| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |
| Who | What | When |
|---|---|---|
| mchehab | @all: I'm currently suffering some e-mail problems... probably e-mails sent to me over at least the last 8 hours didn't arrive
I re-routed my e-mails to a new server, so I'll likely be able to receive stuff again but the ones already sent may have issues | [11:57] |
| ................................................................................. (idle for 6h44mn) | ||
| pinchartl | mchehab: I can't help noticing you keep adding new old/cs_* tags to the linux media main repository. what's going wrong ? | [18:42] |
| mchehab | in this specific case, just a rebase at the fixes branch | [18:50] |
| pinchartl | we should stop rebasing that tree... | [18:56] |
| .... (idle for 15mn) | ||
| mchehab | sfr once asked me to rebase on every new Kernel version
I don't do that on master but from time to time I end needing to do on fixes - specially when we have topic branches to merge like this time After doing that, I realized that I could probably have handled it on a different way anyway, the CoC may actually force us to do rebases (we actually had yesterday a similar request at a ML archive) someone invoking the right to be forget | [19:11] |
| pinchartl | no it won't, let's stop spreading FUD | [19:14] |
| mchehab | it is not FUD. That's the text of the e-mail we received yesterday:
Dear Sir or Madam Right to be forgotten request Hello I'd like to have my data/posts removed from your servers. If you need any more data from me, please let me know as soon as possible. It may be helpful for you to know that data protection law requires you to respond to a request within one calendar month. If you do not normally deal with these requests, please pass this letter to your Data Protection Officer, or relevant staff member. | [19:18] |
| pinchartl | I meant the fact that we would be forced to rebase. the CoC will not force rebases
which mailing list was that sent to BTW ? | [19:19] |
| mchehab | linux-dvb ML
it was an e-mail sent back in 2008 (btw, I didn't see anything there that would justify such request - maybe that's just my eyes) | [19:29] |
| pinchartl | ah, that's why I couldn't find it when looking at the e-mails from yesterday :-) | [19:30] |
| mchehab | it was sent to ML admins | [19:31] |
| pinchartl | I'm not sure I would have complied right away... | [19:31] |
| mchehab | nothing happened yet
IMHO, it takes a lot less time to just remove a single e-mail than to do anything else (that would likely require lawyering-up) | [19:32] |
| pinchartl | in any case, we should minimize rebase operations in the main tree | [19:33] |
| mchehab | agreed | [19:33] |
| pinchartl | if you fear breaking other trees on git.linuxtv.org, how about creating a separate tree with the orphan commits, and referencing it in the alternates ? | [19:34] |
| mchehab | may work... the alternates should be "propagated" to all other trees
(or we should do something different - right now, once a month a script "compacts" all trees, in order to save disk space at the server) | [19:37] |
| pinchartl | the same way that the main tree is referenced in the alternates
ah I thought other trees were using alternates to reference the main tree, isn't that the case ? | [19:38] |
| mchehab | yes it is
and we do a git gc once a month | [19:38] |
| pinchartl | ok | [19:39] |
| mchehab | --aggressive | [19:39] |
| pinchartl | one option would be to pull the missing commits in the trees that use them, remove them from master, and then run git gc to remove them from the clones if they're not needed | [19:39] |
| mchehab | if a tree doesn't find a ref, it breaks after git gc --aggressive
good idea may not work, trough it will notice the alternates and git pull won't actually pick the objects | [19:39] |
| pinchartl | I'm sure there's a way to force the copy
not sure how though :-) | [19:43] |
| mchehab | removing the alternates first
(it might have other ways) it does have ways to retrive a single mising object missing it is painful to use it, though because it would need to re-check with git fsck and get the next missine one(s) a good solution would be to have a machine with enough disk to avoid that :-p | [19:46] |
| pinchartl | :-)
or how about this we could have a clone of Linus' tree referenced by all repositories | [19:49] |
| mchehab | that's what we have
the media-tree doesn't have all objects | [19:50] |
| pinchartl | so we already have two alternates, Linus and media_tree ? | [19:50] |
| mchehab | just the ones that aren't at Linus tree
yep | [19:50] |
| pinchartl | would we consume so much extra disk space if we removed the media_tree alternate, as git gc runs monthly ? | [19:51] |
| mchehab | no idea
when this was setup, we were moving from mercurial, with required a lot less disk than git as the trees had just media drivers on it (it was similar to media_build tree with the drivers on it) | [19:52] |
| pinchartl | yes, I remember
I was already there :-) | [19:53] |
| mchehab | ah, ok!
(don't remember the exact time you joined the project) we had to optimize it a lot, as we had to keep the HQ trees for a while and needed enough space for git trees so, the double-alternates model was added since the beginning with that, each individual tree was a way smaller than the mercurial one (ony a few MB) the gc --agressive was added in order to avoid making us run out of space (yet, we had one or two such events) | [19:53] |
| pinchartl | my first contribution to V4L2 was around 2003 if I remember correctly
the uvcvideo driver started at the beginning of January 2006 before that I worked on the Matrox DC10 driver that's actually likely before 2000 can't remember when exactly though | [19:57] |
| mchehab | I guess you only got an account when we merged uvcvideo, but I may be wrong.
-ELONGTIMEAGO | [20:02] |
| pinchartl | oh yes, I had no account before that
and uvcvideo was developed out of tree for a few years | [20:02] |
| mchehab | I would guess that you got an account near 2007
and we migrated to git about the same time 2007 or 2008 | [20:03] |
| ↑back Search ←Prev date Next date→ Show only urls | (Click on time to select a line by its url) |