--- Log opened Tue Mar 10 00:00:11 2015
<@xor-freenet> ArneBab_: skydrome: thanks :)00:32

Loewenzahn: there are munin scripts which can be used to graph freenet memory usage in case you want to just measure it yourself00:33

Loewenzahn: https://github.com/Artefact2/munin-freen et

Loewenzahn: i would reduce the memory limit in small steps until it keeps using 80-90% in average00:34

Loewenzahn: also, it would be nice if you could contribute your observations by telling us the result etc :)

Loewenzahn: IIRC, its been a long time since someone has done a scientific adjustment of the default memory limit, so those numbers would be helpful00:35

Loewenzahn: also, it might make sense to wait for the next freenet release (build 1468) which should be out the next weeks, because it changes architecture significantly

"<Loewenzahn> [23:39:23] do people linger here, just for answer questions towards freenet. but in the end using that chatroom is bad too, so why have this" in general development discussion happens on this channel, so you're at the right place00:36

"<Dizzle> [23:55:52] When you insert a CHK using the freenet webapp, does it automatically get the filename/path thing at the end?" the URI is generated after the insert is finished i think, yes. this might change with build01468 though00:43
< Loewenzahn> thanks a lot. you will know as fast as we tried, but till then will pass some time. More than just some weeks.:)00:49
< Dizzle> xor-freenet: thanks, is he filename part documented somewhere? It seems like its necessary/immutable for the key to work.00:50
<@xor-freenet> Dizzle: well it depends on what you want to know from the documentation00:51
Dizzle: freenet keys can contain whole directories, so there can be multiple filenames
Dizzle: if you have a CHK@.../ without a filename, it might be a directory
Dizzle: the web interface doesn't provide an integrated UI to view the content of those directories, but the KeyExplorer plugin can view them00:52
< Dizzle> huh, interesting. Is there a way to tell it to download the whole archive of a known CHK? (i.e. leave off the filename)00:53
<@xor-freenet> Dizzle: no, i think there is not. maybe a tool such as pyFreenet can do that, ArneBab_ might know00:54
Dizzle: but also check KeyExplorer maybe
< Dizzle> alrighty, thanks xor
<@xor-freenet> Dizzle: i think the fact that there is no such UI is actually a usability issue which should be fixed someday, latest when we have a filesharing tool00:55
operhiem1: another candidate for stuff which we should at least clone into the official freenet github account is the freenet munin scripts of Artefact2: https://github.com/Artefact2/munin-freen et they allow graphing of freenet memory usage, network usage, etc.... i've been using them for years, they are nice... don't have a screenshot now since i don't run them at the moment, but munin looks like that: http://upload.wikimedia.org/wikipedia/01:05
commons/c/ce/Munin-memory-week.png
operhiem1: un-wrapped link: http://upload.wikimedia.org/wikipedia/co mmons/c/ce/Munin-memory-week.png
operhiem1: i've filed a bug for this at: https://bugs.freenetproject.org/view.php ?id=647901:11
operhiem1: also, what about thaw now? remember, its author said its OK if we put the repository back up as long as we remove his name from the code. we don't even have to remove it from the commit history. "fgrep -Ri flesch thaw-repository" should find all occurences of his name.....i think it were around 30 occurences.. if you don't feel like dealing with removing his name, maybe mail me the backup of the repository so i can do it01:20
< GitHub156> [plugin-WebOfTrust] xor-freenet pushed 1 new commit to master: http://git.io/plFQ01:50

plugin-WebOfTrust/master ec38ccb xor-freenet: XMLTransformer: Fix import of unwanted identities...
 -!- mode/#freenet [+o infinity0_] by ChanServ05:46
< GitHub34> [plugin-WebOfTrust] xor-freenet pushed 2 new commits to master: http://git.io/p8ih06:04

plugin-WebOfTrust/master 69ab18c xor-freenet: (Own)IdentityTest.testClone(): Fix to show failure of preserving...

plugin-WebOfTrust/master 80c1ab0 xor-freenet: (Own)Identity.clone(): Fix failure of preserving mLastChangedDate...
< GitHub116> [plugin-WebOfTrust] xor-freenet pushed 3 new commits to master: http://git.io/p8FU07:25
plugin-WebOfTrust/master c78d995 xor-freenet: KnownIdentitiesPage: Extract function getPageCount()...
plugin-WebOfTrust/master 9b819e1 xor-freenet: KnownIdentitiesPage: Fix current page not being preserved...
plugin-WebOfTrust/master 0a23080 xor-freenet: KnownIdentitiesPage: Improve accuracy of shown date/time values...
< ArneBab> Dizzle: have a look at keyutils, I think it’s shipped and can be activated on the plugins page (you might have to activate advanced mode)07:44

xor-freenet: I think it would be good to provide KeyUtils outside of advanced mode: They are pretty polished I think.07:46

xor-freenet: wasn’t it nextgens who has a backup repo of thaw?07:47
<@xor-freenet> ArneBab: its on the project servers IIRC, so operhiem1 should have access to it. ng has been very annoyed by me randomly trying to assign work to him, so i would prefer to leave him alone07:48
< Dizzle> Alright. Now is the path element considered part of the key technically? We may want to update the documentation for CHK and others in the wiki, as they don't even mention the path part at the end (as far as I know).
<@xor-freenet> Dizzle: the query for the data on the network is only the CHK@.../, not the path, so technically its something different07:49
Dizzle: at the low level, a CHK@../ is just a 32 KiB binary blob of data. the filesystem aspect of paths / filenames is an additional feature which freenet implements on top of that07:50
< Dizzle> alrighty
<@xor-freenet> Dizzle: basically everything with regards to paths/files happens locally at your machine after freenet has downloaded the CHK@.../
Dizzle: (also, the metadata about the paths/files etc. is split into multiple underlying 32 KiB CHKs at the network level)07:51
< Dizzle> hmm, got a doc on that last one? it confuses me. Each file has its own 32KiB CHK? i.e. there can't be more than one file with the first part of the CHK (before the trailing path) being the same?07:53
<@xor-freenet> Dizzle: imagine a folder of sheets of paper. each sheet is called a "CHK@.../", and each sheet is 32 KiB07:54
< Dizzle> ok
<@xor-freenet> Dizzle: what the file/path stuff does is writing text onto the sheets which says "sheet number #2134 contains file filename-1234.txt"
Dizzle: the sheets themselves don't care whats written onto them
Dizzle: the sheets = the freenet network level
< Dizzle> Is the folder also called a "CHK@.../"?07:55
<@xor-freenet> Dizzle: network layer i mean
Dizzle: basically the "CHK@.../" is the number of a sheet, its address, its key
Dizzle: the "folder" is the whole network
< Dizzle> theoretically, a sheet could "contain" multiple files?07:56
<@xor-freenet> Dizzle: effectively, if you enter an adress such as "CHK@.../filename.html" into freenet, it will download the CHK@.../ without the filename. that CHK then contains a list of many other CHKs which the file consists of.... if the file is 320 KB, then it will be 10 additional CHKs in that list.... (and possibly some more if the file list needs more space)07:57
Dizzle: as long as the metadata and the file fit into a sheet, yes
Dizzle: raw CHKs without a filename are sort of like sectors of a harddisk. the files are built on top of that
< Dizzle> ok, and this file-specific CHK is never exposed to the front end?08:00
<@xor-freenet> Dizzle: typically not, you typically only handle the top-level CHK
Dizzle: keyutils can display the individual low level ones probably
< Dizzle> ok, it's starting to make some sense to me08:02
<@xor-freenet> Dizzle: :)
< TheSeeker> xor-freenet: 20 keys... you forgot checkblocks.08:05
<@xor-freenet> TheSeeker: thanks... i actually have very little knowledge about the client layer filesystem stuff08:06
TheSeeker: well i do know that we have redundancy, but the format of the metadata is a mystery to me08:07
< TheSeeker> The top block metadata contains a list of all the blocks in the file, including checkblocks and cross-segment blocks (if over 80 MiB) ... if that metadata is larger than 32 KB, the metadata itself is inserted as a splitfile, and the top block points to that instead. For extremely large files, there can be multiple layers of redirection.08:13

you'll often also have multi-level metadata if the chk was a directory insert, since the relative paths and filenames are included in the top layer. when you have a lot of files and some of them are larger than 2 MB, things start getting a bit more complicated when doing metadata parsing...08:16

then you start getting into whether the file data is internal to an archive, or is a pointer to an external redirect, or a pointer into a different container.08:17

I still get a bit of a headache when trying to debug keepalive.08:19
< ArneBab> TheSeeker: thanks!
(could you flog your last 4 messages?)
< TheSeeker> I think I'd want to be a lot more precise for that kind of thing.08:21
< ArneBab> activity trumps perfection ☺08:25
* ArneBab often drops blog comments on his site with just minor changesg.
<@xor-freenet> TheSeeker: thanks for clarifiying.
Dizzle: read what TheSeeker said in case you're interested in further details
< TheSeeker> Read the keyutils source if you want far more detail.08:27
< Dizzle> cool cool08:31
also, is toad gone?08:32
<@xor-freenet> Dizzle: toad is currently busy with studying computer science. he didn't formally leave the project, and still responds to mail if we have important questions08:40

