IRC Log Viewer » #firebreath » 2012-09-13

IRC Nick Time (GMT-7) Message
AndrewSt 07:09 Hi all! I have problem: I need to post message in forum, but when I login at http://www.firebreath.org in http://forum.firebreath.org/ shows that "You are not logged in.". When I try to login Forums http://forum.firebreath.org/account/signin/ it show 404 error
taxilian 07:09 AndrewSt: Yeah, the forum doesn't work. I need to take it down but I haven't had time to mess with it
use stackoverflow.com or the google group
TruthFairy 07:09 is there something that talks about the best or alternative way to combine plugin types in your app, i.e. multiple inheritance of the types vs. an instance of each type?
taxilian 07:09 what kinds of types are we talking about?
TruthFairy 07:09 AX, NP
taxilian 07:09 I'm still not following what you're after, I'm afraid
TruthFairy 07:09 np
the public log of this channel is like the forum, state wise?
taxilian 07:09 it works, AFAIK
but the guy who was maintaining it hasn't done anything in awhile
TruthFairy 07:09 it didn't the first an only time I looked
only known about fb a couple of weeks
taxilian 07:09 really? I just tried it and it's working for me
what didn't work?
TruthFairy 07:09 OK, so it works but only for search
taxilian 07:09 oh, yeah, there is no browse functionality
only search
TruthFairy 07:09 normally irc logs are allow you to just look at the collected logs by date, that threw me off, search only is fine
taxilian 07:09 would be nice to fix, but… :-/ someday maybe someone will get frustrated enough to do so =]
TruthFairy 07:09 well you seem to be heavily signed on to Atlassian stuff so maybe it'll come in an upgrade of theirs
I use Jira and Fisheye, but not much
taxilian 07:09 could be. the nice thing about atlassian is that it's free for open source projects and I hate it less than I hate the other options I've found =]
TruthFairy 07:09 i like that you can get the software at low cost and run it yourself
taxilian 07:09 yeah
I wish it didn't require so many system resources, though
TruthFairy 08:09 but you can tell they're evil
you are not evil, maybe because of xian or something
taxilian 08:09 heh. and what does "evil" mean in this case?
TruthFairy 08:09 evil = exploiter of programmers
taxilian 08:09 and what is it that makes atlassian evil?
(I've always been fairly impressed with them, other than their tendency to use java which rapes the resources on my box)
TruthFairy 08:09 really you must mean CPU, memory really isn't much of a problem these days, generally
I have my first SSD coming today
a real IDE SSD, not a thumb drive
taxilian 08:09 no I'm talking about both; I had to allocate 4GB of ram to the virtual machine running the firebreath stuff
out of 16GB total where none of my other stuff needs more than 2GB (and that's being generous) it's kinda annoying
TruthFairy 08:09 i've let virtualization migrate out to vendors like AWS, don'e wanna mess with it otherwise
taxilian 08:09 it's pretty easy to set up a vsphere box, which is free for smaller deployments, so that's what I use for my home server, which coincidentally is what firebreath's web resources run on
heh. that is a problem I've never had; I've always had a talent for using ram, I guess
TruthFairy 08:09 except when doing image work or sumthing
taxilian 08:09 this machine has 10 gig and 7 of it is currently in use
but I have a tendency to open everything all at once; multiple browsers, 5 vim windows, tweetdeck, IM/IRC, etc
reichi 08:09 firebreath is a real ram eater
at compile time
taxilian 08:09 during compilation, yes
it's the tradeoff; if we used fewer templates and didn't use boost that would go away, but those are the things that make things so easy to do
reichi 08:09 i'm not using templates ;)
taxilian 08:09 simplifies threading issues, allows JSAPIAuto to be so slick, etc
reichi 08:09 it's almot all about boost
TruthFairy 08:09 I've been running on Windows machine with 2G and I don't think it's even swapping
taxilian 08:09 if you're using FireBreath you're using templates =]
boost is a bit of a beast
I actually tried for a long time to not put it in firebreath, or to only use it for a few small select things
reichi 08:09 that would make you write your own smartpointer
basically
taxilian 08:09 eventually I realized that it was counter-productive. the stability went up a lot when I started using boost components instead of trying to do my own stuff
and yeah, I had my own smart pointer
reichi 08:09 hello refcount
where the hell are you from? ;)
taxilian 08:09 which turned out to be buggy, because writing your own smart pointer is shockingly difficult to do well; it seems like a simple project, but it's not
reichi 08:09 we have your own smartpointer class
sorry
our own
yeah
boost is fine
though it boosts the footprint, too ;)
taxilian 08:09 switching to use boost was pretty much the best decision I've made on this project, but good grief does it boost the foostprint, memory usage, compile time, ....
