IRC Log Viewer » #firebreath » 2011-10-15

IRC Nick Time (GMT-7) Message
linearray 03:10 taxilian_away: does this help?
FireBreathBot 03:10 07af0f0 by Max Amanshauser: Documented SystemHelpers.h
linearray 03:10 when you come back maybe you can tell me how to send a pull request for one specific commit, because frankly I have no idea what I'm doing ;)
taxilian 09:10 linearray: to make a pull request with just one item you need to create a new branch with just that commit in it
and do the pull request on that
FireBreathBot 09:10 Commit 5c3b704 on firebreath-1.6 by Max Amanshauser: "Documented SystemHelpers.h"
Commit 5c3b704 on master by Max Amanshauser: "Documented SystemHelpers.h"
taxilian 10:10 linearray: thanks for doing that
linearray 10:10 does it help? i was unsure ecause it is not a class
no idea how it will ahow up in confluence
taxilian 10:10 it should work; I just ran the reload script and found out there was an error keeping it from completing :-P
so if nothing else, it got me to discover that the docs haven't been updating correctly for awhile =]
linearray: so making a pull request the easiest way is usually to make a new branch based on the branch you want it pulled into (so in this case the master branch from the main repo) and then cherry-pick your change(s) in
taxilian 10:10 grr.. I still don't fully understand python and its unicode errors
not sure how to fix this
taxilian 11:10 …. crap
cripity crap
so the new confluence 4.0 has broken my import script
kylehuff 11:10 anything I can do to help?
taxilian 11:10 probably not unless you want to just fix it for me; it's probably not going to be terrible, but it's a bit of a convoluted script
if you want to take a look at how it works, it's in the source tree doxygen/
kylehuff 11:10 well, I am willing to try
what exactly is broken? er, rather, how can I see the problem?
taxilian 11:10 hmm. not easily =] the problem is visible when trying to import doxygen output into confluence
hmm. found the issue, though; "Wiki formatted content can no longer be retreived using this API"
because they dropped wiki formatting
kylehuff 11:10 I see
taxilian 11:10 "Confluence 4.0 introduced changes to the way page content is stored in Confluence. Page content is no longer stored as wiki markup, but is now stored in an XHTML based storage format."
hmm. yeah, I don't think there is much you can do without having access to the server, and I can't give you that access at the moment :-/
looks like I'll need to update to use the soap api instead
how is your python?
kylehuff 11:10 ah, okay. well, let me know if there something I can do.
python? rusty, but probably better than my C
taxilian 11:10 hehe. well,
need to make a soap client for confluence
or find one
and make it so that as much as possible it has the same interface as the client in the file
if you could figure that out, that'd save me some time; I have a ham radio test session to go to, so I'll be AFK for a couple of hours and then distracted most of hte rest of the afternoon
but if I had that, the rest might be fairly easy
kylehuff 11:10 okay, I will see what I can do in-between screaming at kids
taxilian 11:10 looks like we might be able to use the "convertWikiToStorageFormat" to avoid needing to figure out the new format ourselves
hehe. fair enough :-P
I appreciate your willingness to help
you should be able to use your own login to test things for the most part
I gotta run now
kylehuff 11:10 later
linearray 13:10 ohai
kylehuff 13:10 taxilian: I made the change you suggested (convertWikiToStorageFormat), but as my comment on the following commit describes, the server doesn't like it
FireBreathBot 13:10 f0398b3 by Kyle L. Huff: an attempt to fix the docs2confluence script
taxilian 13:10 according to what I've been reading, I think we'll have to use the SOAP api
looks like they've dropped support for getPage and storePage in the xmlrpc one
kylehuff 13:10 ah, buggar...
taxilian 13:10 yeah
kylehuff 13:10 I did some very basic playing with the SOAP api, but I couldn't get it to load, not sure if the confluence plugin is not enabled or what.
I could convert it over to use the SOAP api, I just couldn't access it via at "/rpc/soap/jirasoapservice-v2?wsdl"
taxilian 13:10 hmm
nir 14:10 hello
kylehuff 14:10 hello
nir 14:10 the generated dll for a simple plugin is 3MB (1MB zipped) is there a way to significantly reduce its size?
kylehuff 14:10 okay, my bad, jirasoapservice works fine, when pointed to, but I think I'm going to need to find the soap service URL for confluence, not jira
nir: what target? have you tried MinSizeRel?
taxilian 14:10 nir: compile it in release mode and it gets a lot smaller
my plugin (not totally simple) is only about 600KB zipped
and that's optimized for speed, not size
kylehuff: this might help:
(sorry, people keep wanting me to help them — start and stop exams, grade their exams, fill out paperwork, etc… fussy people)
kylehuff 14:10 yeah, I just found the needed provider path.
no, don't be sorry. I'm just facing a bit of a learning curve, I know nothing about confluence
taxilian 14:10 thanks for looking into that
nir 14:10 @taxillian, thanks, release mode is 440KB zipped
taxilian 14:10 nir: if you need it smaller, there is an optimization for size that you can turn on
probably won't make it much smaller, but would a bit
the optimization flag isn't documented, but someone will get to it one of these days
linearray 14:10 -D MinSizeRel=1 ?
