Time |
Nick |
Message |
20:03 |
meeting |
Meeting started Sun Mar 20 20:03:00 2011 UTC. The chair is walterbender. Information about MeetBot at http://wiki.debian.org/MeetBot. |
20:03 |
|
Useful Commands: #action #agreed #help #info #idea #link #topic #endmeeting |
20:03 |
silbe |
garycmartin: thanks, see you next week! |
20:03 |
garycmartin |
Thanks all, sorry I have to dash, will look at the log! |
20:03 |
|
garycmartin has quit IRC |
20:03 |
silbe |
walterbender: can you set the topic, please? |
20:03 |
walterbender |
#info gary had to leave... but we will continue for one more topic |
20:04 |
|
#topic What date should the activities list view show for pre-installed activities? |
20:04 |
silbe |
#link http://dev.laptop.org/ticket/10487 OLPC#10487: Activity Installation time wrong (activity list view) |
20:04 |
|
#link http://lists.laptop.org/piperm[…]thread.html#31544 Patch discussion: Do not adjust dates of activities |
20:05 |
walterbender |
IMHO, the relevant date is the Activity modification date... it tells the user how recent a version they are running. |
20:05 |
|
and if they modify the activity themselves, it highlights that fact too |
20:06 |
silbe |
walterbender: but what is the "modification date" for activities that came with the OS image and were neither started nor modified yet? |
20:06 |
christianmarcsch |
walterbender: +1 |
20:06 |
|
silbe: in that case, wouldn't we use the date of installation? |
20:07 |
silbe |
christianmarcsch: that's exactly the question |
20:07 |
christianmarcsch |
sllbe: i think that would make sense. we'd seed them with the date of installation, and override once they have been instantiated |
20:07 |
walterbender |
silbe: is there no way to include the modification date in the build? |
20:09 |
|
m_anish_afk is now known as m_anish |
20:09 |
silbe |
The options I can think of are: 1. use date of last modification from source tarball, 2. use time of OS image build, 3. use time of first Sugar start-up, 4. use some artificial time in the past. Current OLPC implement the last option AFAIK. |
20:10 |
|
s/AFAIK// (just checked again) |
20:10 |
walterbender |
I prefer #1 |
20:12 |
|
and it would be consistant |
20:13 |
christianmarcsch |
wouldn't #2 make more sense for the user though? |
20:13 |
|
maybe it's not totally consistent internally, but to the user the only thing that matters is when they started using the system |
20:13 |
walterbender |
christianmarcsch: why? |
20:14 |
|
christianmarcsch: I would think that they'd want to know the version of TA they were running was 2-years old on their new release.... |
20:15 |
christianmarcsch |
walterbender: i suppose you are right |
20:15 |
walterbender |
so they could berate the developer to get off the stick... |
20:15 |
christianmarcsch |
walterbender: :) |
20:15 |
|
but |
20:16 |
|
once someone starts using the activity, would it switch to the date of last instantiation? |
20:16 |
walterbender |
christianmarcsch: shouldn't... |
20:16 |
christianmarcsch |
walterbender: ok, then it makes sense |
20:16 |
walterbender |
running an activity doesn't update the activity, just the journal instance |
20:17 |
christianmarcsch |
walterbender: perfect, ok. just needed to understand that. |
20:18 |
walterbender |
silbe: correct me if I am wrong, but we are talking about the date of the activities shown in the list view |
20:18 |
silbe |
walterbender: exactly. List View of the Home View. |
20:19 |
walterbender |
christianmarcsch: this is what determines the order in the circle/spiral or the list in the list view |
20:20 |
christianmarcsch |
walterbender: right, that makes sense |
20:20 |
walterbender |
thinks the list view should show most recent first, because those are the ones you'd likely be working on... |
20:21 |
|
silbe: is there anything re #1 that is controversial (or a headache technically?) |
20:22 |
silbe |
walterbender: not to me :) Check http://dev.laptop.org/ticket/10487 for other opinions |
20:22 |
|
walterbender: the technical side of things is solvable |
20:23 |
walterbender |
silbe: I didn't understand the issue with Gary's side-effect. |
20:23 |
silbe |
we just need to agree on how it _should_ work, and I'll take care of making Sugar _do_ work like it |
20:23 |
walterbender |
silbe: if the activity is touched, presumably it is for a reason |
20:24 |
silbe |
walterbender: I'll try to figure out where and maybe why it's touched. Off the top of my head I don't see a technical reason to do so. |
20:24 |
walterbender |
silbe: seems like there was *almost* consensus on #1 in the ticket (as I recall) |
20:25 |
|
silbe: with a few people wanting some additional info available (which I think is needless complexity) |
20:25 |
silbe |
walterbender: iff we want to provide additional information, we should handle that as a separate topic IMO |
20:26 |
walterbender |
silbe: agreed |
20:26 |
silbe |
(I too think it's too complex for the list view; at this point we can just refer users to the Terminal) |
20:26 |
walterbender |
silbe: re Gary's side effect, as an activity developer, I like it... the things I am working on show up at the start of my spiral. |
20:27 |
silbe |
walterbender: what exactly do you call "Gary's side effect"? That an activity gets sorted in front if you just use it, without modifying it? |
20:28 |
|
walterbender: if you modify the activity, #1 would show it at the top (list view) / inner-most (spiral) |
20:29 |
walterbender |
silbe: we need clarification, but he describes "when you use an activity the folder gets touched" but is this true? |
20:29 |
silbe |
tries |
20:30 |
walterbender |
has never seen that behavior |
20:30 |
silbe |
oh, I had the sort order reversed in my mind it seems: the list view currently shows the oldest activity first |
20:31 |
walterbender |
silbe: yes... which I think is the opposite of what we want... but the oldest is on the outside of the spiral, which I think is correct. |
20:32 |
silbe |
walterbender: confirmed, it does get touched. |
20:32 |
walterbender |
silbe: but running an activity has no impact on the order as far as I know |
20:32 |
|
silbe ??? |
20:32 |
|
satellit_afk is now known as satellit_ |
20:33 |
walterbender |
silbe: is that modification time or access time? |
20:33 |
silbe |
walterbender: it gets even stranger: in the list view, the just-run activity gets sorted somewhere in the middle (after a Sugar restart) |
20:34 |
|
walterbender: ls uses mtime by default |
20:34 |
|
walterbender: (i.e. mtime gets changed) |
20:34 |
walterbender |
silbe: hmm. I cannot say I ever noticed that before. It is a bug. |
20:35 |
|
why does running an activity modify the directory? |
20:35 |
silbe |
walterbender: I guess we should just agree on how it's supposed to behave and fix Sugar so it does behave the newly specified way. |
20:35 |
|
The current behaviour definitely is a bug. I can't imagine any spec saying "sort a recently-run activity in the middle" ;) |
20:35 |
walterbender |
silbe: it makes no sense for activities to be modified when they are run |
20:35 |
silbe |
walterbender: +1 |
20:35 |
satellit_ |
any ideas about Give priority to last installed activity by date installed and not version number? |
20:36 |
silbe |
satellit_: the activities list view doesn't look like it's sorted by version number |
20:37 |
walterbender |
satellit_: how do you have both at once? two instances of the same activity? Sugar doesn't support that, does it? |
20:37 |
satellit_ |
but browse-120 from updates precludes running browse-119.xo installed later |
20:37 |
walterbender |
satellit_: does it prevent running or installation? |
20:37 |
silbe |
we sort it by internal order of the bundle registry (I'm sure about that part, filed a bug about it some months ago). IIRC the internal order of the bundle registry is based on the activity mtime... |
20:38 |
satellit_ |
so an update may install a non-fuctioning non removable activity |
20:38 |
silbe |
satellit_: that's a separate issue. |
20:38 |
satellit_ |
prevents running .xo of earlier version |
20:38 |
walterbender |
satellit_: yeah... but that is a different topic, I think |
20:38 |
satellit_ |
ok |
20:38 |
silbe |
satellit_: we should talk about that some time as well, but not now. |
20:39 |
walterbender |
silbe: do we have consensus re modification time and that running an activity shouldn't modify the activity??? |
20:39 |
|
silbe: any idea why the latter is happening? |
20:39 |
silbe |
walterbender: +1 from me. christianmarcsch? |
20:39 |
|
walterbender: not yet, but I'll figure it out. |
20:39 |
christianmarcsch |
silbe: +1 |
20:39 |
walterbender |
looks to see what was touched in Pippy |
20:40 |
silbe |
ok, nice. Can we agree on it? Gary can still chime in via email and we can reconsider if necessary. |
20:40 |
walterbender |
silbe: ah.. it is the .pyc files of course |
20:41 |
|
wonders if there is a way to ignore those. |
20:42 |
silbe |
walterbender: ah, I see. There are ways to either avoid that or at least ignore it, though. |
20:42 |
walterbender |
silbe: I think we should ignore them... it is misleading |
20:42 |
|
wonders how, without having to walk through the entire tree |
20:43 |
silbe |
do we have consensus on how it should work? |
20:44 |
walterbender |
yes |
20:45 |
|
silbe: I will update the ticket with a pointer to this discussion. |
20:45 |
silbe |
walterbender: thanks. Can you add a #agree <description of consensus> so it turns up in the minutes, please? (#agreed is restricted to the chair) |
20:46 |
|
s/#agree/#agreed/ |
20:46 |
walterbender |
silbe: or not, because I cannot write to tickets at OLPC anymore :( |
20:46 |
|
#agreed that modification date is the proper date to list and sort on |
20:46 |
silbe |
walterbender: you should have received a token via email when you logged in. |
20:47 |
christianmarcsch |
walterbender, silbe: i will need to leave soon--were there any other issues you wanted to go over? |
20:47 |
silbe |
walterbender: they recently enabled email address verification to counter spam |
20:47 |
walterbender |
#agreed that we should fix the "bug" that considers .pyc file generation a modification to an activity |
20:47 |
|
silbe: yes... and I never get the verification in my email... |
20:47 |
silbe |
christianmarcsch: not for today. We already took much more time than I wanted. |
20:47 |
|
walterbender: :-/ spam folder? |
20:48 |
walterbender |
christianmarcsch, silbe: next week, same time... |
20:48 |
|
looked there... will look again. |
20:48 |
christianmarcsch |
sounds great. see you next week! |
20:48 |
silbe |
walterbender: thx! (for the #agreed's) |
20:48 |
|
christianmarcsch has quit IRC |
20:48 |
silbe |
thanks everybody for joining, see you next week! |
20:48 |
walterbender |
silbe: not in my spam folder :P |
20:48 |
|
OK. CU |
20:48 |
|
#end-meeting |
20:48 |
meeting |
Meeting ended Sun Mar 20 20:48:49 2011 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot. (v 0.1.4) |
20:48 |
|
Minutes: http://meeting.sugarlabs.org/s[…]-20T20:03:00.html |
20:48 |
|
Log: http://meeting.sugarlabs.org/s[…]11-03-20T20:03:00 |