lekernel changed the topic of #m-labs to: Mixxeo, Migen, MiSoC & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
xiangfu has joined #m-labs
xiangfu has quit [Ping timeout: 264 seconds]
xiangfu has joined #m-labs
xiangfu has quit [Ping timeout: 240 seconds]
xiangfu has joined #m-labs
xiangfu has quit [Remote host closed the connection]
xiangfu has joined #m-labs
rjo_ is now known as rjo
stekern_ is now known as stekern
sh4rm4 has quit [Remote host closed the connection]
sh4rm4 has joined #m-labs
sh[4]rm4 has joined #m-labs
sh4rm4 has quit [Ping timeout: 252 seconds]
aeris has quit [Quit: en a pas]
aeris has joined #m-labs
sh[4]rm4 is now known as sh4rm4
xiangfu has quit [Ping timeout: 245 seconds]
xiangfu has joined #m-labs
xiangfu has quit [Remote host closed the connection]
sb0 has joined #m-labs
Alain_ has joined #m-labs
_florent_ has joined #m-labs
<_florent_> hi
<_florent_> I'm working on the VIVADO support for Mibuild
<_florent_> it's working fine with SimpleSoC and 2014.1 version on the KC705
<_florent_> but I have a question related to Python
<_florent_> here is my platform file:
<_florent_> for now I have to comment my import to use ISE or VIVADO
<_florent_> but I'd like to be able to pass a parameter to the Platform class to choose the toolchain
<LoneTech> yay! python questions I might be able to figure out :)
<ysionneau> \o/
<_florent_> but since it does not seems to be possible to do dynamic inheritance, I'm a stuck...
<_florent_> does someone has idea or a workaround for that?
<LoneTech> dynamic inheritance is quite possible; just create the class dynamically. i.e. you could make a metaclass
<_florent_> in fact here I want to be able to do this:
<_florent_> Platform(toolchain="ise") in make.py --> return Platform that inherit from XilinxISEPlatform
<_florent_> Platform(toolchain="vivado") in make.py --> return Platform that inherit from XilinxVIVADOPlatform
<_florent_> LoneTech : thanks, I'm looking at that
<LoneTech> so a structure like: def Platform(toolchain): class Platform(toolchain): ... return Platform could do that
<LoneTech> i.e. a fresh class produced when that function is called
<LoneTech> memoize it if you want to reuse it for type checks or something
<LoneTech> though I'm not sure this is very elegant from either OO or performance perspectives, it doesn't feel like the most performance critical part
<_florent_> ah yes indeed if Plaform is a function it's easier :)
<sb0> that's a lot of duplicated code
<sb0> moment
<LoneTech> _florent_: I meant moving the entire class definition into a function
<_florent_> ah yes ok I see
<sb0> something like this
<sb0> you may want to have a "toolchain" string argument instead
<sb0> but you can't inherit from Platform or deal with its class attributes, not sure if that's a problem
<sb0> use args/kwargs
<_florent_> ok
<_florent_> implementation seems to be fast
<sb0> XilinxPlatform -> xilinx_platform
<sb0> XilinxVIVADOPlatform -> XilinxVivadoPlatform
<_florent_> less than 2 minutes for SimpleSoC
<sb0> yeah, once in a blue moon xilinx improves their software
<_florent_> last time I tested, It crashed before the end...
<sb0> they don't only push new 'features' like flexlm
<LoneTech> flexlm is an old misfeature
<sb0> they had their own system until ise 11 (iirc)
<sb0> which wasn't as bad as flexlm
<LoneTech> they just had a short lapse of insanity when webpack didn't require it
<sb0> it's kinda puzzling to me wtf are people thinking when they spend money on flexlm for it to cripple their software
<sb0> *plus* the tech support issues it generates
<larsc> 'my manager told me to do this, so I'm going to do it' ;)
<LoneTech> somebody sold them on a greed scheme. I've even encountered it in a simple graphics adaptor, making it impossible to support.
Alain_ has quit [Quit: ChatZilla 0.9.90.1 [Firefox 29.0.1/20140506152807]]
kyak has quit []
kyak has joined #m-labs
_florent_ has quit [Quit: Page closed]
kyak has quit [Ping timeout: 265 seconds]
kyak has joined #m-labs
sb0 has quit [Quit: Leaving]
stekern has quit [Ping timeout: 252 seconds]
stekern has joined #m-labs