Dizzle: from my impression, he hasn't even himself decided yet what to do after studying....08:41
< Dizzle> ahh, cool!
< TheSeeker> hey, we got a mention at least. http://www.dailydot.com/politics/uk-brie fing-tor-child-abuse-minor-role/08:48
< GitHub66> [plugin-WebOfTrust] xor-freenet pushed 1 new commit to master: http://git.io/p4Tc
plugin-WebOfTrust/master a3573a9 xor-freenet: IdentityFetcher/SubscriptionManager: Fix AssertionError...
< TheSeeker> The entirety of the comment in the report being "Freenet is designed as a tool for sharing files anonymously and has been used to distribute censored information. Files shared by users are split into encrypted blocks and stored across the computers of other Freenet users."08:49
<@xor-freenet> progress report for today: 5.5 hours, 7 commits. did more test runs. finished 1 more of the 6 scheduled ones, 2 of 6 are finished now. fixed 5 bugs caught during the test runs, which is also why i only finished 1. but the 5 fixed bugs at least show that the test runs were no waste of time :) the most notable one was that the XMLTransformer would import an identity even if it had a bad score meanwhile, which sort of is a security issue. (only sort09:02
of because it could not affect other people's scores since the score computation algorithm does not rely upon only "good" identities being imported, it can deal with malicious ones as well. this is because whether an identity is good/bad in terms of scores depends upon the view of each OwnIdentity, so the score computation algo must eat bad ones as well. still, a distrusted identity should not be imported since its trust list might contain bloat/
DoS/other malicious stuff). also, while the test runs were executing, i searched my email backlog for WOT stuff and dealt with bugtracker maintenance
also started writing a message for publishing the release candidate / requesting testers.... once the 6 test runs are done, the release candidate will be out.09:06
< ArneBab> TheSeeker: that’s a pretty positive summary ☺10:32

TheSeeker: it’s the gist of the technical underpinnings of Freenet.10:33

xor-freenet: does the importer now discard the bad ID or reset its score?10:35
< n1qn4q> just a little while ago there was someone banned for saying he was trying to get censored books via freenet10:46
how does this fit with the u.k. parliament, also why is this channel doing lawenforcement when there isnt even a hint on what kind of censored books or whatever the guy was talking about
<@operhiem1> "@toad_ | will that be before or after IPv6 is deployed? ;)" -> The things 1468 is blocking on are problems purge-db4o introduced anyway.11:33

xor-freenet: The plugin with login support was https://github.com/Thynix/plugin-DVCS-We bUI11:51
< ArneBab> n1qn4q: lawenforcement is not parliament.11:56
< n1qn4q> ?
< ArneBab> n1qn4q: we can’t follow what parliament thinks about laws but only the laws which exist right now.11:57
n1qn4q: also this channel is not only bound by UK laws, but by the horrible mess of merged copyright restrictions from multiple nations.11:58
n1qn4q: whether you like it or not: If you want Freedom of speech, use Freenet, not the #freenet IRC channel.11:59
<@operhiem1> "this build includes more new commits than WOT *ever* consisted of as of the previous build" doesn't sound like good news to me. When I read that I think "and just IMAGINE all the bugs they introduce!"
< ArneBab> xor-freenet: to add to that: I think the humor in “good news and bad news” will not fall on fertile ground.12:00
<@operhiem1> Dun wanna bundle, in general. (re pyFreenet)

If there's a good reason to bundle we can talk about it. I just don't want to accumulate even more maintenance burden for the base package.12:01

xor-freenet: https://archives.freenetproject.org/gith ub-backups-official/Thaw.git/12:05

n1qn4q: Agreed - they did not say copyright-infringing. I disagree with kicking them until their intent is clearly infringing. (FPI folks are legally obligated to do this.)12:09

Ooh, http://toolkit.translatehouse.org/. There's also http://pootle.translatehouse.org/.12:20

That might be preferable to Transifex because it's open-source / free software instead of a SaaS like Transifex.

Offline translation is http://virtaal.translatehouse.org/12:41

Can someone add me as a developer on the FlogHelper bug tracker please?12:52
< oselotti> operhiem1: Why I'm not anymore in finnish translation team? I used to be a reviewer.13:25
<@operhiem1> oselotti: Because the people who said they'd help with translation have delivered an ongoing fuckup instead.13:27

The project is now under the Open Technology Fund - https://www.transifex.com/organization/o tf/dashboard - but translators did not transfer, and my contact with the organization has gone unresponsive.13:28

(I am no longer a project maintainer as well. They lied to my face on how transfer would work. >:( )13:29

oselotti: Feel free to apply to join the translation team again - I think hernic still has some degree of power in the project?13:30

Sorry about all this. I am upset about it and feel that they lied to me to get the project transferred, then dropped me.
< oselotti> Okay, I thought that I was kicked out from the team.13:31
<@operhiem1> The team did not transfer - they said they'd copy it over but have failed to do so.

I would post an announcement but it doesn't seem like it'd be useful - "raise your hand if you can't hear this" sort of thing.13:32

oselotti: Anything I can help with? Is that enough information?13:35
< ArneBab> operhiem1: can’t you transfer it back?13:40
<@operhiem1> ArneBab: My admin power over the project did not transfer, and they have not added me as an admin again, so no, I can't.13:41
< oselotti> operhiem1: Yes thanks for the info. I was just confused.
<@operhiem1> When I filled out the transfer form, when I submitted it I got a message saying that admin/translation teams would not transfer. I mentioned this to my contact in OTF, who assured me that it would.13:42

-> I think they were telling me whatever I wanted to hear until I transferred the project.

My contact also said they would be responsive to email. They have not responded after several days.13:43

Unless they fix this I'm tempted to publicly post this travesty and try to shame them into either transferring the project back or completing the transfer the way they said it'd work.13:44

(More public than public IRC I mean.)

ArneBab: Any thoughts?13:47

I done goofed, certainly, but having someone lie to my face for prolonged periods is something I don't generally suspect. In retrospect I wish I'd been more cautious.13:48
< ArneBab> I’d warn them first by mail: If you don’t finish the transfer or transfer it back until Wednesday, 23:59 UTC, I will post about this in public.13:51
<@operhiem1> Yes, I will certainly warn them first.13:52
< ArneBab> and if they don’t complete within the deadline, it’s newstime.
I once used pootle for Phex. Can they host the online version for us?13:54
<@operhiem1> Which "they?"13:55
< ArneBab> translatehouse
(selfhosting inflicts a maintenance cost on us)
<@operhiem1> That doesn't look to me like a service they offer.
< ArneBab> that’s sad, because it would be great to have13:58
<@operhiem1> Agreed, but I can also appreciate wanting to limit the scope of their operations.13:59
< ArneBab> yes14:03
< GitHub142> [fred] Thynix created contributing (+1 new commit): http://git.io/pBBA14:26

fred/contributing 4dd0c97 Steve Dougherty: Add contributing guidelines...
<@operhiem1> Thoughts on https://github.com/freenet/fred/pull/338 would be appreciated.14:28
* operhiem1 is off to work.
< ArneBab> operhiem1: you’re already home again?14:36
operhiem1: sounds good, though maybe a bit long14:38
<@nextgens> it's like always, we're going backwards14:49

we got shafted several times from sourceforge before we've decided to run our own hosting14:50

then decided again to "outsource"

whether that's to github, transifex, google-code or elsewhere

won't be the last time we get shafted :)

11:52 <@operhiem1> Can someone add me as a developer on the FlogHelper bug tracker please?14:56

done
< ArneBab> hm, yes…15:05
< Lieke> i always wondered why freenet isn't using freenet to host freenet15:15

infocalypse seems to work fine
< ArneBab> Lieke: it’s less convenient, and I still didn’t get to polishing Infocalypse far enough that it’s more fun to use than github…15:23
< TheSeeker> Freenet does use Freenet to update itself. It's a bit much to ask people to use freenet to get the initial installer though.17:32
< Lieke> i was talking about the source repository and i didn't say it couldn't be mirrored on the clearnet17:45
< TheSeeker> There is an infocalypse (mercurial) source respository in freenet, but I don't know how up to date it is.17:49

USK@3Xvf~RI-N8YC-xFtQIEnEwR~ZLC05UhQjQzK AS0lL4Y,PqTg9Y5Me6~WthiW5FoHG9Mq18c7oRbe 0owRutJJJZc,AQACAAE/fred1.R1/12617:52
< ArneBab> TheSeeker, Lieke: it should be roughly up to date and actually contains additional data not in the main github repo, because I pulled all branches from every fork.17:54
 -!- mode/#freenet [+o infinity0_] by ChanServ19:50

-!- mode/#freenet [+o toad_] by ChanServ20:25