Displaying #traffic-server/2015-11-11.log:

Wed Nov 11 00:11:21 2015  danielxu:Joined the channel
Wed Nov 11 00:41:31 2015  _klk_:Joined the channel
Wed Nov 11 01:26:34 2015  sudheerv:briang: ping
Wed Nov 11 01:26:35 2015  sudheerv:u t
Wed Nov 11 01:42:22 2015  briang:hi sudheerv
Wed Nov 11 01:43:23 2015  sudheerv:briang can you review
Wed Nov 11 01:43:25 2015  sudheerv:TS-4001
Wed Nov 11 01:43:31 2015  sudheerv:TS-4002
Wed Nov 11 01:43:33 2015  sudheerv:ugh
Wed Nov 11 01:43:40 2015  sudheerv:TS-4010
Wed Nov 11 01:43:41 2015  sudheerv::)
Wed Nov 11 01:58:16 2015  sudheerv:zwoop: ping
Wed Nov 11 02:25:09 2015  zwoop:sudheerv pong
Wed Nov 11 02:25:30 2015  sudheerv:zwoop: question about aborted transactions
Wed Nov 11 02:26:14 2015  sudheerv:if there are a large number of requests being served with read-while-write (on a cache miss) and first client that initiated the origin request aborts in the middle of download
Wed Nov 11 02:26:26 2015  sudheerv:will the bg fill help to continue the other requests?
Wed Nov 11 02:26:32 2015  sudheerv:or are all of them aborted as well?
Wed Nov 11 02:26:46 2015  zwoop:that's the intent
Wed Nov 11 02:27:01 2015  zwoop:thats wgt BG fill is required for read-while-write to kick in
Wed Nov 11 02:27:02 2015  sudheerv:ok, cool
Wed Nov 11 02:27:11 2015  sudheerv:makes sense
Wed Nov 11 02:28:07 2015  sudheerv:if the server transaction itself aborts (say, due to outbound transaction active timeout), all the requests would get aborted as well?
Wed Nov 11 02:28:23 2015  bcall:as I remember there is a percentage that you have to go over for BG to kick in
Wed Nov 11 02:28:29 2015  bcall:but that is configurable
Wed Nov 11 02:28:39 2015  sudheerv:bcall: yes, we set it to always trigger (meaning 0.0%)
Wed Nov 11 02:28:57 2015  bcall:yeah, I am would assume they get aborted
Wed Nov 11 02:29:09 2015  bcall:I would
Wed Nov 11 02:29:30 2015  sudheerv:yeah, i wonder if bg fill should be allowed to continue in that scenario
Wed Nov 11 02:29:41 2015  sudheerv:not sure how easy it'd be though
Wed Nov 11 02:29:51 2015  thumbs:Joined the channel
Wed Nov 11 02:29:56 2015  es:Joined the channel
Wed Nov 11 02:31:38 2015  bcall:it would have to start another request in its place and do a range request - also you would have to make sure it is from the same object on the origin
Wed Nov 11 02:31:46 2015  bcall:kinda tricky
Wed Nov 11 02:31:46 2015  briang:sudheerv yah I'll look at it for you
Wed Nov 11 02:31:52 2015  sudheerv:briang: cool thanks :)
Wed Nov 11 02:31:55 2015  sudheerv:bcall: umm, why so?
Wed Nov 11 02:32:12 2015  sudheerv:i was thinking of simply aborting the ua side of the transaction and continue filling the cache via bg fill
Wed Nov 11 02:32:16 2015  sudheerv:just like it was a ua_abort
Wed Nov 11 02:32:20 2015  briang:sudheerv, 4001 or 4002?
Wed Nov 11 02:32:22 2015  briang:or both?
Wed Nov 11 02:32:25 2015  sudheerv:4010 :)
Wed Nov 11 02:32:27 2015  sudheerv:TS-4010
Wed Nov 11 02:32:35 2015  sudheerv:your baby!
Wed Nov 11 02:32:46 2015  zwoop:bcall it has to be set to 0% for read-while-write to work
Wed Nov 11 02:32:55 2015  bcall:yeah
Wed Nov 11 02:32:58 2015  zwoop:basically, it has to always be allowed to BG fill, otherwise we don't allowe read-whilew-rite
Wed Nov 11 02:33:05 2015  sudheerv:yeah
Wed Nov 11 02:33:27 2015  sudheerv:i am now wondering if bg fill can be made to work even for transaction active timeout on the outbound side too :)
Wed Nov 11 02:33:40 2015  sudheerv:it certainly works (or should work) for client side abort
Wed Nov 11 02:33:41 2015  bcall:sudheerv: I thought you were talking about if the origin server aborts if it could continue
Wed Nov 11 02:33:53 2015  sudheerv:bcall: yeah, that'd be slighlty harder
Wed Nov 11 02:34:19 2015  sudheerv:or may be not even possible ;)
Wed Nov 11 02:34:23 2015  bcall::)
Wed Nov 11 02:34:40 2015  bcall:anything is possible it is just how much code you want to write for it :)
Wed Nov 11 02:34:44 2015  sudheerv:yeah, i was thinking about the transaction active timer (outbound) side
Wed Nov 11 02:34:49 2015  sudheerv:lol, true
Wed Nov 11 02:35:45 2015  bcall:sudheerv: is your active timeout an issue?
Wed Nov 11 02:36:10 2015  sudheerv:bcall: yeah....
Wed Nov 11 02:36:22 2015  sudheerv:so, for live streaming, we set the transaction out timer to 8 sec
Wed Nov 11 02:36:55 2015  sudheerv:basically to not let any request last longer than 8 sec (didn't want to do it based on the ua side timer, since that can not be overridden etc)
Wed Nov 11 02:37:27 2015  sudheerv:but, then, all the concurrent requests that are latching on to a given request, will be impacted if the download takes longer than 8 sec for that request
Wed Nov 11 02:37:35 2015  sudheerv:the concurrent requests could be very new..
Wed Nov 11 02:37:42 2015  sudheerv:just 1 sec old for example
Wed Nov 11 02:38:04 2015  sudheerv:bg fill (if we can make it work even in this case) could help
Wed Nov 11 02:40:05 2015  bcall:zwoop: ATS compiles on Fedora 23
Wed Nov 11 02:40:27 2015  zwoop:woot!
Wed Nov 11 02:40:43 2015  zwoop:it compiles on Ubuntu 15.10 too (I setup a VM for that on Rackspace)
Wed Nov 11 02:40:44 2015  bcall:always a concern when I upgrade - wondering if I will have to fix things before I can do work
Wed Nov 11 02:40:57 2015  PSUdaemon:+1
Wed Nov 11 02:41:05 2015  PSUdaemon:played with dnf
Wed Nov 11 02:41:12 2015  PSUdaemon:as that is the upgrade path now to 23
Wed Nov 11 02:41:20 2015  PSUdaemon:but wanted to wait until after the summit to upgrade
Wed Nov 11 02:41:24 2015  PSUdaemon:lest i lose my dev machine
Wed Nov 11 02:42:38 2015  bcall:I have a backup box I upgrade first as a guinea pig
Wed Nov 11 02:42:54 2015  bcall:and if I can get it to zfs I upgrade my main box
Wed Nov 11 02:43:27 2015  bcall:zfs is a pain with linux to get it working sometimes - but it is worth the trouble
Wed Nov 11 02:44:23 2015  zwoop:PSUdaemon dnf is just yum. I think the "new" upgrade tool you are thinking of is fedup
Wed Nov 11 02:44:26 2015  zwoop:?
Wed Nov 11 02:44:42 2015  PSUdaemon:no
Wed Nov 11 02:44:47 2015  PSUdaemon:fedup is gone
Wed Nov 11 02:45:01 2015  PSUdaemon:https://fedoraproject.org/wiki/DNF_system_upgrade
Wed Nov 11 02:45:06 2015  PSUdaemon:you have to use dnf now
Wed Nov 11 02:45:12 2015  zwoop:you've always been able to use yum to update
Wed Nov 11 02:45:17 2015  bcall:yep
Wed Nov 11 02:45:18 2015  zwoop:but what they recommend is to use fedup
Wed Nov 11 02:45:34 2015  PSUdaemon:i don't think you can use yum this time
Wed Nov 11 02:45:39 2015  zwoop:yum doesn't exist
Wed Nov 11 02:45:40 2015  bcall:they recommend dnf now
Wed Nov 11 02:45:45 2015  zwoop:really
Wed Nov 11 02:45:54 2015  bcall:yep
Wed Nov 11 02:45:54 2015  PSUdaemon:yeah
Wed Nov 11 02:45:55 2015  zwoop:well, dnf is just yum in disguise :)
Wed Nov 11 02:45:55 2015  PSUdaemon:all dnf
Wed Nov 11 02:46:04 2015  PSUdaemon:it's more ike yum-ng
Wed Nov 11 02:46:06 2015  zwoop:well, that's good (one path)
Wed Nov 11 02:46:11 2015  bcall:https://fedoraproject.org/wiki/DNF_system_upgrade
Wed Nov 11 02:46:20 2015  bcall:it was pretty easy
Wed Nov 11 02:46:21 2015  PSUdaemon:good link
Wed Nov 11 02:46:44 2015  bcall:4 commands for a simple upgrade
Wed Nov 11 02:46:49 2015  bcall:a lot more for zfs
Wed Nov 11 02:47:10 2015  zwoop:I bricked one of my boxes using "yum update" :-/
Wed Nov 11 02:47:17 2015  PSUdaemon:i just hate messing with a box i have no access to
Wed Nov 11 02:47:21 2015  PSUdaemon:or limited access
Wed Nov 11 02:47:29 2015  zwoop:ya, me too
Wed Nov 11 02:47:30 2015  PSUdaemon:and don't want to deal with it before the summit
Wed Nov 11 02:47:43 2015  PSUdaemon:but right after the summit i am upgrading
Wed Nov 11 02:48:13 2015  bcall:my Dell Monitor crapped out on my today - it is a QHD monitor - they wanted to ship me a 1080p monitor because they don't make my model anymore
Wed Nov 11 02:49:16 2015  bcall:they didn't even tell my I was getting a 1080p until I looked at the paperwork - wonder how many times that works on people
Wed Nov 11 02:50:06 2015  bcall:and their tech support is worthless
Wed Nov 11 02:51:19 2015  zwoop:gonna try installing F23 on our RAX setup
Wed Nov 11 03:00:59 2015  thumbs:Joined the channel
Wed Nov 11 04:50:33 2015  jpeach:bcall: why did you do FIPS?
Wed Nov 11 05:08:46 2015  _klk_:Joined the channel
Wed Nov 11 06:06:32 2015  reveller1:Joined the channel
Wed Nov 11 09:12:46 2015  lrea:Joined the channel
Wed Nov 11 09:14:29 2015  Lethalman:Joined the channel
Wed Nov 11 13:24:57 2015  bahumbug:Joined the channel
Wed Nov 11 13:30:58 2015  JSeymour:Joined the channel
Wed Nov 11 13:57:15 2015  shinrich1:Joined the channel
Wed Nov 11 14:13:25 2015  reveller:Joined the channel
Wed Nov 11 14:13:33 2015  reveller:Left the channel
Wed Nov 11 14:15:57 2015  JSeymour:Joined the channel
Wed Nov 11 14:20:35 2015  reveller:Joined the channel
Wed Nov 11 14:39:37 2015  esproul:Joined the channel
Wed Nov 11 15:42:29 2015  felicity:is there a way to stop TS writing to records.config?
Wed Nov 11 15:45:33 2015  sudheerv:felicity: traffic_manage is the one that writes to it, if you are running traffic_server alone, records.config should not be modified
Wed Nov 11 15:47:41 2015  felicity:hmm, i didn't know you could start traffic_server on its own
Wed Nov 11 15:48:18 2015  sudheerv:lol, you could, but, not recommended :)
Wed Nov 11 15:48:18 2015  felicity:that means traffic_line -b won't work, right?
Wed Nov 11 15:48:20 2015  sudheerv:yeah
Wed Nov 11 15:48:28 2015  sudheerv:no management port/access anymore
Wed Nov 11 15:48:40 2015  sudheerv:you will still have the proxy function, but, no heartbeats, management access
Wed Nov 11 15:49:07 2015  shinrich1:I think this changed recently. Checkout TS-3884
Wed Nov 11 15:49:35 2015  shinrich1:Oh nevermind. Just identified requirements and closed as dup
Wed Nov 11 15:49:51 2015  shinrich1:TS-315 is still there
Wed Nov 11 15:50:02 2015  felicity:TS-315
Wed Nov 11 15:51:48 2015  sudheerv:interesting…bcall seems to have added proxy.config.disable_configuration_modification
Wed Nov 11 15:51:58 2015  sudheerv:not sure why the ticket is still open though
Wed Nov 11 15:52:07 2015  felicity:aha, looks like it's implemented in master, perhaps i can pull that into our local 6.0 branch
Wed Nov 11 15:52:14 2015  sudheerv:oh, it's recent
Wed Nov 11 15:52:18 2015  sudheerv:yeah, sounds good
Wed Nov 11 15:52:34 2015  felicity:sudheerv: i guess the last comment means the patch is incomplete
Wed Nov 11 15:52:38 2015  sudheerv:felicity: umm..may be not
Wed Nov 11 15:52:39 2015  sudheerv:yeah
Wed Nov 11 15:52:44 2015  sudheerv:https://issues.apache.org/jira/secure/ViewProfile.jspa?name=jpeach%40apache.org would like the ability to reload configuration files, so I will need to modify the change.
Wed Nov 11 15:52:52 2015  felicity:that's okay, we never reload the config, we use traffic_line -b
Wed Nov 11 15:52:56 2015  felicity:assuming 'reload the config' means -x
Wed Nov 11 15:53:15 2015  sudheerv:hmm..but, then you will be running on a possible throw-away patch
Wed Nov 11 15:53:39 2015  sudheerv:as long as you follow up and replace the patch with the newer version, should be okay i suppose
Wed Nov 11 15:53:41 2015  felicity:if there's another commit i can just revert it and apply the new one
Wed Nov 11 15:54:29 2015  sudheerv:yeah, but, not sure if that patch will be equivalent in the functionality you are looking for
Wed Nov 11 15:54:40 2015  sudheerv:in any case, i think in this specific case, it seems okay
Wed Nov 11 15:54:53 2015  sudheerv:the open item doesn't seem to contradict the current patch
Wed Nov 11 15:55:00 2015  sudheerv:it seems to ask for more functionality, afaict
Wed Nov 11 16:13:31 2015  ben3:Joined the channel
Wed Nov 11 16:16:12 2015  [Jok]_:Joined the channel
Wed Nov 11 16:18:25 2015  \ask:Joined the channel
Wed Nov 11 16:20:09 2015  felicity:https://github.com/torchbox/trafficserver/blob/6.0T/LAYOUT "Traffic Server 3.x" :)
Wed Nov 11 16:20:16 2015  felicity:maybe that file should have the version number removed
Wed Nov 11 16:20:33 2015  sudheerv:lol
Wed Nov 11 16:25:44 2015  blattj:Joined the channel
Wed Nov 11 16:26:01 2015  bcall:sudheerv: I left it open because I need to make an updated to TS-315 - there is a commend at the end of the ticket
Wed Nov 11 16:26:09 2015  bcall:comment
Wed Nov 11 16:26:20 2015  sudheerv:bcall: yeah, saw that, thanks :)
Wed Nov 11 16:26:31 2015  sudheerv:i think felicity wanted to use that patch in its current state
Wed Nov 11 16:26:40 2015  sudheerv:i hope that'd be *safe*?
Wed Nov 11 16:26:50 2015  bcall:yes, it is
Wed Nov 11 16:27:05 2015  sudheerv:cool..
Wed Nov 11 16:27:25 2015  bcall:but you can't do traffic_line -x
Wed Nov 11 16:27:48 2015  bcall:that is what jpeach wanted
Wed Nov 11 16:28:12 2015  _iwc:Joined the channel
Wed Nov 11 16:28:56 2015  sudheerv:oh ok - i think felicity says, she doesn't need that part
Wed Nov 11 16:34:26 2015  lrea:Left the channel
Wed Nov 11 16:50:04 2015  es:Joined the channel
Wed Nov 11 16:58:48 2015  es1:Joined the channel
Wed Nov 11 17:09:07 2015  gancho:Joined the channel
Wed Nov 11 17:29:30 2015  niq:Joined the channel
Wed Nov 11 17:29:30 2015  niq:Joined the channel
Wed Nov 11 18:20:46 2015  maskit:hi, if malcious HTTP header causes an additional request to an origin server, would it be a security issue?
Wed Nov 11 18:22:12 2015  felicity|mobile:Joined the channel
Wed Nov 11 18:22:42 2015  felicity|mobile:Is anyone using stale-while-revalidate with 6.0? I remember someone here a couple of days ago reporting crashe
Wed Nov 11 18:25:37 2015  zwoop:maskit depends on what it is, email security@trafficserver.apache.org with the details please
Wed Nov 11 18:27:59 2015  maskit:got it. I will send it soon.
Wed Nov 11 18:29:59 2015  zwoop:thanks
Wed Nov 11 18:32:49 2015  blattj:Joined the channel
Wed Nov 11 18:34:06 2015  maskit:I've just sent the email.
Wed Nov 11 18:34:50 2015  felicity|mobile:speaking of security, is there an advance notification list for TS security releases?
Wed Nov 11 18:38:24 2015  sudheerv:zwoop: what are the two kinds of overridable abilities
Wed Nov 11 18:38:28 2015  sudheerv:per-transaction
Wed Nov 11 18:38:33 2015  sudheerv:and the second one per-remap
Wed Nov 11 18:38:41 2015  sudheerv:what's the correct term for the second one?
Wed Nov 11 18:39:56 2015  zwoop:there's no difference
Wed Nov 11 18:40:03 2015  zwoop:there's only one "overridable" ability
Wed Nov 11 18:40:12 2015  sudheerv:umm…
Wed Nov 11 18:40:13 2015  zwoop:either it's overridable per transaction, or it's not
Wed Nov 11 18:40:28 2015  sudheerv:the per-remap ones should all be able to be overridable per txn?
Wed Nov 11 18:40:52 2015  zwoop:what is "per remap" ?
Wed Nov 11 18:40:57 2015  sudheerv:like per origin
Wed Nov 11 18:41:03 2015  zwoop:there is no such thing
Wed Nov 11 18:41:09 2015  sudheerv:umm..
Wed Nov 11 18:41:21 2015  zwoop:you use a plugin in a remap rule, that uses the per transaction overridable API
Wed Nov 11 18:41:24 2015  sudheerv:for example, max_number_of_origin_connections
Wed Nov 11 18:41:36 2015  sudheerv:right, i understand that
Wed Nov 11 18:41:52 2015  sudheerv:but, isn't there a restriction based on the config whether it makes sense to allow per txn vs per origin?
Wed Nov 11 18:42:31 2015  zwoop:not sure I understand the question, if you need that, you'd need to keep track of it yourself
Wed Nov 11 18:42:58 2015  sudheerv:umm..sorry, i'm not asking it clearly
Wed Nov 11 18:43:12 2015  bcall:they are all per transaction just depends on how they are applied based on configuration
Wed Nov 11 18:43:15 2015  zwoop:no :)
Wed Nov 11 18:43:16 2015  sudheerv:not all the configs that are part of Overridable params can be overriden per txn
Wed Nov 11 18:43:17 2015  sudheerv:?
Wed Nov 11 18:43:19 2015  zwoop:bcall yeah
Wed Nov 11 18:43:22 2015  sudheerv:yes
Wed Nov 11 18:43:28 2015  sudheerv:i guess, that's my question :)
Wed Nov 11 18:43:38 2015  zwoop:what is "Overridable params" ?
Wed Nov 11 18:43:44 2015  sudheerv:the overridable param struct
Wed Nov 11 18:43:53 2015  sudheerv:the definition in apidefs
Wed Nov 11 18:43:54 2015  zwoop:if it's in that struct, it's overridable per transaction
Wed Nov 11 18:44:07 2015  sudheerv:well, it won't work depending on the point at which you apply that setting, no?
Wed Nov 11 18:44:13 2015  zwoop:yeah, of course
Wed Nov 11 18:44:15 2015  sudheerv:so, functionally, that's to me sounds like not overridable?
Wed Nov 11 18:44:23 2015  sudheerv::)
Wed Nov 11 18:44:28 2015  zwoop:you expect it to go back in time? :)
Wed Nov 11 18:44:38 2015  sudheerv:no, i'm saying, just bcoz a config is in that struct
Wed Nov 11 18:44:48 2015  sudheerv:doesn't guarantee that it can be *changed* per txn
Wed Nov 11 18:44:54 2015  zwoop:so, something uses the configuration in read-request-header hook, and you change the parameter in read-response-header-hook, yet expect that change to apply in read-request-header-hook ?
Wed Nov 11 18:44:54 2015  sudheerv:right?
Wed Nov 11 18:44:55 2015  felicity|mobile:sudheerv: but this applies to any stage of processing, not pre/post remap
Wed Nov 11 18:45:17 2015  sudheerv:felicity: umm..well, there's a small difference
Wed Nov 11 18:45:22 2015  zwoop:sudheerv I think you are saying, you have to be aware when a config is used, and when you set it
Wed Nov 11 18:45:29 2015  zwoop:if so, then yes, you ahve to be aware of that
Wed Nov 11 18:45:33 2015  sudheerv:the ones in per-remap (per-origin), you are aware of the change at TXN_START
Wed Nov 11 18:45:38 2015  felicity|mobile:read headers, remap, send request, read response... at each stage there are more configs you can't override
Wed Nov 11 18:45:43 2015  sudheerv:whereas, plugins can't override everything at TXN_START
Wed Nov 11 18:46:05 2015  zwoop:I have no idea what the question / comments are
Wed Nov 11 18:46:10 2015  sudheerv:or may be not..even conf_remap hooks at something
Wed Nov 11 18:46:10 2015  zwoop:what is per-remap / per-origin
Wed Nov 11 18:46:13 2015  sudheerv:yeah, nm..
Wed Nov 11 18:46:22 2015  zwoop: :)
Wed Nov 11 18:46:24 2015  sudheerv:stupid question, sorry
Wed Nov 11 18:46:47 2015  zwoop:well, it is a valid point that you have to know when it is too late to change an overridable config
Wed Nov 11 18:46:52 2015  zwoop:worthwhile pointing out in the docs probably
Wed Nov 11 18:47:23 2015  sudheerv:yeah..you are only saying that to make me feel better :)
Wed Nov 11 18:47:58 2015  zwoop:yeah
Wed Nov 11 18:47:59 2015  zwoop: ;-P
Wed Nov 11 18:48:16 2015  sudheerv::)
Wed Nov 11 18:48:17 2015  zwoop:there are no stupid questions, only stuipid answers from zwoop
Wed Nov 11 18:48:21 2015  sudheerv:thanks, it worked :)
Wed Nov 11 18:50:54 2015  jpeach:or maybe the SetOverrideable() API could know and return an error if it is too late?
Wed Nov 11 18:51:22 2015  felicity|mobile:or abort, in debug mode?
Wed Nov 11 18:51:36 2015  felicity|mobile:because it's a programming error, not a runtime error
Wed Nov 11 18:51:40 2015  jpeach:also helpful
Wed Nov 11 18:53:42 2015  zwoop:jpeach that's an idea, a lot of leg work though, but technically just adding some more info into the existing structs
Wed Nov 11 18:54:18 2015  zwoop:jpeach i.e. someone would have to analyze every configuration, and decide what is the first and last hook it can be set in to be effective.
Wed Nov 11 18:54:41 2015  zwoop:jpeach make it so, #1!
Wed Nov 11 18:54:46 2015  sudheerv:and the problem is that analysis can change over time
Wed Nov 11 18:55:01 2015  zwoop:(fwiw, we have many APIs that suffer from the same problem, and they have no support for this either :/)
Wed Nov 11 18:55:04 2015  jpeach:currently plugin developers need to do that however
Wed Nov 11 18:55:06 2015  sudheerv:yeah
Wed Nov 11 18:55:11 2015  zwoop:jpeach yep
Wed Nov 11 18:55:37 2015  zwoop:jpeach it might be worth adding a general concept of this knowledge into every API (at least those related to the HttpSM and hooks)
Wed Nov 11 18:55:48 2015  sudheerv:i think adding documentation on a given setting to indicate at what hook it's available to modify, that's probably good?
Wed Nov 11 18:56:23 2015  jpeach:that requires the same amout of knowledge as a code change, and is less testable :)
Wed Nov 11 18:56:33 2015  zwoop:it'd be better with compile time errors (if/when possible) and runtime errors / warnings. But, a lot of work.
Wed Nov 11 18:56:34 2015  jpeach:but it is still an imprivement
Wed Nov 11 18:59:19 2015  zwoop:if someone goes through this effort, we should aim for both docs and code support
Wed Nov 11 19:03:14 2015  zwoop:ah, maybe not, maybe it's only some well defined resolutions it supports ?
Wed Nov 11 19:07:48 2015  blattj:Joined the channel
Wed Nov 11 19:28:25 2015  _klk_:Joined the channel
Wed Nov 11 19:51:04 2015  _klk_:Joined the channel
Wed Nov 11 19:57:24 2015  blattj:Joined the channel
Wed Nov 11 19:57:43 2015  _klk_:Joined the channel
Wed Nov 11 20:36:19 2015  _klk_:Joined the channel
Wed Nov 11 20:57:50 2015  shinrich1:Joined the channel
Wed Nov 11 21:18:32 2015  zwoop:hmmm, that made no sense, since it was in the wrong window ....
Wed Nov 11 21:19:43 2015  niq:zwoop: the start of a ghost story?
Wed Nov 11 21:20:07 2015  zwoop:indeed
Wed Nov 11 21:44:22 2015  frank-:Joined the channel
Wed Nov 11 23:02:19 2015  bahumbug:Joined the channel
Wed Nov 11 23:33:51 2015  _klk_:Joined the channel

Comments