[19:42] <veebers> barry: ping, following up my email :-)
[19:58] <barry> veebers: hi.  the problem is that pybootchartgui is i think unmaintained and has no effective test suite.  i really don't know much about the code, so my fix was just a bandaid to get it past the given sample data.  if you have other sample data that it's crashing on, i can try to take a look, but i'd need that sample data
[19:59] <barry> i see comment #8
[19:59] <veebers> barry: Ah I see. I have the sample data here, would you like me to email it to you?
[19:59]  * veebers looks
[19:59] <veebers> barry: ack
[20:00] <barry> veebers: i'll see what i can do about that failure
[20:04] <veebers> barry: have emailed the data, let me know if you get it (~8MB should get through, right? :-))
[20:05] <barry> veebers: maybe ;)  can you attach it to LP: #1381177?
[20:05] <veebers> barry: that's a much better idea, doing that now
[20:12] <barry> veebers: looks like the email got through
[20:16] <veebers> barry: ah cool, sorry was checking that there wasn't any private data in that data file before attaching
[20:18] <barry> veebers: no worries, i see the crash.  this is probably going to be a onsie-twosie type fix.  e.g. you give me data it crashes on, i tried to fix it knowing so little about the code, rinse, repeat ;)
[20:18] <veebers> barry: heh, hopefully we get lucky and this is the one time
[20:20] <barry> veebers: yeah.  this code is ancient, obscurely written, etc.
[20:21] <veebers> oh . . . I see :-P
[20:23] <barry> veebers: http://paste.ubuntu.com/8925948/
[20:32] <barry> veebers: do you want to test this before i upload to vivid?
[20:45]  * veebers look
[20:45] <veebers> s
[20:46] <veebers> barry: sure, I'll pull the branch and build itn ow
[20:54] <veebers> barry: d'oh, now I get this error: http://pastebin.ubuntu.com/8926432/
[21:00] <veebers> barry: it looks like that parse code is called a couple of times?
[21:02] <barry> veebers: maybe.  i just ran it like so:
[21:02] <barry> python pybootchartgui.py ../bootchart.tgz
[21:03] <barry> haven't seen that latest pastebin bug
[21:05] <barry> can you tell me how you got it?  i wonder if this isn't just a losing battle
[21:05] <barry> veebers: https://code.google.com/p/pybootchartgui/
[21:05] <barry> veebers: https://github.com/mmeeks/bootchart
[21:07] <veebers> barry: sorry I thought that I had shared the command I use: http://pastebin.ubuntu.com/8926726/
[21:08] <barry> veebers: let me try that in a chroot with the patched package
[21:15] <barry> veebers: that command with the posted patch wfm
[21:16] <veebers> barry: let me try again
[21:17] <veebers> barry: Actually I'm still running utopic and just built the package locally, but that shouldn't be an issue. . . .right?
[21:17] <barry> veebers: shouldn't be, since there's not much changed since then, but i am testing it in a vivid chroot
[21:17] <barry> veebers: i can give you a deb if that's helpful
[21:18] <veebers> barry: I might as well try that, yes please
[21:18] <veebers> (make sure we're testing the same thign
[21:18] <veebers> )
[21:21] <barry> veebers: dget http://barry.warsaw.us/debian/pybootchartgui_0+r141-0ubuntu5_amd64.changes
[21:21] <barry>  
[21:21] <barry> then install the .deb
[21:21] <veebers>  barry ack, thanks
[21:27] <veebers> barry: ok, installing that deb worked for me (although I get 'False' printed out, but meh). I must have screwed something up when I built from scratch
[21:29] <veebers> barry: also, thanks for those other links, it looks like there will be work items coming up to transition to the newer stuff
[21:32] <barry> veebers: cool.  yeah, no idea what that False is but i see it too.  if this works for you, i'll upload this to vivid and we'll wait for the next failure ;)
[21:32] <veebers> barry: heh, awesome, thanks
[21:35] <thomi> veebers: just curious, do you know why we need the gui module at all?
[21:35] <thomi> I mean, we never display the actual gui graph anywhere, right?
[21:36] <thomi> or do we scrape it for the data points or something?
[21:36] <veebers> thomi: Um, that's a good question. Perhaps nuclearbob can weigh in here. I'm not sure, I just know there is this bug
[21:36] <thomi> fair nuff
[21:37] <nuclearbob> thomi veebers: I don't know, actually.  If there's a non-gui tool to generate those data points, that would probably be better.  The tests themselves haven't really be examined in a while, mostly just the way they're run