TruthFairy 08:09 if you try to use the whole magilla, yeah
taxilian 08:09 ahh, well
TruthFairy 08:09 most pkgs are header only, so in that sense zero foot print
reichi 08:09 firebreath plugins somehow become incredibly big
TruthFairy 08:09 you must be doing a lot of stuff
taxilian 08:09 actually in many ways the header-only ones are almost worst, because generally they are template based which increases compile time *and* usually ends up with multiple copies of all of the methods getting generated
reichi 08:09 well
TruthFairy 08:09 the one I'm building is only about 1M in release form
taxilian 08:09 FireBreath plugins aren't really that big when they are compiled optimized, particularly when you compress it
reichi 08:09 min has 2M
iirc
+e
taxilian 08:09 our binary download (which does use firebreath) is about 700K
reichi 08:09 well
i've implemented the HBBTV Core functionality
but i don't consider that "a lot of things"
TruthFairy 08:09 yeah, mine is just a mime type display utility so also not a resource hog
reichi 08:09 cloc says it's about 3k loc
i do consider 3000 lines of code reasonably small
TruthFairy 08:09 looks like a lot of stuff, but maybe not in the client
yeah 3 KSLOC is small
reichi 08:09 but nevertheless
TruthFairy 08:09 not big anyway
reichi 08:09 FB was an incredibly helpful framework for me :)
TruthFairy 08:09 the SSD I got is 30G and I think it's a few years old.
do newer ones have longer lifetimes?
or just high capacity?
reichi 08:09 well
i think it's VERY hard to say
Mine works fine for a year now
TruthFairy 08:09 this OCZ Vertex apparently lasts about 2 years
reichi 08:09 i've got an agility 3
my pc only has SATA II 300 so getting anything faster wasn't much of a point
TruthFairy 08:09 you may have about a year left on it
reichi 08:09 well
i tend to get intel next time
but long live SSD are quite expensive
you can definitly tell by just looking at the price ;)
TruthFairy 08:09 i expected the lifetime thing would be addressed by now
reichi 08:09 i don't know
i don't see why they would fail so soon
NAND flash i pretty robust
way more than it's said to be
taxilian 08:09 time will tell =
=]
TruthFairy 08:09 but the one i got is cheap, don't mind replacing every couple of years if it cuts disk ip time in half or better
*io
which vs. 7200RPM non-raid it should
i had 10K cheetahs I got in the late 90's they lasted about 6-7 years
(SCSI)
which I used in RAID 0. I expect this SSD to be faster, at least by a little
so my earlier question was is there a recommended way to use the ActiveX and NPAPI function in the same plugin?
taxilian 08:09 umm; you mean, have the same plugin work on both npapi and activex?
TruthFairy 08:09 y
taxilian 08:09 the recommended way is to write it using FireBreath
after that it just works
TruthFairy 08:09 lol
do you plan to maintain FB thru this decade?
taxilian 08:09 oh, probably
I use it myself at my job
and don't expect that to change anytime soon
lately I've been a little slower on the maintenance bit, though; time crunch :/
TruthFairy 08:09 do you have a SSD?
taxilian 08:09 in my laptop, yes
plan to get one in my imac when I replace it, but waiting for the next release cycle
TruthFairy 08:09 is it 2X or more faster on disk IO?
taxilian 08:09 *way* more
TruthFairy 08:09 so 4X maybe?
