Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2015-08-19

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 overclk joined #gluster-dev
00:13 jrm16020 joined #gluster-dev
00:53 lpabon joined #gluster-dev
01:06 yliu joined #gluster-dev
01:09 dlambrig joined #gluster-dev
01:13 dlambrig joined #gluster-dev
01:47 ilbot3 joined #gluster-dev
01:47 Topic for #gluster-dev is now Gluster Development Channel - http://gluster.org | For general chat go to #gluster | Patches - http://review.gluster.org/ | Channel Logs - https://botbot.me/freenode/gluster-dev/ & http://irclog.perlgeek.de/gluster-dev/
01:50 dlambrig joined #gluster-dev
01:58 _Bryan_ joined #gluster-dev
02:13 dlambrig joined #gluster-dev
02:15 baojg joined #gluster-dev
02:48 baojg joined #gluster-dev
02:55 badone_ joined #gluster-dev
03:20 Byreddy joined #gluster-dev
03:34 krishnan_p joined #gluster-dev
03:36 shubhendu joined #gluster-dev
03:40 kanagaraj joined #gluster-dev
03:42 dlambrig joined #gluster-dev
03:42 itisravi joined #gluster-dev
03:45 nthomas joined #gluster-dev
03:46 nbalacha joined #gluster-dev
04:02 atinm joined #gluster-dev
04:03 dlambrig joined #gluster-dev
04:05 kkeithley1 joined #gluster-dev
04:07 kdhananjay joined #gluster-dev
04:13 ppai joined #gluster-dev
04:17 gem joined #gluster-dev
04:17 gem_ joined #gluster-dev
04:22 skoduri joined #gluster-dev
04:32 hgowtham joined #gluster-dev
04:32 ashiq joined #gluster-dev
04:35 asengupt joined #gluster-dev
04:43 jiffin joined #gluster-dev
04:46 ndarshan joined #gluster-dev
04:52 deepakcs joined #gluster-dev
04:53 Bhaskarakiran joined #gluster-dev
05:01 vimal joined #gluster-dev
05:04 sakshi joined #gluster-dev
05:07 aravindavk joined #gluster-dev
05:10 Manikandan joined #gluster-dev
05:11 poornimag joined #gluster-dev
05:23 Bhaskarakiran joined #gluster-dev
05:32 kotreshhr joined #gluster-dev
05:34 rafi joined #gluster-dev
05:36 skoduri joined #gluster-dev
05:39 atalur joined #gluster-dev
05:40 atalur joined #gluster-dev
05:43 vmallika joined #gluster-dev
05:44 anekkunt joined #gluster-dev
05:47 aspandey joined #gluster-dev
05:50 nbalacha joined #gluster-dev
05:58 Saravana_ joined #gluster-dev
05:58 baojg joined #gluster-dev
06:04 raghu joined #gluster-dev
06:06 kshlm joined #gluster-dev
06:09 aspandey joined #gluster-dev
06:15 nthomas joined #gluster-dev
06:15 shubhendu joined #gluster-dev
06:16 ndarshan joined #gluster-dev
06:17 asengupt joined #gluster-dev
06:20 ggarg joined #gluster-dev
06:20 ndevos hchiramm_: ah, that reminds me, could you PLEASE create a specs repository in Gerri?
06:20 ndevos +t
06:20 hchiramm_ its available already
06:21 ndevos hchiramm_: oh, thanks, I must have missed the announcement :)
06:21 hchiramm_ http://review.gluster.org/#/a​dmin/projects/glusterfs-specs
06:21 hchiramm_ ndevos, need to clone the repo ^^^
06:21 hchiramm_ rastar_afk, ^^ can u please try the same ?
06:21 Bhaskarakiran joined #gluster-dev
06:22 hchiramm_ rastar, ^^^
06:22 dlambrig joined #gluster-dev
06:22 rastar ndevos: hchiramm_ I can access it
06:22 rastar Will try sending a spec today
06:23 rastar We can try it out a bit more before announcement?
06:23 hchiramm_ we need to push https://github.com/gluster/glusterfs-specs to review.gluster.org
06:23 hchiramm_ rastar, yes
06:23 rastar hchiramm_: and then make github/gluster/glusterfs-specs read-only?
06:24 hchiramm_ it should be the case .
06:24 mchangir joined #gluster-dev
06:24 krishnan_p joined #gluster-dev
06:24 hchiramm_ there should be a mirroring plugin which do that
06:25 mchangir I'm building upstream master with gcc 5.1.1 on Fedora 22 and facing runtime issues with unresolved references ... how do we resolve this issue?
06:25 hchiramm_ lets make both up2date now  , then work on mirroring part
06:25 rafi joined #gluster-dev
06:25 kshlm mchangir, What exactly are the errors you're getting?
06:26 hchiramm_ mchangir, http://gluster.readthedocs.org/en/latest​/Administrator%20Guide/Compiling%20RPMS/
06:26 rastar hchiramm_: makes sense
06:26 hchiramm_ mchangir, may be above document can help u
06:26 kshlm hchiramm_, He's speaking about runtime errors, not build errors.
06:26 hchiramm_ ah.. missed that ..
06:27 mchangir "unresolved reference" to function name (for inline functions) ... once inline function is changed to static inline, the problem disappears ... however I'm still can't get glusterfind working ...getting error in mem_get0()
06:28 hchiramm_ rastar, please let me know once u push that .
06:28 rastar hchiramm_: sure
06:30 mchangir rather getting log printed from mem_get0() about memory pool pointer being NULL
06:30 anoopcs hchiramm_, rastar : I got the following warning on cloning glusterfs-specs vis ssh
06:30 anoopcs warning: remote HEAD refers to nonexistent ref, unable to checkout.
06:30 anoopcs s/vis/via
06:31 hchiramm_ anoopcs, checkiing
06:31 ndevos anoopcs: same error here :-/
06:31 rastar mchangir: For inline function related errors, you can use std=gnu89 in CFLAGS until all references to inline functions in glusterfs are changed to static inline
06:32 rastar mchangir: For the mem_get0 issue please save a backtrace of that
06:32 ndevos rastar: you checked out the specs repo from review.gluster.org or from github?
06:32 rastar mchangir: we have not been able to get a bt for it whenever a bug has been filed on that
06:32 kshlm mchangir, Most likely you are facing an issue with unclean installations. You've done installations from multiple sources (yum, git etc) and have not cleaned up correctly between installations.
06:33 kshlm anoopcs, ndevos, hchiramm_, you get that error when you've just created a git repo, but haven't commited anything.
06:33 kshlm ie. its just a bare repo.
06:33 mchangir rastar, should I put std=gnu89 in Makefile.in?
06:33 hchiramm_ kshlm, true..
06:33 rastar ndevos: anoopcs Nope, it is not working for me either..Better to do a sync from github first
06:33 ndevos kshlm: yeah, that is what I thought too, but hchiramm_ has created it?
06:33 kshlm Someone with push access to the gerrit repo needs to push changes.
06:33 ndevos yeah, thats hchiramm_'s task :D
06:33 rastar kshlm: ndevos anoopcs hchiramm_ : Lets do the merge from github first
06:34 anoopcs mchangir, Or if you don't want to compile entire glustefs with GNU89 standard, you can go with -fgnu89-inline which will use the 89 standard only for inline functions
06:34 hchiramm_ yep.. rastar is trying the same ..
06:34 anoopcs hchiramm_, rastar, kshlm :Ok.. I am waiting for the merge.
06:34 rastar mchangir: anoopcs : For now you can just do CFLAGS="std=gnu89" ; make
06:34 Bhaskarakiran joined #gluster-dev
06:34 mchangir rastar, ok
06:35 * ndevos does not get any compile errors on Fedora 22?
06:35 anoopcs ndevos, We still have issues when compiled in debug mode
06:36 ndevos anoopcs: hmm, whats the BZ?
06:36 * anoopcs is checking . . .
06:38 ndevos kkeithley_blr: you want to send a patch for the @@ -> @ change, or shall I include that in Jiffins lookupat patch?
06:38 anoopcs ndevos, Following mail thread have complete explanations in case you missed it. http://www.gluster.org/pipermail/g​luster-devel/2015-July/046211.html.
06:39 kaushal_ joined #gluster-dev
06:39 anoopcs ndevos, BZ 1245331
06:41 anoopcs ndevos, And we have http://review.gluster.org/#/c/11769/ to address all inline usages in glusterfs.
06:41 rastar hchiramm_: First we will have to do a sync..
06:46 ndevos anoopcs: thanks!
06:48 anoopcs ndevos, np
06:48 hchiramm_ rastar, yes..
06:48 ndevos hchiramm_ or raster: are your doing that now?
06:50 hchiramm_ anoopcs, http://username@review.gluster.org/glusterfs-specs can u clone this
06:50 badone__ joined #gluster-dev
06:50 hchiramm_ ndevos, trying it
06:50 anoopcs hchiramm_, nop :(
06:51 anoopcs hchiramm_, Same error as before
06:51 mchangir rastar, no can do ... [2015-08-19 06:41:06.839500] W [MSGID: 101095] [xlator.c:189:xlator_dynload] 0-xlator: /usr/local/lib/libgfdb.so.0: undefined symbol: gf_sql_str2sync_t
06:53 mchangir rastar, after a CFLAGS="std=f89" make install   and even after a   CFLAGS=-fgnu89-inline make install  after a "make clean"  ... I get get the unresolved symbol error in the brick log
06:54 ppai_ joined #gluster-dev
06:54 anoopcs mchangir, Yuu need to give CFLAGS before configuring I guess
06:54 mchangir oh!
06:55 mchangir ok
06:55 anoopcs mchangir, Or else you can directly edit configure.ac to include either of the above mentioned options and do ./configure and make and make install.
06:55 ggarg joined #gluster-dev
06:56 atalur joined #gluster-dev
06:56 ndevos mchangir: you can also just cherry-pick the inline-fix and build your change on top of that ;-)
06:56 anoopcs mchangir, You can change line no:271 from configure.ac to look like CFLAGS="${CFLAGS} -g -O2 -fgnu89-inline"
07:00 anoopcs mchangir, But if you are compiling without debug mode, we should not hit these undefined symbol issues and we don't need those changes. Only when glusterfs is compiled in debug mode, we have similar problems.
07:00 mchangir ok
07:01 kkeithley_blr I hope we're only thinking of -fgnu89-inline as a short term work-around.
07:02 anoopcs kkeithley_blr, Yes. Because we have Jeff's patch under review.
07:02 mchangir for now I've configured the sources with the CFLAGS on the command-line
07:02 kkeithley_blr okay
07:07 kshlm joined #gluster-dev
07:10 nbalacha joined #gluster-dev
07:16 nthomas joined #gluster-dev
07:19 ndarshan joined #gluster-dev
07:21 hchiramm_ anoopcs, ndevos the repo should be fine now
07:25 rastar mchangir: that is another bug to fix...please configure with sqlite
07:28 rastar mchangir: I mean don't use disable-tiering
07:34 mchangir rastar, ok
07:39 anoopcs hchiramm_, Works now.
07:39 anoopcs hchiramm_++
07:39 glusterbot anoopcs: hchiramm_'s karma is now 12
07:39 anoopcs rastar++
07:39 glusterbot anoopcs: rastar's karma is now 11
07:43 ggarg joined #gluster-dev
07:45 atalur joined #gluster-dev
07:48 shubhendu joined #gluster-dev
07:48 ppai_ joined #gluster-dev
07:55 ndarshan joined #gluster-dev
07:55 RedW joined #gluster-dev
08:11 deepakcs joined #gluster-dev
08:21 timbyr_ joined #gluster-dev
08:32 chan5n joined #gluster-dev
08:47 badone__ is this common?
08:47 badone__ ./tests/basic/afr/arbiter-statfs.t (Wstat: 0 Tests: 22 Failed: 6)
08:47 badone__ Failed tests:  15-17, 20-22
08:49 hchiramm_ itisravi, ^^^
08:49 hchiramm_ badone, it shouldnt be common
08:49 hchiramm_ :)
08:49 badone hchiramm_: I made a minor code change on master head and compiled and installed on a standalone system and got that
08:50 itisravi badone: does the test pass without your change?
08:50 badone itisravi: I haven't tested. Let me do that
08:50 hchiramm_ yeah, that need to be checked
08:51 hchiramm_ badone, sometime it can be spurious failure irrespective of the change.. but not sure the case here.
08:51 badone hchiramm_: are there logs anywhere?
08:51 itisravi badone: well it should pass or else it wouldn't have been merged post regressions.
08:51 itisravi hchiramm_: yeah
08:52 badone I'd like to grab the logs before I run again?
08:52 hchiramm_ badone, if u run standalone ( means in your system) there wont be any log
08:52 badone okay, let me checkout master and try again
08:52 hchiramm_ but u could comment out cleanup() function and get the logs
08:53 itisravi hchiramm_: why not? It should get logged in /var/log/gluster no?
08:53 hchiramm_ itisravi, I think we do a cleanup
08:53 hchiramm_ need to cross check though
08:53 itisravi hchiramm_: don't think cleanup removes the logs.
08:54 hgowtham anoopcs++ rafi++
08:54 glusterbot hgowtham: anoopcs's karma is now 17
08:54 glusterbot hgowtham: rafi's karma is now 28
08:55 shubhendu joined #gluster-dev
08:55 hchiramm_ arbiter-statfs.t:cleanup;
08:55 hchiramm_ arbiter-statfs.t:cleanup;
08:55 hchiramm_ arbiter.t:cleanup;
08:55 hchiramm_ arbiter.t:cleanup
08:55 hchiramm_ itisravi, ^^^ isnt it doing the cleanup ?
08:56 itisravi hchiramm_: no, I mean the function cleanup in tests/include.rc doesn't remove the logs.
08:58 hchiramm_ ok.. looks like we do cleanup of "/var/lib/."
08:59 badone there is no /var/log/gluster on this system
08:59 badone it's just a fedora dev system
08:59 hchiramm_ but logs will be there.
09:02 itisravi badone: /var/log/glusterfs/
09:03 badone ls /var/log/gluster/
09:03 badone ls: cannot access /var/log/gluster/: No such file or directory
09:03 itisravi either there or /usr/local/var/log/glusterfs if it is glusterfs-3.5 I think
09:03 itisravi glusterfs not gluster
09:04 badone doh! my bad, sorry
09:06 badone itisravi: can't find any logs in there for test runs
09:06 chan5n left #gluster-dev
09:07 badone itisravi: master fails as well
09:08 badone so without my commit it fails anyway
09:08 badone a462e1f3146ad1c0e1099bbc4d0f5269a1c02545
09:09 chan5n joined #gluster-dev
09:12 pranithk joined #gluster-dev
09:14 badone anyway, I can submit my patch because it's not the problem.
09:23 aspandey joined #gluster-dev
09:43 josferna joined #gluster-dev
09:46 ppai_ joined #gluster-dev
10:02 pranithk joined #gluster-dev
10:05 shubhendu joined #gluster-dev
10:08 kshlm joined #gluster-dev
10:12 kshlm joined #gluster-dev
10:12 kdhananjay1 joined #gluster-dev
10:18 poornimag joined #gluster-dev
10:19 vipulnayyar joined #gluster-dev
10:20 primusinterpares joined #gluster-dev
10:27 baojg joined #gluster-dev
10:29 dlambrig joined #gluster-dev
10:44 josferna joined #gluster-dev
11:12 atinm joined #gluster-dev
11:16 jrm16020 joined #gluster-dev
11:21 gem sakshi++
11:21 glusterbot gem: sakshi's karma is now 5
11:27 krishnan_p joined #gluster-dev
11:32 rafi1 joined #gluster-dev
11:40 hchiramm_ kshlm++ thanks !
11:40 glusterbot hchiramm_: kshlm's karma is now 32
11:49 ndevos hchiramm_, rafi1, atinm, kkeithley_blr: any idea who is hosting todays meeting?
11:49 rafi1 ndevos: meghna will host todays meeting
11:49 rjoseph joined #gluster-dev
11:49 ndevos rafi1: oh, great!
11:50 kshlm hchiramm_, https://review.gluster.org/11958
11:50 hchiramm_ kshlm, checking
11:51 meghanam_ joined #gluster-dev
11:51 hchiramm_ kshlm, merged
11:51 hchiramm_ kshlm++
11:51 glusterbot hchiramm_: kshlm's karma is now 33
11:53 meghanam_ REMINDER : Gluster Weekly  Community Meeting will start in ~10 minutes on #gluster-meeting.
11:56 firemanxbr joined #gluster-dev
11:56 kshlm hchiramm_, one more https://review.gluster.org/11959
11:56 hchiramm_ kshlm, checking
11:57 hchiramm_ kshlm, done
11:58 kshlm Now all we need to do is setup replication!
11:59 hchiramm_ indeed ..
12:04 jdarcy joined #gluster-dev
12:06 gp joined #gluster-dev
12:07 Guest4573 joined #gluster-dev
12:08 itisravi joined #gluster-dev
12:12 rjoseph joined #gluster-dev
12:12 raghu joined #gluster-dev
12:51 kkeithley1 joined #gluster-dev
12:56 kaushal_ joined #gluster-dev
13:01 asengupt joined #gluster-dev
13:03 shyam joined #gluster-dev
13:03 kotreshhr left #gluster-dev
13:05 rafi1 meghanam_++
13:05 glusterbot rafi1: meghanam_'s karma is now 2
13:07 aravindavk joined #gluster-dev
13:08 atinm joined #gluster-dev
13:08 vimal joined #gluster-dev
13:12 ashiq joined #gluster-dev
13:13 lpabon joined #gluster-dev
13:14 krishnan_p joined #gluster-dev
13:15 kotreshhr joined #gluster-dev
13:16 vipulnayyar left #gluster-dev
13:19 skoduri joined #gluster-dev
13:20 rafi joined #gluster-dev
13:24 ashiq joined #gluster-dev
13:30 ppai_ joined #gluster-dev
14:07 shubhendu joined #gluster-dev
14:11 kbyrne joined #gluster-dev
14:28 kshlm joined #gluster-dev
14:32 itisravi joined #gluster-dev
14:33 deepakcs joined #gluster-dev
14:35 nthomas joined #gluster-dev
14:35 itisravi_ joined #gluster-dev
14:50 _Bryan_ joined #gluster-dev
15:12 _Bryan_ joined #gluster-dev
15:38 wushudoin| joined #gluster-dev
15:45 atinm joined #gluster-dev
15:53 ggarg joined #gluster-dev
15:54 cholcombe joined #gluster-dev
16:26 skoduri joined #gluster-dev
17:00 jrm16020 joined #gluster-dev
17:30 gem joined #gluster-dev
17:38 cholcombe nixpanic_, were you able to figure out how to decode the glusterfs auth / cred struct in your wireshark code?
17:44 ndevos cholcombe: yes, see the two functions starting at line 1180 https://code.wireshark.org/review/gi​tweb?p=wireshark.git;a=blob;f=epan/d​issectors/packet-rpc.c;hb=HEAD#l1180
17:45 cholcombe ndevos, ah thanks!  I've been digging to try and figure out how to encode these and send them to gluster.  I have filler padding that just so happens to work
17:46 ndevos cholcombe: you're welcme :)
17:46 cholcombe ndevos, btw i found a bug in gluster's rpc decoding that crashes quotad every time i send it the packet
17:47 ndevos cholcombe: oh, wow, did you file a bug for that?
17:47 cholcombe not yet
17:47 ndevos wait, I think that rings a bell, something with a incorrect path/string or such?
17:47 cholcombe it decodes the packet and it gets asked for 140TB of ram.  that causes it to sigsev
17:47 cholcombe https://gist.githubusercontent.com/cholcomb​e973/64a0fda5646aa5d3d064/raw/2cb1ead4226fd​e1ab6b03e994f04ca58a0f21c0d/gistfile1.txt
17:47 ndevos oh, maybe not...
17:47 cholcombe there's the crash
17:48 cholcombe it's interesting because the function says no i can't give you that much ram and then it panics and crashes
17:48 cholcombe so something is not handling the error properly but i'm having trouble figuring out what
17:48 ndevos hah, yes
17:48 cholcombe i suspect glusterd would crash if i sent it this packet also
17:49 ndevos hmm :-/
17:49 cholcombe basically what happened was i sent it a dict but forgot to turn it into a string first for xdr
17:49 cholcombe so it started reading the wrong data and thought it needed a giant buffer and blew up
17:50 cholcombe i think there's another thread that is crashing and not the one i'm watching on gdb
17:50 ndevos oh, thats bad, definitely needs some hardening there
17:50 cholcombe yeah i suspect the rpc code was never sent fuzzy data
17:50 cholcombe and me trying to figure it out caused all kinds of bad data to be generated haha
17:51 ndevos yeah, I've seen bad rpc behaviour too, and have no doubt that there are more issues in there
17:51 cholcombe me too.  i'm sure of it
17:51 cholcombe i'd love to figure out the patch myself but i'm kinda stuck at the moment
17:52 ndevos I guess rpc hardening should be a general thing to look into for a next major release (3.8 or 4.0)
17:52 cholcombe yeah. i'm sure i could help :)
17:53 cholcombe the opaque strings are the problem
17:53 ndevos that would be most welcome!
17:53 cholcombe if the RPC mechanism just passed a struct instead of a string it would be fine
17:53 ndevos yeah, I hate the dicts in the protocol
17:53 cholcombe me too.
17:53 cholcombe i think it was a lazy hack
17:54 ndevos definitely!
