramindu> I think for this week, let's get Alex to review it as well

[11:14] <mayank_jain> I have completed the event-manager part (including addition of category)

[11:14] <ramindu> I think it'll be a good idea to get his input once we've merged each module

[11:15] <mayank_jain> I have started working on managing resource pages. So far I have been able to do the following :

[11:15] <mayank_jain> i)Options for choosing a) Template b) Pages c) None of these to create a page.

[11:15] <mayank_jain> ii) I started with the Template part first.

[11:16] *sneha:* I would've started you first, but dilantha's not here yet.  Take a look at what I posted 1/2 hour ago to discuss@, particularly Greg's doc

[11:16] == No such nick/channel: sneha:

[11:16] *sneha*  I would've started you first, but dilantha's not here yet.  Take a look at what I posted 1/2 hour ago to discuss@, particularly Greg's doc

[11:17] <mayank_jain> iii) The editor can select a Template. Based on his selection , the title and the content of the content of the page changes (in the interface)

[11:17] <mayank_jain> iv) He can edit and save the pages.

[11:17] <kaushikaruk> Hi ramindu..

[11:18] <kaushikaruk> Sry for the delay..

[11:18] <kaushikaruk> Still onthe way to home..

[11:18] <ramindu> It's ok

[11:18] <mayank_jain> I used sample templates from Red Cross's site temporarily.

[11:18] <ramindu> Got my earlier messages?

[11:18] <kaushikaruk> Still more to commit some code..

[11:18] <kaushikaruk> Yes i saw..

[11:18] <mayank_jain> The ones which louiqa suggested

[11:19] <kaushikaruk> Will commit them as well..

[11:19] <ramindu> Oh ok

[11:20] <glennp> mayank: how many templates so far?

[11:20] <ramindu> We'll both try to reach out to Alex this week

[11:20] == dilantha [70879136@gateway/web/freenode/ip.] has joined #sahana-meeting

[11:20] <mayank_jain> I simply used the existing pages from here :

[11:21] <kaushikaruk> Ok ramindu..

[11:21] <ramindu> There are two more modules in Kilauea to be merged, right

[11:21] <mayank_jain> of all possible disaster events

[11:21] <kaushikaruk> Will get a feedback from his as well

[11:21] <mayank_jain> I haven't started to work on creating Templates. I planned of doing it in 2nd phase

[11:22] <mayank_jain> There are four template for each disaster. about/prepare/respond/recovery

[11:22] <ramindu> kaushikaruk, any other blockers/ideas?

[11:22] == triune [~triune@unaffiliated/triune] has joined #sahana-meeting

[11:23] <glennp> Looks like 22 disaster categories, and and as you say, 4 tabs per category

[11:24] <glennp> But don't you need a few templates now to test your code?

[11:24] <mayank_jain> I didn;t use Pet safety and power outage .. 4 of each

[11:25] <kaushikaruk> Yes.. bt some more remaining on cluent module as well..

[11:25] <mayank_jain> I used these templates itself to test the code ..

[11:25] <kaushikaruk> Did u able to look at that date js thing..

[11:25] <mayank_jain> Like created a new flood event. Used the flood templates to create four diff pages. Modified the content a bit to reflect relationship with the event.

[11:26] <ramindu> I'll take a look and contact you regarding that

[11:27] <glennp> OK, that sounds good.   So when you say "I haven't start to work on creating Templates [yet]", you mean from scratch, rather than reusing existing HTML?

[11:28] <mayank_jain> Changes like adding local emergency contact nos. to specific flood event.

[11:28] <mayank_jain> Yes! I simply copied them with minor changes from Red cross.

[11:28] <mayank_jain> Sahana might have different needs.

[11:29] <kaushikaruk> Ok ramindu..

[11:29] == Sneha1 [~sneha@] has joined #sahana-meeting

[11:29] <mayank_jain> I am focusing now on creating pages first. Then will move on to templates.

[11:29] <glennp> That's true.  And different installs of Vesuvius might eventually have different style guides for these pages

[11:29] == sneha [73fd301f@gateway/web/freenode/ip.] has quit [Quit: Page closed]

[11:29] <kaushikaruk> I hv completed the search as well

[11:29] <mayank_jain> Pages => System and Resource

[11:30] <mayank_jain> Yes.

[11:30] <kaushikaruk> Will continue on edit paerson and other related forms..

[11:30] <ramindu> Great!

[11:31] <mayank_jain> I have a doubt regarding the display of resource pages. Currently since the pages are not mapped with events all pages are displayed. So even if a person is not inside an event he can see all pages

[11:31] <mayank_jain> What should happen now given that the pages are mapped with events?

[11:33] <glennp> We want to get to where pages mapped to events are seen just with those events

[11:34] <mayank_jain> Redirect him to a page where he will have to select an event?

[11:34] <glennp> Other pages are either - not mapped to any event because they are drafts or future items

[11:35] <glennp> or pages intended to be seen with all events (without having to specify them all)

[11:36] <glennp> So a person not inside an event should see all of those last type of pages.

[11:36] <mayank_jain> ohk. Only display Not event-specific pages(system pages) and pages which r mapped to all events.

[11:36] <mayank_jain> Fine.

[11:37] <glennp> I'm thinking the non-event-specific system pages and pages mapped to all events are almost the same thing

[11:37] <mayank_jain> Also I noticed now : How does one create a System Page. I didn't see any specific mechanism of designating a page Resource or System.

[11:38] <mayank_jain> So I am explicitly asking the user about whether he wants to create a resource page or system page.

[11:39] <glennp> Just checked with triune about that.

[11:39] <kaushikaruk> Thanks ramindu..

[11:39] <glennp> He says you can't do it through GUI.  Instead, create a new resource page, then go to database, change page ID from positive to negative.  Magic!

[11:40] <glennp> Ramindu, kaushikaruk... are you done?

[11:40] <mayank_jain> I have  added an option for choosing the type.

[11:41] <mayank_jain> This will resolve this

[11:41] <ramindu> kaushikaruk, I guess we're done.

[11:41] <glennp> mayank: ok, sounds good

[11:41] <ramindu> Please do drop the weekly update e-mail.

[11:41] <kaushikaruk> Ok sure..

[11:42] <kaushikaruk> Thanks for the time..

[11:42] <kaushikaruk> Bye..

[11:42] <mayank_jain> ok. :)

[11:42] <glennp> guru, triune, chat time.

[11:42] <guru___> k

[11:43] <guru___> i have committed some codes on launchpad, and updated design.

[11:43] <guru___>

[11:43] <triune> guru has already provided me some updated design documents and some code to begin looking at :D

[11:44] <guru___>

[11:44] <guru___> yes i mention here for others..  :)

[11:44] <glennp> mayank: do the templates and or your planned pages have a way to pick up the database fields yet?

[11:45] <glennp> like disaster title

[11:45] <mayank_jain> glennp: Sorry. I didn't understand your question

[11:46] <mayank_jain> Yeah! When you select a Template , the page's title automatically get's replaced with the template's. If that's what u meant

[11:47] == dilantha [70879136@gateway/web/freenode/ip.] has quit [Quit: Page closed]

[11:47] <triune> guru___: any burning questions at this point? I still have not completed reading everything over yet

[11:48] <glennp> But when the page is viewed in the context of a particular event, e.g., pakistani flood 2012, then we might want to auto-show that, not just "flood"

[11:49] == dilantha [70879136@gateway/web/freenode/ip.] has joined #sahana-meeting

[11:49] <triune> but here's a few things i was thinking: 1) sahana.conf should have a $conf['sync_api_key'] (64 character string key) that a foreign instance will need to have in order access the web service; this should be managed via the gui sync page you are working on near the wsdl location field

[11:49] <mayank_jain> The event name is also retrievable along with the template title. I can merge them in a way to make sensible title ..

[11:50] <guru___> yes

[11:50] <triune> 2) there should be a script that you can call/load as a cron task that will check if there are an servers or events that need to be sync'd (are configured to sync) and sync them at an interval

[11:50] <mayank_jain> Pakistan Flood 2012 : Recovery.. somthng like this

[11:51] <triune> I'm not sure how you want to handle these notes, perhaps just addendum to your design docs ;)

[11:51] <mayank_jain> When the user selects the template page, i am retrieving all its contents from the db. Info about the event can also be pulled up.

[11:51] == kaushikaruk [~AndChat32@] has quit [Ping timeout: 240 seconds]

[11:52] <glennp> OK.  Play around with a modified template that does something like that.

[11:52] <guru___> cron for the person data sync ryt?

[11:53] <mayank_jain> Ok.

[11:53] <mayank_jain> I will work on it

[11:53] <glennp> Sounds like you have a reasonable plan for the week ahead.  Anything further?

[11:54] <triune> optimally you will have a table that you can store your "syncable events" along with their respective keys and wsdl locations, perhaps have a column for errors too

[11:54] <mayank_jain> Nopes. Thanks :)

[11:54] <guru___> ok

[11:54] <guru___> will add that too

[11:54] <guru___> triune: i have encounter a prob with incident_id configuration..

[11:54] <glennp> Thanks, mayank.  Sneha1, dilantha, hi, guys.

[11:55] <triune> this table will be looked up via cron and the events with appropriate data will be sync'd .... if there are errors encountered, they will be captured so that cron can throttle back its sync period on that instance/event

[11:55] <Sneha1> hi glennp

[11:55] <triune> guru___: please explain

[11:55] <dilantha> hi glennp

[11:56] <glennp> Sneha: were you able to track down that INSTALL doc?

[11:56] <Sneha1> yes.

[11:56] <Sneha1> but the installer codes in the existing code base, isnt working.

[11:57] <Sneha1> i read the mails and greg's doc

[11:57] <glennp> This is true.  Probably only useful as a starting point for stealing code snippets

[11:57] <Sneha1> for the demo manager.

[11:58] <guru___> here i'm using incident_id for perticular event, but incident_id will not going to be same for each instance,

[11:58] <guru___> is there any global variable to identify events..?

[11:59] <glennp> Greg's put most of the effort into the shell script stuff, so his GUI design is a sketch.

[12:00] <triune> guru___: that would be the global variable ... but its only global to the vesuvius instance, not the world

[12:00] <dilantha> Sneha1: do you have any unclear points in the greg's document ?

[12:01] <guru___> ya so when we call sync, does it ok to use incident name...

[12:01] <triune> guru___: could your syncPerson class override incident_id so that it is a static variable to the local instance that is being sync'd into?

[12:02] <glennp> He has suggested to simplify by working on the mechanism, with a simple GUI run by the admin.  It's a different vision than mine, but maybe you could design your GUI in such a way that it supports Greg's approach first but can grow towards mine?

[12:02] <guru___> triune: Not like that, will have to change..

[12:02] <Sneha1> what I understood is, the shell script will 1. copy the existing 'master' vesuvius codes into the codes of the particular instance, 2. get the database from the eixsting codes?

[12:02] <glennp> Where Greg's is "admin run", mine is "user run"

[12:03] <Sneha1> yes. the admin is responsible for running the shell script from the UI

[12:03] <triune> glennp & Sneha1 its all about milestones... working demo manager first and all the bells and whistles for a fancy demo manager gui later ;)

[12:04] <triune> guru___: that's just a thought.... i think sync process will somehow have to make sure incident_id is set to the local event's incident_id before person->save() is called

[12:05] <Sneha1> once the demo is initialised then all the other functionalities can come with it. right?

[12:05] <glennp> triune: not disagreeing, just saying the design of the working demo manager should be influenced by the goal of eventually turning it over to end users, so don't make it way-geeky to start.

[12:06] <triune> Sneha1 & glennp : 3 buttons is pretty simple in my opinion, and yes more functionality later

[12:06] <triune> sorry, 3 per instance and just 1 to create a new instance :)

[12:07] <guru___> triune: yes, can be managed.. and i had trouble on using admin user

[12:07] <guru___> it breaks with fatal errors in the middle..

[12:08] <triune> you can email me errors if u need help debugging them ;)

[12:08] <guru___> k

[12:09] <glennp> triune:  well, presumably, your "create new instance" button then starts the wizard to select which preload, etc., so there's more to the design than 3 buttons.

[12:11] <triune> that is true. perhaps the instances themselves are just links to a deeper page with all the respective actions: the three buttons I mentioned, and other tools to manage the instance

[12:12] <glennp> sneha: I'm not sure what you're asking.

[12:12] <Sneha1> the preloading the database will be a feature for the user right? not the admin?

[12:13] <Sneha1> so after the instance is created (with the help of the admin using the wizard) the end user will get the chance to preload his instance database right?

[12:13] <triune> guru___: any other questions? If not, I'll review your code changes and design doc and email you any concerns

[12:14] <Sneha1> i mean the wizard for the admin will be a different from the wizard for the end demo user.

[12:14] == roshan [~roshan.h@] has quit [Quit: Leaving]

[12:14] <guru___> k

[12:14] <guru___> thats all

[12:14] <glennp> This bifurcation between end user and admin is problematice.

[12:15] == ramindu [~rami@] has quit [Read error: Connection reset by peer]

[12:15] <triune> glennp: i agree... just assume for now that admin and user are the same thing

[12:15] == mayank_jain [~mayank@] has quit [Quit: Leaving]

[12:15] <triune> design a single interface

[12:15] <Sneha1> ok

[12:15] <glennp> guru, triune... thanks.  mohit21, akanduru, chat time.

[12:15] <akanduru> Hello Mohit

[12:16] <mohit21> Hi akanduru

[12:16] <glennp> Sneha:  another approach:  make an end-user friendly interface that just solicits and packages up the end-user request, and then send it to the admin, who excutes Greg's design

[12:17] <mohit21> I was working on searchSolrForEvents

[12:17] <mohit21> I have changed the method for building parameters

[12:17] <akanduru> Okay.

[12:17] <triune> Sneha1: as far as preloading event data, i believe that should also accompany other instance gui actions...

[12:18] <mohit21> It returns a return type PLSolrParams only

[12:18] <mohit21> But I've changed the content

[12:18] <mohit21> It now returns something like Image Type:jpeg

[12:19] <akanduru> Okay. type can be any file type, right?

[12:19] <glennp> triune: I think you just said the admin does it, but the end-user must communicate the desired choice to the admin (which you could do with what i just suggested).

[12:20] <mohit21> Sorry I didn't get you

[12:20] <akanduru> The image file type depends on the file extension of the image. It can be jpg, png, gift etc.

[12:20] <mohit21> Yes

[12:21] == ramindu [~rami@] has joined #sahana-meeting

[12:21] <mohit21> If I select jpeg in the side bar this will be added to the searchParams

[12:21] <triune> i cant say much more on the user/admin stuff other than what I had envisioned was the demo manger being the base class, if you will, and other functionality can later be added (extending that class) after its completed

[12:21] <triune> nothing wrong with a very simple system that could easily be incremented upon

[12:22] <triune> glennp: sounds reasonable :)

[12:22] <akanduru> Right. We discussed a lot of issues yesterday. any new issues cropped up?

[12:23] <Sneha1> ok.. so the user GUI for now will send the request to the demo manager

[12:24] <Sneha1> and the demo manager will create and initialse the instance

[12:24] <Sneha1> (using the shell script)

[12:24] <mohit21> I searched a bit on the problem with the url

[12:25] <glennp> Sounds like a reasonable direction.  The communication between the 2 systems could be passed through a new database table (probably less work than email, web services, etc.)

[12:25] <mohit21> I couldn't figure out a method using which the external directory can be mapped in all servers

[12:25] <akanduru> For the URL issue, first make sure that you can access one URL from http request.

[12:26] <akanduru> If you can resolve that, you should be good.

[12:26] <mohit21> I guess if I am able to map the image directory to a url that should work

[12:26] <akanduru> right.

[12:27] <akanduru> Your image directory should be accessible to the web server.

[12:27] <glennp> What needs to be passed is the initial user profile info (for setting up an admin account) and choice of preloads, at the least

[12:28] <mohit21> I was trying to find a way so that by configuring the xml files of the application I can map the directory in different types of servers alike

[12:28] <mohit21> Without any changes in the server configurations

[12:29] <mohit21> I couldn't find how to do that

[12:29] == guru___ [~guruthars@] has quit [Quit: Leaving]

[12:29] <akanduru> That is needed. But can you display the images if you hardcode the path now?

[12:29] <mohit21> No

[12:30] <mohit21> I haven't configured it

[12:30] <mohit21> The application is not able to show the image

[12:30] <akanduru> Test with hardcoded path.

[12:31] <mohit21> What do you mean by hardcoded path?

[12:31] <akanduru> If that works you can figure out using env variables.

[12:31] <glennp> Also, requested name of instance, probably minimal info about organization, maybe initial disaster event name [tho that might be in preload]

[12:31] <mohit21> I can access a directory within the application

[12:31] <mohit21> But accessing external directory is a problem

[12:32] <Sneha1> all the functions that triune has enlisted for the shell script like creating the database, the mysqldump. editing htacess and sahana.conf etc will be done from the shell?

[12:32] <akanduru> Using http get request?

[12:32] <glennp> The GUI launches the shell operations.

[12:33] <Sneha1> ok.

[12:33] <akanduru> I am not talking about application access.

[12:33] <akanduru> I am talking about web server access.

[12:33] <dilantha> Yes.. you will have to execute the shell from the background..not manually

[12:34] <glennp> Not just launches, but also configures the parameters of the shell commands

[12:34] <triune> right!

[12:34] <mohit21> I have left the issue for now ... I'll work on it a bit and then if I'm not able to solve it I'll talk to you

[12:34] <glennp> So its not a fixed script, but generated (might be multiple scripts)

[12:35] <mohit21> Right now I'm trying to write the methods for search correctly

[12:35] <mohit21> Once I am able to search the images I'll try to work on the issue

[12:36] <akanduru> Sure.

[12:36] <Sneha1> multiple scripts for the different purposes?

[12:36] <Sneha1> or the same one called from the GUI?

[12:36] <mohit21> I had this question that sort is used for solr

[12:36] <mohit21> So do I need to sort the results as well?

[12:37] <glennp> Up to you if its one big script, or multiple little ones.  Generated and called from the GUI in either case.

[12:37] <akanduru> Yes. There needs to be some sort method. First you have to figure out what sort methods you can provide.

[12:37] <Sneha1> ok.

[12:38] <akanduru> I can think of two methods.

[12:38] <akanduru> alphabetical order of filenames, file timestamp

[12:38] <glennp> Error detection & recovery may be easier with multiple smaller scripts, but probably would run slower.

[12:40] <glennp> Just guessing here.

[12:40] <akanduru> It will be good if you can do file extensions, but that may not be that important since we provide a way to filter by file extension.

[12:40] <mohit21> How would I get the timestamp?

[12:40] <akanduru> You should eb able to get file timestamps from java io classes.

[12:40] <mohit21> ok

[12:41] <mohit21> And what should be the imageid?

[12:42] <akanduru> Is that integer?

[12:42] <mohit21> yes

[12:42] <triune> As I mentioned in the doc, has 3 parameters passed into it.... but what is involved in cloning an instance shouldn't really change much others than those 3 parameters :: also, I think the question was asked in email: should demo manager be running off of a branch? I would think that it should be based off of a stable release, not a trunk per se

[12:42] <akanduru> Ahh. We may have a bit of problem here.

[12:43] <glennp> Probably one way to proceed is to get a simple form of script working without the GUI(s) first, and values hard-coded.

[12:43] == ramindu [~rami@] has quit [Quit: Bye]

[12:44] <akanduru> That integer type is used because the image names are stored in mysql/solr.

[12:44] <Sneha1> ok.

[12:44] <akanduru> You may have to change the type to a string.

[12:45] <rami> ramindu is now known as /me

[12:45]  * rami replaces ramindu

[12:45] <Sneha1> that should probably be my first step for now. creating the

[12:45] <rami> :D

[12:45] <mohit21> Then the image names might not be unique

[12:46] <glennp> "simple" as in without error recovery.  Then next step, work to have Demo mgr GUI generate that script with parameters passed.

[12:46] <akanduru> but the sql/solr queries may be expecting integers, so you may need to convert to integer types in PL interface.

[12:47] <akanduru> You want to keep the image id separate from image names.

[12:47] <akanduru> Sometimes, image name could imply name of person in the picture.

[12:48] <glennp> Yup, getting that functionality working is important.

[12:48] <akanduru> I am talking filename mapping to imageid and make imageid a string in PLRecord.

[12:49] <mohit21> Then what would be imageid?

[12:50] <akanduru> imageid can be concatination of event name (in your case it it is directory) and filename.

[12:51] <mohit21> Should I keep PLRecord as it is ... or should I remove the variable which are not to be used by me

[12:52] <glennp> regarding triune's comment about based off the stable release... that should be eventually, but for now, Sneha, you can use whatever Vesuvius branch you are currently working with as the "master" to clone for demos.

[12:53] <akanduru> You mean the fields inside the PLRecord, let them there. It becomes a generic structure for PL and you ssearch items.

[12:53] <glennp> ...says triune

[12:54] <glennp> sneha, dilantha... anything else this week?

[12:54] <mohit21> ok

[12:55] <triune> :D

[12:55] <mohit21> And the color_channels is causing a problem

[12:55] <mohit21> The default value is -1

[12:56] == Akila [~akila@] has joined #sahana-meeting

[12:56] <mohit21> So It generates the alert bar "No skin annotation is required"

[12:56] <akanduru> What is the problem?

[12:56] <akanduru> Hmm

[12:56] <mohit21> The condition is that it should be greater than equal to 3

[12:56] <dilantha> that's all from my side

[12:57] <glennp> I see Akila has just joined us.  Rami, are you still there?

[12:57] <Sneha1> the cloning will be done from the current branch?

[12:57] <Akila> Hi

[12:58] <glennp> Sneha - Yes, that's fine for the development period.  At actual deployment, post-GSoC, a stable branch would be better.

[12:58] <akanduru> That message implies that it is black and white picture.

[12:59] <mohit21> hmm

[12:59] <Sneha1> ok.

[12:59] <rami> yup, still here

[12:59] <rami> Akila, how goes things?

[12:59] <Akila> Hi Ramindu

[12:59] <akanduru> By default you can set more than three for your search results.

[13:00] <Akila> I was working on module specific dynamic app creation in the last week

[13:00] <mohit21> ok

[13:00] <akanduru> I have to think about it.

[13:00] <akanduru> Perhaps 3 might also work.

[13:00] <Akila> what I did was filter out folders to be copied in the 'mod' folder

[13:00] <Akila> not sure whether that approach is okay

[13:00] <akanduru> You assume everything color...

[13:01] <mohit21> okay

[13:01] <Akila> but still the whole database dump is created

[13:02] <glennp> sneha - I've got to run.  Good chat this week, all.  Rami, could you take over hosting duties, until you, akanduru, etc. are done?  Thanks.

[13:02] <akanduru> You are making good progress.

[13:03] <dilantha> ok.glennp thanks for your time

[13:03] <akanduru> You have anything else to talk?

[13:03] <dilantha> and the support :)

[13:04] <mohit21> I am trying to complete the search methods.

[13:04] <glennp> :D  bye

[13:04] <dilantha> bye

[13:05] <mohit21> Hope I can make the application working by the end of this week

[13:05] <dilantha> Sneha1: lets windup then ?

[13:05] <Sneha1> sure

[13:06] <Sneha1> this week making the clone script is the target?

[13:06] <Sneha1> with the simple gui?

[13:06] <mohit21> No that's all

[13:06] <akanduru> Okay. Ping me if you face more problems. There is more to it than just makeing the search work.

[13:06] <dilantha> try to complete the shell script first..and make it work

[13:07] <dilantha> after than we can do the GUI

[13:07] <Sneha1> ok.

[13:07] <mohit21> Ya sure.

[13:07] == rami [] has quit [*.net *.split]

[13:07] == triune [~triune@unaffiliated/triune] has quit [*.net *.split]

[13:07] <akanduru> There is sqllite part that needs work.

[13:07] <Sneha1> ill need to update the workflow also with the new developments.

[13:08] <dilantha> yes

[13:08] <dilantha> and dont forget to put your weekly update to the mailing list

[13:08] <mohit21> hmm

[13:08] <akanduru> Also take some time to update sahana wiki, if you haven't already.

[13:09] <Sneha1> ok.

[13:09] == ramindu_ [~ramindu@] has joined #sahana-meeting

[13:09] <mohit21> Yes I'll be doing that

[13:09] <ramindu_> hi Akila, sorry, got disconnected

[13:09] <ramindu_> so where were we?

[13:09] <ramindu_> ah yes, copying over the relevant module folders from the /mod folder is the corect approach

[13:09] == triune [~triune@unaffiliated/triune] has joined #sahana-meeting

[13:09] == rami [] has joined #sahana-meeting

[13:09] <Akila> no problm... on module specific portable app

[13:10] <Akila> yup

[13:10] <ramindu_> because Vesuvius checks which folders are available there and loads them as modules

[13:10] <akanduru> Okay, good bye mohit.

[13:10] <ramindu_> bye, mohit!

[13:10] <mohit21> Goodbye !! :)

[13:10] <ramindu_> and akanduru, see you next week! :)

[13:11] <ramindu_> anyway Akila, what do you tihnk about limiting the database based on the modules that are included in the portable app as well?

[13:11] <ramindu_> does that seem feasible?

[13:11] <dilantha> bye Sneha1

[13:11] <ramindu_> there are database tables used based on particular modules

[13:11] <akanduru> See you ramindu!

[13:11] <Akila> it's going to be very complicated

[13:12] <ramindu_> maybe we'll only have the tables that are needed by the modules being installed?

[13:12] == akanduru [820efe18@gateway/web/freenode/ip.] has quit [Quit: Page closed]

[13:12] <ramindu_> yeah I guess we'll just pull the whole database

[13:12] <Akila> what about user accounts?

[13:12] <ramindu_> it'll be more annoying if someone decides to put one of the modules back in to a particular instance as well (if we don't have the database tables)

[13:12] <ramindu_> what do you mean?

[13:13] == dilantha [70879136@gateway/web/freenode/ip.] has quit [Quit: Page closed]

[13:13] <Akila> user login credentials are exposed through db dump

[13:13] <Akila> should we add only a default login user account or something?

[13:14] <Akila> I mean in the db dump

[13:16] <ramindu_> not really

[13:16] <ramindu_> it's a hash right

[13:16] <ramindu_> with a salt

[13:16] <ramindu_> not really reversible

[13:17] <ramindu_> :)

[13:17] <ramindu_> usernames are exposed but not passwords

[13:17] <Akila> true, but might be a security risk

[13:17] <ramindu_> maybe we could obfuscate the whole DB dump?

[13:17] <Akila> salt is also stored in the db ryt?

[13:18] <Akila> obfuscate as in encrypt the file?

[13:18] <ramindu_> yup

[13:18] <ramindu_> yes to both questions :)

[13:19] <Akila> but encryption key should also be available to the user

[13:19] <ramindu_> nah it can be in the code itself

[13:20] <ramindu_> but again there's some risk there

[13:20] <ramindu_> hmmmm...

[13:20] <Akila> code is also exposed

[13:20] <ramindu_> maybe ask whomever is installing the portable app to create the super admin account?

[13:20] <ramindu_> without having credentials available in the dump?

[13:21] <Akila> u mean like put a default login?

[13:21] <Akila> like in default vesuvius installation sql

[13:21] == mohit21 [7ab1a30c@gateway/web/freenode/ip.] has quit [Ping timeout: 250 seconds]

[13:23] == r0bby [~wakawaka@guifications/user/r0bby] has joined #sahana-meeting

[13:23] <ramindu_> nah, you have a first run check on the portable app, right?

[13:23] <ramindu_> maybe ask them to create an admin user during that phase?

[13:24] <ramindu_> and store that in the DB?

[13:24] <ramindu_> how does that sound?

[13:25] == r0bby_ [~wakawaka@guifications/user/r0bby] has joined #sahana-meeting

[13:26] == r0bby [~wakawaka@guifications/user/r0bby] has quit [Disconnected by services]

[13:26] == robbyoconnor [~wakawaka@guifications/user/r0bby] has quit [Disconnected by services]

[13:26] == r0bby_ has changed nick to robbyoconnor

[13:26] <Akila> sounds good

[13:27] <Akila> so the users table will not be imported? (only structure)

[13:27] == rami1 [~ramindu@] has joined #sahana-meeting

[13:27] == rami1 has changed nick to ramindu

[13:27] <ramindu> Akila, I might have missed some of your messages

[13:27] <Akila> oh

[13:27] <ramindu> something wrong with Internet connection ;)

[13:28] <Akila> so we will exclude the users table data in db dump process?

[13:28] == ramindu_ [~ramindu@] has quit [Ping timeout: 240 seconds]

[13:29] <ramindu> yup

[13:29] <ramindu> that's what I'd suggest

[13:29] <Akila> sounds good

[13:29] <ramindu> the security concern you raised is quite valid

[13:30] <Akila> so in the first run it will prompt the user to create an admin account

[13:30] <Akila> then create a new user in Vesuvius

[13:30] <ramindu> but I'd like to suggest one thing

[13:30] <Akila> (admin user)

[13:30] <Akila> yup?

[13:30] <ramindu> let's assign all the missing person reports to the admin user when that user is created

[13:30] <ramindu> I mean the reports that already exist in the system

[13:31] <ramindu> otherwise those reports would lose visibility

[13:31] <Akila> do we need existing reports in the portable app?

[13:32] <ramindu> wouldn't we, if we're creating the app to include existing data on a particular instance?

[13:32] <Akila> well...I'm not quite sure about filtering data for the portable app

[13:32] <ramindu> I see.

[13:32] <ramindu> for now, let's have this approach:

[13:33] <Akila> sure

[13:33] <ramindu> all users get deleted during data dump for Portable App creation

[13:33] <ramindu> admin user is created upon re-instantiation of that app elsewhere

[13:33] <Akila> loud and clear :)

[13:35] <Akila> seems like Greg can't make it for the Hangout on 22nd

[13:39] <ramindu> yeah

[13:40] <ramindu> ok then, shall we wrap this up?

[13:40] <Akila> sure

[13:40] <ramindu> :)

[13:40] <ramindu> hope to see your weekly update

[13:40] <ramindu> keep up the good work!

[13:40] <ramindu> goodnight

[13:40] <Akila> I'm on it

[13:40] <Akila> goodnight!

[13:42] == ramindu [~ramindu@] has left #sahana-meeting []



QR Code
QR Code agasti:vesuvius:gsoc2013:july16irc (generated for current page)