nir 14:10 @taxillian, actually optimizing for size resulted in a bigger dll :(
linearray 14:10 oh, i see it
tell me about the four build configurations
so that a certain someone can put it on the wiki
I get minsizerel... what exactly are debug, relwithdebinfo & release
hm, there's some info in CMakeCache.txt, I'll use that
taxilian 14:10 linearray: relwithdebinfo is a release mode that generates symbols
release is just normal release mode
nir: I doubt you did it right, then
but I could be wrong; I've just never seen that happen
linearray 14:10 ah, debug info... I was thinking about .deb packages
nir 14:10 I also doubt I did it right :) I have been lucky to forget my C++
coming back to vc++ and staring at the tons of configuration options make my head spin
taxilian 14:10 hehe
linearray 14:10 interesting... MinRelSize should not be much smaller than Release
first time in years I actually read the gcc manpage
taxilian 14:10 nor is optimizing for speed much faster
normal release is a compromise between the two
linearray 14:10 -Os enables all -O2 optimizations that do not typically increase code size. However, instructions are chosen for best performance, regardless of size.
that just can't work too well hehe
when would one want to use RelWithDebInfo?
taxilian 15:10 linearray: i want to use it all the time
so that I can save the debug symbols and if there is a crash and I can get the crash dump I can load it in visual studio and see the stack
linearray 15:10 very good
taxilian 15:10 linearray: awesome. I added docs for the FB_OPTIMIZE parameter to your build configurations page
kylehuff 15:10 I don't care for this soap interface...
getPage() returns an object that does not have a __setitem__ method, so you cannot modify the "content", and if you build your own page object using the info from the return of getPage, it complains about illegalArgumentException, but doesn't say which argument.
taxilian 15:10 fun
kylehuff 15:10 if I can get it accept the page, I'm pretty much done with the conversion to using the SOAP interface, and with minimal changes to the original file too
taxilian 15:10 cool
kylehuff 16:10 okay, ugh.. got it working... the problem was with the "id" and "parentId" fields being converted in python from an int to a long
FireBreathBot 16:10 2e909b5 by Kyle L. Huff: Converted to use SOAP interface
taxilian 16:10 kylehuff: so you think it'll just work if I pull that commit in and run it?
kylehuff 16:10 well, no, I have to fix something, in testing I messed up the parentId. but otherwise, yes...
taxilian 16:10 ahh okay
let me know =] if you've just solved the problem for me I'll be very happy indeed =]
kylehuff 16:10 okay, fixed.
it should run, everything validates for me, I just can't tell for sure because I get a permission denied. it appears to be working otherwise.
FireBreathBot 16:10 83f4138 by Kyle L. Huff: Converted to use SOAP interface
kylehuff 16:10 that converts it to using the SOAP interface, and also converts the contents to using convertWikiToStorageFormat
taxilian 16:10 cool
I'll pull and see
it's a bit of a time consuming process, though
kylehuff 16:10 due to the changes? or usually just time consuming all together?
taxilian 16:10 all together
it has to generate the doxygen stuff, then strip out unneeded files, then go through and do a replace for all URLs, then...
lol. and then fail because I forgot to install soappy
kylehuff 16:10 lol
taxilian 16:10 is the rest of the script
that calls that
kylehuff 16:10 in the instructions you linked earlier for soappy, it had a bunch of stuff I didn't do. I just installed from apt, and it worked fine.
taxilian 16:10 the most time consuming piece in there is hte little for loop that removes any documents that no other files reference
yeah, I used pip
easy_install would probably work as well
it is running again
there is a unicode error I had to track down earlier that may bite me again, though
hmm: <Fault soapenv:Server.userException: java.lang.IllegalArgumentException: <SOAPpy.Types.structType detail at 83425544>: {'hostname': 'shaoden', 'faultData': ''}>
kylehuff 16:10 I added a processing step if a UnicodeDecodeError exception is reached, were it attempts it again with an encode to utf-8
taxilian 16:10 that explains why I didn't hit that error, then =]
sounds like your solution was better than mine
yeah, I keep hitting that exception
not sure why
kylehuff 16:10 hmm, not sure whats up with the IllegalArgument... I *was* getting that, but fixed it by calling int() for the fields "id" and "parentId"
taxilian 16:10 hmm
it's htis that fails: page = self.rpc.storePage(self.token, npage)
kylehuff 16:10 hmm, I don't see it, but I only get let it attempt a few. let me try something.
taxilian 16:10 is there any way to get more useful information out of that call?
the exception?
kylehuff 16:10 not that I found... but, if you have it print out the variable npage when it reaches the exception, you could see if the "id" and parentId" fields are an integer, and not a long or a string
taxilian 17:10 hmm. something strange is occuring, at least
kylehuff 17:10 I have it processing on my laptop now, and so far I have not reached any errors regarding the input arguments, just permission errors
yeah, I can't make it fail on my machine, Linux, Python 2.6.6
taxilian 17:10 hmm. well, I gotta run; I guess I could give you the password for the automation, but there is a reasonably good chance that doing it from there would leave things in a non-optimal way
the document is: {'title': 'file logging.h', 'space': 'documentation', 'content': '<p><ac:macro ac:name="doxygen_init" /></p>\n<ac:macro ac:name="html-include"><ac:parameter ac:name="url"></ac:parameter></ac:macro>', 'version': 2, 'parentId': 1280310, 'id': 1281392}
well, will probably have to look at it more later
thanks for your help
I have a date with my wife tonight
so I need to run
kylehuff 17:10 np. have fun!
kylehuff 18:10 taxilian_away: I made some changes that should fix the issues you were having with the docs script. I was able to update via the script without incident. changes are here: