Displaying #traffic-server/2015-11-10.log:

Tue Nov 10 00:01:50 2015  bcall:jpeach: doesn't look like you tested -f with asan
Tue Nov 10 00:02:06 2015  bcall:jpeach: fixing it now
Tue Nov 10 00:03:35 2015  jpeach:bcall: huh? -f?
Tue Nov 10 00:03:42 2015  bcall:yes
Tue Nov 10 00:03:50 2015  bcall:traffic_server -f
Tue Nov 10 00:04:26 2015  jpeach:ah
Tue Nov 10 00:05:33 2015  bcall:I am in asan hell today trying to get the regression to work with it
Tue Nov 10 00:10:26 2015  gancho:Joined the channel
Tue Nov 10 00:23:04 2015  adam____:Joined the channel
Tue Nov 10 00:23:33 2015  thumbs:Joined the channel
Tue Nov 10 00:38:20 2015  _klk_:Joined the channel
Tue Nov 10 02:32:05 2015  adam_____:Joined the channel
Tue Nov 10 02:36:44 2015  danielxu:Joined the channel
Tue Nov 10 03:10:10 2015  zwoop:what does this mean:
Tue Nov 10 03:10:11 2015  zwoop:Nov 10 02:56:29 s_sys@usdal2-edge-bx-005.ts.apple.com syslog-ng[32757]: The current log file has a mismatching size/inode information, restarting from the beginning; state='affile_sd_curpos(/var/log/trafficserver/apple.log)', stored_inode='267619048', cur_file_inode='269752658', stored_size='707335', cur_file_size='623312', raw_stream_pos='631753072'
Tue Nov 10 03:10:14 2015  zwoop:there's a lot of them
Tue Nov 10 03:10:29 2015  zwoop:does it mean it's reading the log file from the beginning over and over again ? Or is that when the log file is rotated ?
Tue Nov 10 03:11:05 2015  zwoop:meh
Tue Nov 10 04:26:27 2015  es:Joined the channel
Tue Nov 10 04:48:23 2015  danielxu:Joined the channel
Tue Nov 10 04:51:01 2015  adam____:Joined the channel
Tue Nov 10 04:59:33 2015  reveller:zwoop: ^^^
Tue Nov 10 05:01:35 2015  jpeach:dxu: ayt?
Tue Nov 10 05:27:06 2015  jpeach:hey bcall
Tue Nov 10 05:27:11 2015  bcall:KEY
Tue Nov 10 05:27:15 2015  bcall:hey :)
Tue Nov 10 05:27:40 2015  jpeach:in TS-4001, is the link hack really needed?
Tue Nov 10 05:28:19 2015  jpeach:we should be able to do this http://apaste.info/hHe
Tue Nov 10 05:30:00 2015  bcall:yeah, I don't think it is need
Tue Nov 10 05:30:13 2015  bcall:I thought that was odd myself - I will test
Tue Nov 10 05:31:01 2015  bcall:yeah, works for me
Tue Nov 10 05:31:14 2015  jpeach:it is fine on OSX and my fedora is still building
Tue Nov 10 05:31:18 2015  jpeach:you wanna commit?
Tue Nov 10 05:31:23 2015  bcall:sure
Tue Nov 10 05:31:29 2015  jpeach:thanks
Tue Nov 10 05:51:01 2015  bcall:jpeach: ^
Tue Nov 10 05:51:07 2015  bcall:can you take a look at that
Tue Nov 10 05:54:06 2015  masaori:bcall: jpeach: hi, works fine without forceLink?
Tue Nov 10 05:54:16 2015  masaori:http://apaste.info/FEO
Tue Nov 10 05:54:17 2015  bcall:yeah
Tue Nov 10 05:54:43 2015  masaori:in my box, tests are never called
Tue Nov 10 05:55:30 2015  masaori:hmm, I'll check my box
Tue Nov 10 05:59:14 2015  jpeach:ok I was expecting a link fail .. will double-check
Tue Nov 10 05:59:15 2015  jpeach:bcall: what does the RFC say about exact match on IMS?
Tue Nov 10 05:59:15 2015  bcall: If the variant has not been modified since a valid If-
Tue Nov 10 05:59:15 2015  bcall: Modified-Since date, the server SHOULD return a 304 (Not
Tue Nov 10 05:59:15 2015  bcall: Modified) response.
Tue Nov 10 05:59:15 2015  jpeach:that sounds a lot like <= to me then
Tue Nov 10 05:59:35 2015  jpeach:but I think we should be consistent with all the conditionals
Tue Nov 10 05:59:38 2015  bcall: Note: When handling an If-Modified-Since header field, some
Tue Nov 10 05:59:38 2015  bcall: servers will use an exact date comparison function, rather than a
Tue Nov 10 05:59:38 2015  bcall: less-than function, for deciding whether to send a 304 (Not
Tue Nov 10 05:59:38 2015  bcall: Modified) response. To get best results when sending an If-
Tue Nov 10 05:59:39 2015  bcall: Modified-Since header field for cache validation, clients are
Tue Nov 10 05:59:40 2015  bcall: advised to use the exact date string received in a previous Last-
Tue Nov 10 05:59:41 2015  bcall: Modified header field whenever possible.
Tue Nov 10 06:01:00 2015  bcall:so you are supposed to send the same value as the proxy gave you - so they should be equal most of the time
Tue Nov 10 06:01:08 2015  bcall:and you should get a 304 back
Tue Nov 10 06:02:00 2015  bcall:[bcall@homer trafficserver]$ sudo /usr/local/bin/traffic_server -R 3 -r HPACK
Tue Nov 10 06:02:00 2015  bcall:traffic_server: using root directory '/usr/local'
Tue Nov 10 06:02:00 2015  bcall:REGRESSION_TEST initialization begun
Tue Nov 10 06:02:00 2015  bcall:REGRESSION_TEST DONE: PASSED
Tue Nov 10 06:02:11 2015  bcall:that is what I am getting now
Tue Nov 10 06:02:15 2015  jpeach:so it sounds like we are compliant?
Tue Nov 10 06:02:22 2015  bcall:yes
Tue Nov 10 06:02:33 2015  masaori:bcall: tests are not called.
Tue Nov 10 06:02:58 2015  bcall:yeah, that seems really odd that we would need a hack like that
Tue Nov 10 06:03:30 2015  masaori:I saw same hack in HttpTransact.cc https://github.com/apache/trafficserver/blob/master/proxy/http/HttpTransact.cc#L9083
Tue Nov 10 06:04:12 2015  jpeach:that's odd, it should be linked because of the static constructors
Tue Nov 10 06:05:44 2015  bcall:http://paste.fedoraproject.org/288723/14471355
Tue Nov 10 06:05:59 2015  bcall:that is with the hack back in
Tue Nov 10 06:06:18 2015  masaori:if I understand correctly, when linker links objects, functions which not called from main() will be removed. This is why we need this hack.
Tue Nov 10 06:06:46 2015  dxu:Joined the channel
Tue Nov 10 06:07:39 2015  danielxu:Joined the channel
Tue Nov 10 06:10:32 2015  jpeach:that's sad :(
Tue Nov 10 06:10:45 2015  jpeach:would it make sense to move the tests into HPACK.cc?
Tue Nov 10 06:13:22 2015  masaori:IMO, it's easy to understand when some body add tests if those test are separated in a file.
Tue Nov 10 06:14:21 2015  jpeach:alright let's revert ... bcall?
Tue Nov 10 06:14:24 2015  masaori:jpeach: it is better than those tests are in HTTP2.cc
Tue Nov 10 06:14:32 2015  bcall:yeah, I can revert it
Tue Nov 10 06:15:03 2015  bcall:done :)
Tue Nov 10 06:15:06 2015  masaori:bcall: jpeach: thanks
Tue Nov 10 06:15:08 2015  jpeach:maybe another alternative is to build a test_hpack program and run it at "make check"
Tue Nov 10 06:15:58 2015  masaori:sounds good
Tue Nov 10 06:16:13 2015  jpeach:bcall: I'm going to commit TS-4006
Tue Nov 10 06:16:18 2015  jpeach:TS-4006
Tue Nov 10 06:16:32 2015  bcall:sure
Tue Nov 10 06:16:42 2015  bcall:I am going to bed :)
Tue Nov 10 06:17:04 2015  jpeach:nite nite
Tue Nov 10 06:17:35 2015  masaori:nite
Tue Nov 10 06:34:59 2015  dxu:jpeach: I'm back, what's up?
Tue Nov 10 06:46:20 2015  gancho:Joined the channel
Tue Nov 10 06:49:12 2015  danielxu:Joined the channel
Tue Nov 10 08:06:50 2015  adam____:Joined the channel
Tue Nov 10 08:29:44 2015  Zerpex:Joined the channel
Tue Nov 10 08:31:51 2015  Zerpex:Yo, if I have "deb http://ppa.launchpad.net/apache-helpdesk/trafficserver-ppa/ubuntu trusty main" .. how do I install version 5.3.2 ? - version 6 keeps giving segfaults..
Tue Nov 10 09:18:39 2015  Lethalman:Joined the channel
Tue Nov 10 09:41:18 2015  briang:Zerpex, can you build from source?
Tue Nov 10 09:41:36 2015  Zerpex:I prefer not to :/
Tue Nov 10 09:42:10 2015  briang:I think daemonkeeper maintains the debian packages
Tue Nov 10 09:42:15 2015  briang:maybe he can comment.
Tue Nov 10 09:42:39 2015  briang:but if you're interested in building from source I can help w/ that.
Tue Nov 10 09:43:56 2015  briang:I think all you need to do is force it to install the 5.3.2 package
Tue Nov 10 09:44:04 2015  briang:from what I can see it's alreayd built: https://launchpad.net/~apache-helpdesk/+archive/ubuntu/trafficserver-ppa/+sourcepub/5391444/+listing-archive-extra
Tue Nov 10 09:44:10 2015  briang:you can also just download the package and install it
Tue Nov 10 09:44:32 2015  briang:just pull the amd64 5.3.2 package
Tue Nov 10 09:45:41 2015  sommarnatt:Joined the channel
Tue Nov 10 10:33:56 2015  adam____:Joined the channel
Tue Nov 10 11:28:00 2015  danielxu:Joined the channel
Tue Nov 10 11:43:09 2015  JSeymour:Joined the channel
Tue Nov 10 11:47:24 2015  lrea:Joined the channel
Tue Nov 10 12:22:20 2015  bahumbug:Joined the channel
Tue Nov 10 12:22:20 2015  bahumbug:Joined the channel
Tue Nov 10 12:37:41 2015  pyry:Joined the channel
Tue Nov 10 13:05:21 2015  reveller:jpeach: thanks for committing TS-4006
Tue Nov 10 13:51:31 2015  adam____:Joined the channel
Tue Nov 10 13:56:12 2015  adam____:Joined the channel
Tue Nov 10 14:09:13 2015  shinrich1:Joined the channel
Tue Nov 10 14:30:16 2015  zwoop:gah, who put me down to do Git during the hackathon?
Tue Nov 10 14:30:26 2015  zwoop:That was supposed to be James and Phil?
Tue Nov 10 14:36:34 2015  esproul:Joined the channel
Tue Nov 10 14:37:10 2015  reveler:Joined the channel
Tue Nov 10 14:53:57 2015  es:Joined the channel
Tue Nov 10 15:12:30 2015  gancho:Joined the channel
Tue Nov 10 15:19:32 2015  davet_:Joined the channel
Tue Nov 10 15:28:07 2015  amc:PSUdaemon - ping
Tue Nov 10 15:44:42 2015  PSUdaemon:amc: pong
Tue Nov 10 15:47:08 2015  amc:Hey, I have slides for you and theGeffon to look at. Hold on.
Tue Nov 10 15:47:34 2015  Humbedooh:I though it was the Gaff
Tue Nov 10 15:47:34 2015  amc:These are just a sketch, really, because my plan is to introduce the slide and then say "Phil, take it away!"
Tue Nov 10 15:47:49 2015  Humbedooh:as you should
Tue Nov 10 15:48:15 2015  amc:"Yes, yes, that's a great question ... for Brian"
Tue Nov 10 15:55:34 2015  PSUdaemon:ok, will take a look
Tue Nov 10 15:55:42 2015  PSUdaemon:fwd'd to jrushford too
Tue Nov 10 16:24:54 2015  lrea:Left the channel
Tue Nov 10 16:26:53 2015  reveller:Joined the channel
Tue Nov 10 16:34:29 2015  shinrich2:Joined the channel
Tue Nov 10 16:40:15 2015  bahumbug:Joined the channel
Tue Nov 10 16:45:10 2015  blattj:Joined the channel
Tue Nov 10 16:52:21 2015  amc:PSUdaemon - If you can think of any other basic things we should cover, I can add those.
Tue Nov 10 17:20:09 2015  jrushford:Joined the channel
Tue Nov 10 17:33:28 2015  blattj:Joined the channel
Tue Nov 10 17:42:38 2015  gancho:Joined the channel
Tue Nov 10 18:01:14 2015  ASFBot:Joined the channel
Tue Nov 10 18:13:17 2015  _klk_:Joined the channel
Tue Nov 10 18:33:01 2015  es:Joined the channel
Tue Nov 10 18:36:29 2015  blattj:Joined the channel
Tue Nov 10 18:40:33 2015  _klk_:Joined the channel
Tue Nov 10 19:11:40 2015  jrushford:Joined the channel
Tue Nov 10 21:33:55 2015  _klk_:Joined the channel
Tue Nov 10 22:03:52 2015  es1:Joined the channel
Tue Nov 10 22:45:13 2015  niq:Joined the channel
Tue Nov 10 22:52:06 2015  blattj:Joined the channel
Tue Nov 10 23:01:46 2015  blattj:Joined the channel

Comments