#MAPBENDER IRC Log - 2008-06-09

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
04:17:52 ticheler: seven: When you're there, please ping me on Bolsena
04:56:28 seven: There.
04:56:43 seven: Here.
04:56:48 seven: Where.
04:56:55 seven: ticheler: ping
04:56:57 seven: :- )
05:24:32 ticheler: seven: Here
08:01:16 Testbaudson: 1700
08:01:23 Testbaudson: time for our weekly IRC meeting
08:01:40 Testbaudson: let's start with the roll call
08:01:47 Testbaudson: Christoph
08:02:18 LenKne: Len
08:03:24 ur: Uli
08:04:04 Testbaudson: Today's agenda can be found here:
08:04:34 Testbaudson: http://www.mapbender.org/IRC_Meeting#Meeting_Agenda_for_June_9.2C_2008
08:04:35 sigq: Title: IRC Meeting - MapbenderWiki ( at www.mapbender.org )
08:04:47 Testbaudson: any additions to the agenda?
08:05:10 spsneo: Hello devs!
08:05:19 spsneo: I am sorry I am bit late
08:05:30 spsneo: Here theres some problem with net
08:05:44 Testbaudson: no problem spsneo
08:05:48 Testbaudson: we have just begun
08:05:48 ur: hello
08:05:53 spsneo: ohk
08:06:07 : * spsneo is reading the agenda
08:06:14 Testbaudson: Let's begin with topic 1
08:06:23 Testbaudson: ( 1 ) GSoC Q&A
08:06:38 Testbaudson: Let's start with Len this time...
08:06:45 spsneo: Today I don't have much to discuss.
08:06:56 Testbaudson: Len has sent a motion to the developer list
08:06:57 spsneo: I am clear about my this week's work
08:07:08 spsneo: so lets carry on with len's motion
08:07:52 Testbaudson: LenKne, did you read my comment?
08:07:59 LenKne: I have just sent an email to the dev list in response to Testbaudson comment
08:08:49 Testbaudson: http://lists.osgeo.org/pipermail/mapbender_dev/2008-June/001223.html
08:08:50 sigq: Title: [Mapbender-dev] Motion to enhance the user/group model ( at lists.osgeo.org )
08:09:00 : * Testbaudson is reading
08:09:34 : * ur too
08:11:49 : * Testbaudson is thinking
08:12:31 mandsch: sorry, I#m late
08:12:39 ur: I think to separate services and applications is very difficult
08:12:59 ur: that was the reason to manage service-permissions in "guis"
08:13:27 ur: if you have modules like addVendorSpecific
08:13:56 ur: the target is a service and this service must be available in the application
08:15:16 Testbaudson: so if we don't add "wms_mb_user" and "wms_mb_group"...
08:15:32 Testbaudson: we only cascade read access and not write access...
08:15:45 Testbaudson: so only the original uploader of that WMS can change it?
08:16:34 Testbaudson: what priviliges are implied by owning a service?
08:16:41 Testbaudson: deletion for sure
08:16:47 Testbaudson: what else
08:16:49 Testbaudson: ?
08:17:08 LenKne: edit
08:17:19 LenKne: update capabilities
08:17:46 Testbaudson: ok
08:17:55 Testbaudson: well, it's a restriction...
08:17:55 LenKne: This can make some big changes to an application if the wms changes the layers it has
08:18:44 Testbaudson: but I think that won't stop us, right?
08:19:29 LenKne: An owner of an application could add their own WMS, thus having control of it...
08:20:11 LenKne: this means that the same WMS may be listed mulitiple times, but does give an admin the ability to manage their application
08:21:50 Testbaudson: hmmm....I don't like that
08:22:46 : * Testbaudson is thinking
08:23:13 LenKne: How about a group called service administrator?
08:23:54 LenKne: Members would have access to all WMS configurations.. wait, never mind, this won't work
08:24:08 Testbaudson: so how about this...
08:24:49 Testbaudson: we also cascade write access
08:25:06 Testbaudson: and add mb_role
08:25:50 Testbaudson: wouldn't that be better in the end?
08:26:08 : * Testbaudson is torn
08:26:33 LenKne: Where does mb_role fit in?
08:27:02 Testbaudson: the table wms has a field wms_owner
08:27:06 Testbaudson: where owner is a user
08:27:15 Testbaudson: it cannot be both user and group
08:27:22 Testbaudson: but if it was the id of a role
08:27:35 Testbaudson: it would work
08:27:50 Testbaudson: ur?
08:28:54 ur: I don't see the difference between another/additional id "role" and an additional id "group"...
08:29:35 Testbaudson: mb_role would replace woth group and user
08:29:40 Testbaudson: mb_role would replace both group and user
08:29:50 ur: ok
08:30:00 Testbaudson: like in postgres
08:30:06 ur: and the roles are defined in which way?
08:30:16 Testbaudson: id + type
08:30:28 ur: and the specific permissions?
08:30:28 Testbaudson: where type is "user" or "group"
08:30:49 Testbaudson: just like in the current model
08:31:05 ur: ok, i*ve got it :- )
08:31:42 ur: that should work
08:31:56 ur: we need only one more table ...
08:32:44 Testbaudson: we could *save* tables
08:33:09 ur: ?
08:33:34 Testbaudson: now there are gui_mb_user + gui_mb_group
08:33:43 Testbaudson: then we would end up with only gui_mb_role
08:34:17 Testbaudson: mb_role would replace both mb_user and mb_group
08:34:50 ur: true
08:36:03 Testbaudson: I think it would simplify Mapbender....
08:36:15 Testbaudson: do we want to deal with the necessary changes?
08:36:45 LenKne: Just to make sure I got this right...
08:36:53 LenKne: If I add a users...
08:37:04 LenKne: I would then have to add them to several rols
08:37:27 Testbaudson: no, the interface would remain the same
08:37:28 LenKne: gui_mb_role, wms_mb_role, etc.?
08:37:42 Testbaudson: you would only have another data model
08:38:06 : * Testbaudson hopes he understood what LenKne wanted to ask
08:38:49 LenKne: Wouldn't a user need to be manually added to a role for an application and WMS?
08:39:27 Testbaudson: yes, but it won't change
08:39:34 Testbaudson: the interface remains the same
08:39:39 Testbaudson: only the SQLs change
08:39:56 Testbaudson: instead of inserting it in gui_mb_user...
08:40:02 Testbaudson: you insert into gui_mb_role
08:40:15 Testbaudson: ( there is no more gui_mb_user )
08:41:26 LenKne: How would they have access to the WMS, if wms_mb_role is used?
08:42:13 Testbaudson: at the moment, read access is handled via applications...
08:42:23 Testbaudson: and write access via the column "wms_owner"
08:42:42 Testbaudson: so wms_owner would be the ID of mb_role
08:42:50 Testbaudson: and not of mb_user
08:43:45 LenKne: Ok, so wms_mb_role will only control "write" access to the application...
08:44:07 LenKne: and we keep "read" access as it is, through the application
08:44:28 Testbaudson: yes...by wms_mb_role do you mean the column "wms_owner"?
08:44:39 LenKne: Sorry, that should be write access to the WMS
08:44:55 Testbaudson: of course
08:45:06 LenKne: yes
08:45:58 LenKne: I think this can work well
08:46:11 Testbaudson: ( This change would be a nice challenge for spsneo's update script... )
08:46:25 : * Testbaudson is not afraid of changes...
08:46:31 Testbaudson: how about ur and seven?
08:46:46 Testbaudson: or, mandsch of course
08:46:57 mandsch: ;- )
08:47:07 seven: gui_mb_user would be deprecated?
08:47:17 Testbaudson: yes
08:47:22 Testbaudson: along with mb_user
08:47:25 Testbaudson: mb_group
08:47:30 Testbaudson: gui_mb_group
08:47:32 Testbaudson: ...
08:47:32 seven: so it is sort of a renaming of gui_mb_user to _gui_mb_role?
08:47:38 Testbaudson: yes
08:47:51 Testbaudson: but we would have to add a column "type"
08:47:59 seven: like reaniming gui_* into application_*...
08:48:07 Testbaudson: containg a user or group flag
08:48:37 ur: why are mb_user and mb_group deprecated?
08:48:49 : * seven is confused and suggests that we look at this on a piece of paper in the office tomorrow...?
08:48:51 Testbaudson: replaced by mb_role
08:49:13 : * seven is slow to understand. Once I got it everybody else will also. :- )
08:49:25 Testbaudson: maybe it's just too hot
08:50:00 Testbaudson: so...defer?
08:50:15 Testbaudson: but we should really discuss it tomorrow...
08:50:36 ur: but if we handle only the wms_owner as role
08:50:37 seven: Yes. :- )
08:50:57 ur: we don't need to change the user/group model
08:51:01 ur: wrong?
08:51:14 : * Testbaudson is thinking
08:52:55 Testbaudson: it might work...
08:53:20 Testbaudson: however, you would have to add
08:53:25 Testbaudson: mb_role_mb_user
08:53:36 ur: why
08:53:39 Testbaudson: and replace all user-related foreign keys
08:53:50 ur: if the owner is a role
08:53:59 ur: he could be a user or a group
08:54:29 ur: that*s all
08:54:36 : * Testbaudson was thinking you meant: mb_user and mb_group *extend* mb_role
08:54:50 Testbaudson: if not, we would have 3 entities...
08:55:51 Testbaudson: LenKne: seven is a sponsor of the Mapbender project...
08:56:17 Testbaudson: we should consider what she suggests
08:56:30 : * seven is the mother of Mapbender.
08:56:45 seven: And mothers are protective and emotional and irrational.
08:56:47 seven: :- )
08:57:03 LenKne: Ok
08:57:06 ur: seven is irrational?
08:57:13 seven: well...
08:57:18 Testbaudson: so: deferred until tomorrow
08:57:21 Testbaudson: time?
08:57:26 ur: irrational borgs - aha
08:57:27 seven: once you come in
08:57:31 seven: lol
08:57:34 seven: You got me.
08:58:02 Testbaudson: what time tomorrow?
08:58:16 seven: once you come in
08:58:19 ur: 23:15
08:58:29 seven: We can work out something intelligible and write it down.
08:58:42 Testbaudson: ok
08:58:45 seven: If LenKne does not get it we can cont the chat tomorrow.
08:58:51 seven: LenKne: Ok?
08:59:00 LenKne: Ok
08:59:10 seven: LenKne: If you have questions up front write mail them to Testbaudson
08:59:18 seven: or better to mapbender_dev
08:59:22 Testbaudson: no, to the dev list
08:59:38 Testbaudson: 2 minutes for 3 more topics...
08:59:43 Testbaudson: ( 2 ) Version control, tools ( Meld )
08:59:56 Testbaudson: as each developer will have to maintain its own branch in the future
09:00:04 Testbaudson: we need to do a lot of merging
09:00:12 Testbaudson: but I find it rather complicated...
09:00:22 Testbaudson: with the Subversion merge process...
09:00:29 Testbaudson: maybe it's just me...
09:00:42 Testbaudson: but I have looked around and found a nice tool
09:00:45 Testbaudson: meld
09:01:06 Testbaudson: http://www.mapbender.org/Meld
09:01:06 seven: Testbaudson: it's not "just" you, its' "only" you.
09:01:07 sigq: Title: Meld - MapbenderWiki ( at www.mapbender.org )
09:01:45 Testbaudson: you can easily merge from/to the branch...
09:01:49 Testbaudson: and it has a nice gui
09:01:54 Testbaudson: fyi
09:02:06 Testbaudson: ( 3 ) Bolsena dev sprint
09:02:11 Testbaudson: any news from ticheler?
09:02:51 : * Testbaudson has asked Christian to do some research on GeoNetwork...
09:03:03 Testbaudson: I guess he will have a presentation on it this week
09:03:27 Testbaudson: as at least Gregor and I are not familiar with it
09:03:43 Testbaudson: so no news from ticheler, seven?
09:04:29 seven: Ahm.
09:04:31 seven: Well.
09:04:34 seven: Yes.
09:04:40 seven: You will get the Kardinal Suite!
09:04:41 seven: :- )
09:05:05 seven: And ticheler and I will slepp at the camp ground.
09:05:14 seven: So you can book everything now.
09:05:20 Testbaudson: You're so modest....
09:05:33 Testbaudson: but thanks
09:05:55 Testbaudson: so we will book first thing in the morning
09:05:58 Testbaudson: ( 4 ) new stuff in trunk
09:06:18 Testbaudson: i was planning to add two things to the trunk
09:06:27 Testbaudson: first, the splash screen
09:06:42 Testbaudson: a new feature, which displays a splash screen...
09:06:46 Testbaudson: as long as mapbender loads
09:06:58 Testbaudson: mandsch has also done some work on this, I assume?
09:07:04 mandsch: yep
09:07:15 mandsch: it's working pretty nice
09:07:24 Testbaudson: sweet
09:07:26 mandsch: i will come back with the code you sent me
09:07:35 mandsch: with some suggestions of my side ..
09:07:43 Testbaudson: ok
09:07:47 mandsch: i need to test the include thing ...
09:08:22 Testbaudson: ok
09:08:30 Testbaudson: we will have this new feature in Mapbender 2.6
09:08:41 Testbaudson: another thing I plan to do is this
09:08:50 Testbaudson: we have a lot of files in the http folder...
09:09:07 Testbaudson: that are never ever requested via http
09:09:16 Testbaudson: so I plan to move these files
09:09:38 Testbaudson: ( most are javascript libraries )
09:09:47 Testbaudson: to the folder /lib
09:10:01 Testbaudson: now, they are in the folder /http/javascripts
09:10:08 Testbaudson: example: map.js
09:10:12 Testbaudson: map_obj.js
09:10:15 Testbaudson: point.js
09:10:17 Testbaudson: etc
09:10:45 Testbaudson: I think our goal should be to keep the http folder contents small
09:11:35 Testbaudson: any objections?
09:11:50 Testbaudson: I also splitted map.js in several small js units....
09:11:54 Testbaudson: like ajax.js
09:12:00 Testbaudson: button.js
09:12:04 Testbaudson: etc
09:13:48 Testbaudson: well, so I just proceed...you don't seem to care ;- )
09:13:56 : * Testbaudson motions to close the meeting
09:13:58 Testbaudson: +1
09:14:11 mandsch: second
09:14:14 mandsch: +1
09:14:14 LenKne: +1
09:14:19 ur: +1
09:14:22 seven: +1
09:16:46 mandsch: seven: ur: do still have aminute?
09:17:05 ur: ok
09:17:17 mandsch: tesbaudson left so quick ...
09:17:29 mandsch: i asled seven about sponsoring the treeconf
09:17:48 mandsch: well changes or enhancments to treeconf
09:18:09 seven: mandsch: Tomorrow?
09:18:14 seven: Sorry, I did not follow up.
09:18:23 mandsch: ok, here?
09:18:27 seven: Yes.
09:18:34 seven: in the afternoon.
09:18:40 mandsch: ok, then tomorrow. cu
09:18:57 seven: Thank you. Cu
09:19:00 mandsch: bye