taxilian 08:09 huge noticeable difference; benchmarks showed it as much as 300x on some types of operations, slower ones were probably still more than 4x faster
(this is compared to my old 7200 rpm)
I may have the numbers somewhere still
TruthFairy 08:09 you mean 300% don't you?
taxilian 08:09 no
TruthFairy 08:09 not that a disk io job that took an hour now takes less than a minute
taxilian 08:09 that was only some types
let me find the numbers
TruthFairy 08:09 i'm actually going to use it with drupal not fb
taxilian 08:09 https://gist.github.com/1287513
well, fb would be sped up but not much; most of fb compilation is not disk i/o
random uncached writes on the SSD vs non-SSD : 220.48 MB/sec vs 1.21MB/sec
so maybe not 300x difference, but still pretty darn significant
granted, that's not really a common type of disk i/o
but still
TruthFairy 08:09 2 orders of mag
taxilian 08:09 on that gist the bottom is the same type of laptop but w/ a 7200 rpm standard drive
uncached sequential read is the smallest difference, and even that was about 1.5x
TruthFairy 08:09 actually based on that only looks like 2X unless I'm misreading
taxilian 08:09 which?
TruthFairy 08:09 the link you gave, the middle is the SSD right?
taxilian 08:09 right
and it depends on which benchmark you're looking at
the smallest difference is the uncached read (which makes sense, if you think about it), and that's only about 1.5x faster
uncached sequential read, that is
TruthFairy 08:09 looks like 2-3 or so depending
taxilian 08:09 but the uncached write, which is the biggest difference, is ~200x
TruthFairy 08:09 which at the people level would be dramatically faster than 7200
taxilian 08:09 so average it's guessing about 2-3
that's what it comes down to, yes =]
TruthFairy 08:09 so not super faster than the 2 cheetahs with raid0 but way cheaper, quieter
taxilian 08:09 main reason I like it on my laptop is it isn't susceptible to movement related data corruption
anyway, I need to get back to work
TruthFairy 08:09 never heard of it
taxilian 08:09 I'm too distractable =]
basically I had issues with my laptop hard drive having minor data problems due to the hard drive always being jostled like crazy (protecting it from my kids, I suppose)
TruthFairy 08:09 ah
jshanab_wcw 12:09 How do i return an image from the plugin to javascript so it can put it as a backgroud of an html element?
taxilian 12:09 jshanab_wcw: the only way I know of us using a data url
and that won't work on ie6
jshanab_wcw 12:09 I do not think we support ie6 :-)
Does that technique use HTTPService?
taxilian 12:09 no
a data url means base64 encoding the image and sending it as a text file
then you can construct a data url to put it in the img src
.lmgt data url
FireBreathBot 12:09 Let me google that for you: http://tinyurl.com/8cr9ea9
jshanab_wcw 12:09 Could I use the HTTPService, that looks like it would allow me to just put a uri in our html instead of replacing a src with a huge bas64 encode piece
taxilian 12:09 you could, though keep in mind that you may (likely will) get errors/warnings if you pull things from the http internal web server from an https page
jshanab_wcw 12:09 humm, that could be an issue. If I use the data url do i just set it to nothing and then use a javascript funtion to replace the src=attribute with the base64 encoded return value from my javascriptAPI call?
taxilian 12:09 basically
there is a size limit on some browsers as to how big that image can be, though
Gordon_LN 14:09 In the WIX installer where does the "UpgradeCode" GUID come from?
taxilian 14:09 I believe it is generated by fbgen
yep, that's it
Gordon_LN 14:09 I _think_ I need it to generate two GUIDs and then move it into the PluginConfig.cmake.
taxilian 14:09 would you care to explain why?
Gordon_LN 14:09 One for 32bit and one for 62bit
64 even
taxilian 14:09 are you planning on making two seperate wix installers?
different MSIs?
Gordon_LN 14:09 ...yes... unless there is a better way.
taxilian 14:09 Hmm. I guess that would be the easiest
if it were me, I'd probably build the msi without using firebreath at that point so I could include both .dlls in the MSI and only install the 64 bit version on machines that support it
Gordon_LN 14:09 In the past I am used to having two build fodlers "MyProject" + "MyProject64"
taxilian 14:09 but I guess the easiest way would be to let firebreath do it and since we can't combine 64 and 32 bit into a single solution your only option is that
Gordon_LN 14:09 Yes its a bit messy.
Windows + Mac or the issue as they need the ability to install 32+ 64 side by side.
*are
(its late here)
I also need to rename the "PLUGIN_NAME" dynamically to make life easier on OSX
if ( CMAKE_SIZEOF_VOID_P EQUAL 8 ) set ( ARCH_NAME "x86_64" ) else () set ( ARCH_NAME "i386" ) endif () set ( PLUGIN_NAME "${PLUGIN_NAME}_${ARCH_NAME}" )
(hmm that didn't paste well)
taxilian 14:09 hmm… I would probably not rename PLUGIN_NAME, I'd change the filename
FBSTRING_FileOutput
you could pretty easily change the wix template to pull a guid from your pluginconfig.cmake, though
Gordon_LN 14:09 I didn't know about "FBSTRING_FileOutput" I will change that.
Yes moving it to pluginconfig.cmake is what I was planning on doing and putting it in the if group then.
taxilian 14:09 go look at the .wxs file in your Win/WiX dir
that's the template
Gordon_LN 14:09 Yes, ok in which case I might update the Mac/projectdef.cmake to use it for the plugin name also...
it is currently ${releaseDirectory}/${projectName}.plugin
taxilian 14:09 ahh
Gordon_LN 14:09 (which is where I was coming from!)
Ok that looks like a plan and will be fairly clean also. thx.
(mind you a plan for another day) - I just got bit with "Chrome" only being 32bit on OSX world....
taxilian 14:09 yep
catery 15:09 anyone there?
taxilian 15:09 nope
though in general,
.ask
FireBreathBot 15:09 If you need help, just ask your question and wait for people to come back.
catery 15:09 so how do I launch an exe can i do that with NPAPI
taxilian 15:09 what platform?
catery 15:09 windows on ie, chrome, firefox
taxilian 15:09 okay, so 90% of your question should be reworded as "how do I launch a process using C++ in windows"
You'll probably want to use CreateProcessEx
catery 15:09 I would like the user to pick a chess analysis program like Rybka and then have my webpage use it to analyze a bunch of positions and show the result to the user and save it to a database
html pased
html based
if possible
taxilian 15:09 and is there a reason you can't just do this with javascript and localstorage?
catery 15:09 i thought you couldn't run an exe with javascript
taxilian 15:09 oh, I see, you want a specific program
honestly if I were you I'd try for a more server-side soluiton
where you submit the information to a server which would run it and then return
catery 15:09 Rybka is something the user owns not me so I don't think thats gonna be possible
taxilian 15:09 I see. well, you can probably do it, though you'll want to do some research on low integrity processes and what that means
because anything you launch from IE w/ UAC on will be in low integrity mode
catery 15:09 by probably do it you are referring to what again?
taxilian 15:09 you can probably do what you want to do
catery 15:09 with what precisely?
taxilian 15:09 use CreateProcessEx to launch the process
learn about low integrity mode, it may bite you
catery 15:09 you mean in the context of NPAPI call createprocessex
taxilian 15:09 I mean in your plugin, yes
be very, very careful -- you don't want to accidentally make it so that anyone can use your plugin to launch an arbitrary executable file
that would be a Very Very Bad Thing
catery 15:09 would it be possible to use activex for ie instead, npapi for just firefox and chrome,
taxilian 15:09 that would be what FireBreath does
ie doesn't support npapi
firebreath plugins work as both
catery 15:09 given the constraint that i need to run an exe is NPAPI even the best solution out of a bunch of cumbersome solutions, could Java work?
taxilian 15:09 I don't know if java would work or not; I'm not a java expert. possibly, if they have java installed
reichi 15:09 a plugin that just launches an exe
that sound's like a invitation to system intrusion
taxilian 15:09 seriously? I don't think what you're doing is a good idea. there is a lot of room for mistakes that could be huge security vulnerabilities
yeah, what he said
there are very good reasons that you can't launch an exe from javascript
be very careful
catery 15:09 but the end point i should takeaway is that it is impossible to launch an exe without asking the user to dl some plugin or something
reichi: so how do i accomplish what i want with the chess analysis
reichi 15:09 that's not the point
the point is, the moment the user has your plugin isntalled
anyone cann abuse it
to launch an exe file
and if you make a slight mistake
ANY exe
catery 15:09 OK given that I require the user to pick a chess analysis program like Rybka and then have my webpage use it to analyze a bunch of positions and show the result to the user and save it to a database, reichi how would you best make this happen
even though it is a horrible thing to have to try to implement how would you do it
reichi 15:09 find a lib i can link against
would be a the preferred way ;)
catery 15:09 Once i have the exe running i can communicate it with , read and write to stdin and stdout with javascript, is there a way?
reichi 15:09 that's even worse
you cannot let javascript talk directly to the exe
that's like totally opening a system
catery 15:09 how do i communicate with the exe then?
reichi 15:09 with an abstracted api
catery 15:09 thats based around stdin and stdout
reichi 15:09 yeah but not on the javascript side
but still
you plan is incredibly dangerous :/
taxilian 15:09 wow. yeah, don't *ever* expose that kind of low level thing to javascript
catery 15:09 so the javascript calls into the api imlpmented in npapi
taxilian 15:09 you want to expose as little as possible to the page
and do as much as you can in the plugin
that minimizes the possibility of security risks
catery 15:09 so what are the bad things that could theoretically happen
taxilian 15:09 worst case? all your personal data stolen and your computer trashed
viruses installed
reichi 15:09 1 mistake an one could execute arbitary console commansd
to their total liking
taxilian 15:09 etc
people could take over your computer
reichi 15:09 (it's not a theoretical threat)
catery 15:09 my web server you mean?
reichi 15:09 no
taxilian 15:09 no, we mean anyone who installs your plugin
reichi 15:09 the computer with the npapi plugin
taxilian 15:09 can you say lawsuits?
reichi 15:09 npapi runs on the client
catery 15:09 so i can't just say npapi plugin works on my site only
taxilian 15:09 nope
reichi 15:09 npapi is client side
it runs in the browser process
catery 15:09 i can't write if (url != mysite) { // do nothing } ???
taxilian 15:09 there are some options, yes, but they can be spoofed
reichi 15:09 is there no way to talk to those tools via socket?
that would help a lot
catery 15:09 http://en.wikipedia.org/wiki/Universal_Chess_Interface
they all use text based protocol
so if the user is selecting the exe to run, then bad things can't happen right?
because the user is the one that selects the exe right?
reichi 15:09 what if he selects the wron exe?
catery 15:09 but thats not gonna install a virus right?
reichi 15:09 i could be ANYTHING
but still
that's not the major issue
the major issue is that, if you do only ONE mistake
people could scan for your plugin on clients
and possibly launc ANY exe
catery 15:09 so what is the hypothetical one mistake i could make
reichi 15:09 there a tons...
taxilian 15:09 catery: hypothetically there are an infinite number of mistakes
catery 15:09 I haven't actually looked at how NPAPI worked
reichi 15:09 there's so much to do wrong
it's hard to do it right
taxilian 15:09 the fact that you ask as though you think that it would be unlikely frightens me more than I can say
our point is just that you are playing with fire and you don't seem to know it
npapi is just an application that you run in the browser
catery 15:09 you guys are making me depressed, making me feel like i have to do an old school win32 application
taxilian 15:09 let me ask you this
what kinds of problems can come from a poorly written win32 application?
catery 15:09 not the ones where some website can scan the plugin
reichi 15:09 bingo
;)
worst case, your applications crashes
catery 15:09 ok so how about i write the plugin (i was planning it to be open sourced anyway)
and you guys verify its correctness
reichi 15:09 no
i'm not gonna take that responsibilty
taxilian 15:09 an NPAPI plugin can do all the same things that a win32 app can, but then add on the problem that you have arbitrary code and input that you never see
catery 15:09 ok
reichi 15:09 it's really incredibly hard to write something like that properly
catery 15:09 plugin just launches an exe it should be really short in terms of lines of code how hard could it be to verify correctness
taxilian 15:09 catery: this is your project; don't try to foist it on someone else. We're not saying you can't do it, but you need to know exactly what you're getting into. It's not a simple problem
reichi 15:09 really good people fail on it regularly
taxilian 15:09 catery: it isn't our problem, we're not going to take responsibility for it
catery 15:09 is there maybe some way to know what the problems others have encountered
reichi 15:09 you could write a small program the user as to install
taxilian 15:09 .lmgt security problems npapi plugins
FireBreathBot 15:09 Let me google that for you: http://tinyurl.com/8fr4qtc
reichi 15:09 that opens a scoket
and use your npapi plugin to talk to that socket
taxilian 15:09 anyway, I'm swamped; need to get back to work. good luck
reichi 15:09 that would require the user to start a small program before your website can do it's work
catery 15:09 reichi eww i don't like how the user has to remember to do that
reichi 15:09 but take away a lot of risks (you still can do it totally wrong)
well
taxilian 15:09 it's possible to do what you want safely, but you need to be very very careful
reichi 15:09 you need to know what you#re doing :/
catery 15:09 if i tell you i'm making the plugin as small as possible
taxilian 15:09 and if you can do it any other way, that would be good to look into first
catery 15:09 will that help
taxilian 15:09 no, not really
reichi 15:09 it's basically the same problems as with sql injections
taxilian 15:09 rule of thumb: if there is another option, don't use a plugin
reichi 15:09 except you don't have any tool to assist you
you have to do all the filtering and sanity checks yourself
and be sure to do them all
and to them all correctly
catery 15:09 what if i require the user to use Chrome
and do things through a Chrome app
reichi 15:09 you don't understand the problem
as long as you launch any exe out of a browser
you have to make sure it cant be exploited
and requiring the user to install a special browser isn't any more comfortable than starting a program
catery 15:09 but a chrome app isn't vulnerable to outside websites right?
reichi 15:09 you can even tell the user the program is not running
can it talk to the webiste?
that's currently open?
catery 15:09 I'm not familiar with how chrome app work
reichi 15:09 i think they're javascript
catery 15:09 I'm not familiar with their abilities and constraints
reichi 15:09 i don't think you can launch a process out of a chrome app
and even if
if theres a connection from the website to the app
the issues remain
catery 15:09 why is that a problem
its my website
reichi 15:09 other webiste
catery 15:09 how would another website get involved
the chrome app hardcodes talking to http://mysite.com
reichi 15:09 if your website can talk to the app
well
i don't know chrome apps
i can't tell
but i'm almost certain you won't be able to launch extermal processes
catery 15:09 uggh every answer you gives me depresses me
taxilian 15:09 chrome can embed a npapi plugin in a CRX and have it only accessible to the extension
and that does make it less dangerous
catery 15:09 so why is making a chrome app more dangerous than just a standard win32 application
reichi 15:09 as is said
i don't know ANYTHING about it
so i'm out now
(it's almost midnight here)
catery 15:09 here is another crazy idea port http://www.stockfishchess.com/ to NACL
and run it in the browser
reichi 15:09 yes
taxilian 15:09 no idea
reichi 15:09 it's an own engine
use that
catery 15:09 the problem is that all the best engines are windows only
so you can't get them to NACL
and they are closed source
reichi 15:09 stockfish is a free uci chess engine
so why not take that?
catery 15:09 but not the best
users want to use the best chess engine
reichi 15:09 well
sorry
bu i really don't have the time
catery 15:09 anyways you guys have been helpful, gonna try to research chrome app time
reichi 15:09 stockfish would be perfekt
as you could integrate it completley into your npapi plugin
without even using uci
catery 15:09 with stockfish i wouldn't need npapi right?
reichi 15:09 sure
catery 15:09 if i have NACL no need for npapi?
taxilian 15:09 correct
reichi 15:09 ah, ok
a ready made solution
(but again with the drawback to only support one browser)
aynways, gn8
taxilian 16:09 good night
catery 16:09 taxilian tell me that the npapi idea is ok with no security problems
taxilian 16:09 the npapi idea is frightening and has the potential to be a security nightmare, but could probably be done correctly if you know what you're doing
catery 16:09 what about chrome app
can chrome app get to exe?
taxilian 16:09 no idea
I'm the plugin guy
catery 16:09 can you give me examples of what can go wrong
taxilian 16:09 no
I'm in the middle of a time-sensitive project
and I don't have time
.lmgt plugin security issues
FireBreathBot 16:09 Let me google that for you: http://tinyurl.com/8phwkb8
taxilian 16:09 that's a good place to start
bryan-_ 20:09 can firebreath plugins be compiled with visual studio 2012? I ask because I'm having trouble with prep2012.cmd. Have had great success with FB on mac, but I'm new to FB on windows. Thanks.
taxilian 20:09 i don't think there is a prep2012.cmd
but if the latest cmake supports it you could probably make one
bryan-_ 20:09 there isn't. I tried the prep2010.cmd
I have cmake 2.8.9
and tried changing 10 to 12 in prep2010.cmd
taxilian 20:09 type cmake /?
the correct generator string would be listed there
if it supports vs2012
bryan-_ 20:09 interesting, not listed
must be the culprit
taxilian 20:09 guess cmake doesn't support it yet
you could check a beta version
bryan-_ 20:09 latest nightly build doesn't list 2012 either
taxilian 20:09 huh
dunno. that's a cmake issue
bryan-_ 20:09 guess I'll have to get get vs2010
downloading visual studio express and following the instructions on the website
^your website
works, except for a link error, google tells me I need SP1
looks like I'm on my way
thanks.
catery 20:09 so how are npapi plugins installed? do users have to go through a bunch of clicks/
taxilian 20:09 download an msi, run it
catery 20:09 taxilian: so there is no way to avoid the user going through this unpleasant download process
of click click click
taxilian 20:09 um, no
catery 20:09 elaborate?
please?
taxilian 20:09 plugins must be registered in the windows system registry
you can't do that from a browser
nor can you put a .dll file in place
you need an installer
catery 21:09 taxilian: so one advantage of IE with activex is you don't need to download to use activex?
taxilian 21:09 dream on
activex is no different
ie does have .cab files which can be slightly less painful; similarly you can put a plugin inside a firefox or chrome extension. however, any of those only work on one browser
and I don't recommend it; I've had too many issues with it
bryan-_ 21:09 ok after updating vs2010, I'm making progress but running into problems with prep2010.cmd finding the MS DDK. I've downloaded and installed into c:\winddk like it says
CMake Error at cmake/Win.cmake:122 (message):
FireBreath on windows requires ATL/MFC libs to be installed. Please
download the Microsoft DDK and install the build environments in C:\WinDDK
Call Stack (most recent call first):
cmake/common.cmake:31 (include)
everything is in c:winddk
any ideas?
found a post on your forum about setting DDK_PATH. did that, still no go
I installed it into c:\winddk as the website said, but now I'm trying in C:\WinDDK\7600.16385.1\ , as the default install location (also what the website said, somewhat conflicting)
aaaaand, that worked
:)
taxilian 21:09 hurray!
always nice when I can ignore a problem long enough that it goes away
:-P
bryan-_ 21:09 LOL
taxilian 21:09 (was on the phone)
bryan-_ 21:09 no worries. I was kind of giving the play by play as I tried things…. bored, I guess
you put a great package out there. any support is icing on the cake
taxilian 21:09 glad it's helpful
bryan-_ 21:09 are you still developing it for facebook?
taxilian 21:09 no
not for over a year now
working for GradeCam now
bryan-_ 21:09 ah… that was about the time I wrote my first plugin
oh yeah, saw a reference to that on the site
wow, gradecam looks neat
taxilian 21:09 it is pretty neat
bryan-_ 21:09 I'm out.. thanks again