17:54 cholcombe it also creates undocumented crap that i have to find all over the place
17:54 ndevos I would like to see a dict entry have a type, so that we can at least verify the structure
17:54 cholcombe each dict seems to have a slightly different encoding too which is really frustrating
17:54 cholcombe yeah a type at the very least would help
17:55 cholcombe at the moment it's just a raw buffer
17:55 ndevos some entries have GFIDs in the binary UUID format, and that breaks little <-> big endian communication :-(
17:56 cholcombe ugh
17:56 cholcombe lol
17:56 cholcombe what a mess
17:57 ndevos indeed, and I have no idea how we can fix that decently - hence the type for a new dict on the protocol (but that is 4.0 material)
17:57 cholcombe i think if the dict went away it would eliminate huge swaths of code and improve things
17:59 cholcombe ndevos, here's what i have so far: https://github.com/cholcombe973/Gl​uster/blob/master/src/rpc.rs#L734
17:59 ndevos yes it would, but I doubt I can convince everyone of that
17:59 cholcombe ndevos, maybe i should email the list and irritate everyone with this idea
18:00 ndevos cholcombe: oh, ew, is that ruby?
18:00 cholcombe ndevos, it's Rust
18:00 ndevos cholcombe: yes, please share your opinion on the list :D
18:00 cholcombe lol
18:00 cholcombe i have quite a strong one after trying to decode this crap
18:01 ndevos ah, yes, ruby is weirder, this is actually quite readable
18:01 cholcombe :)
18:01 cholcombe thanks!
18:01 cholcombe i have a lot of comments in there because the cred/auth thing is confusing
18:02 ndevos I think the functions in the wireshark sources should explain it well?
18:02 ndevos there is also a header file in the gluster sources that contains more details
18:03 cholcombe ah. maybe i should check out that header file
18:03 cholcombe ndevos, yeah i'll try and trace through your code when i have a minute
18:03 cholcombe it looks quite a bit more readable without all the jumping around the glusterfs src has
18:04 ndevos oh, yes, getting all the bits was rather a lot of jumping through the code
18:04 cholcombe indeed
18:07 ndevos cholcombe: https://github.com/gluster/glusterfs/blob​/master/rpc/rpc-lib/src/xdr-common.h#L38 for the AUTH_GLUSTERFS_V2 structure
18:08 cholcombe ah sweet
18:08 cholcombe that's a great comment
18:09 cholcombe ndevos, so i was close
18:09 ndevos cholcombe: yes, very close :)
18:09 cholcombe 5 xdr units.  that makes sense.  I figured lk_owner_val and groups_val was a string
18:11 ndevos the lkowner is an opaque value, it is not interpreted by the gluster server-side
18:11 cholcombe ndevos, creds and verf seem to be the same thing?
18:11 cholcombe ndevos, then what is it used for?
18:11 ndevos I think the verifier is never used
18:12 ndevos but, its quite a while since I looked at those bits, so I could be wrong
18:12 cholcombe ndevos, i think so also.  i gave it some kinda bogus data and it just let me through no problem
18:12 cholcombe i just passed it 0's and it was like cool, ok proceed
18:13 ndevos the lkowner is set by applications (think nfs-ganesha or samba) so that a specific NFS/Samba client can be identified - used for locks
18:13 cholcombe oh i see
18:13 cholcombe it's interesting in the code i see that the lk_owner_val is a "" with 0x00 0x00 0x00 0x00 as the values
18:13 cholcombe i'm like what the hell is this? haha
18:14 ndevos right, a zero-length opaque
18:15 cholcombe it's not zero length though
18:15 cholcombe gdb says it's len = 4
18:15 cholcombe which has me scratching my head
18:16 ndevos oh, is that not the 4 bytes for the 1 xdr-unit that contains the length?
18:16 cholcombe oh.. i see
18:16 cholcombe yes i misread that
18:16 cholcombe it's a string and you always start an xdr string by giving 4 bytes to the len
18:17 ndevos yes, correct
18:17 cholcombe ndevos, ok i need to rejig my code a little then
18:17 ndevos well, "string" is not really correct, it can be anything, but the size of the <anything> is prefixed
18:18 cholcombe right
18:18 cholcombe i think xdr doesn't know what opaque means so it just becomes a string
18:18 ndevos oh, that could be
18:26 * ndevos drops off, its close to midnight here
18:27 dlambrig joined #gluster-dev
18:35 jrm16020 joined #gluster-dev
19:11 _Bryan_ joined #gluster-dev
20:07 jrm16020 joined #gluster-dev
21:18 badone_ joined #gluster-dev
22:18 wushudoin joined #gluster-dev
22:55 wushudoin| joined #gluster-dev
23:00 wushudoin| joined #gluster-dev
23:56 nishanth joined #gluster-dev

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary