| GitHub192 | [migen] sbourdeauducq pushed 1 new commit to master: http://git.io/ktSoRg | 13:55 |
|---|---|---|
| GitHub192 | migen/master cc5ff7a Sebastien Bourdeauducq: add tree visualizer | 13:55 |
| mumptai | moin | 16:47 |
| mumptai | lekernel, where do i send that patch for mibuild? | 16:49 |
| lekernel | devel at lists.milkymist.org | 16:49 |
| mumptai | done | 17:02 |
| lekernel | thx | 17:09 |
| mumptai | btw, is there some development plan for mibuild/migen? | 17:12 |
| lekernel | right now nengel|work is adding direct synthesis (edif) output | 17:12 |
| lekernel | for xilinx | 17:12 |
| lekernel | other features I'm thinking about are having a proper dataflow system (with support for automatic actor sharing, scheduling, etc.) | 17:13 |
| lekernel | which would be great for DSP applications | 17:13 |
| mumptai | fsm with (latex)visualization would also be neat | 17:14 |
| lekernel | another thing is a preprocessor that cleans up the ugly FHDL syntax hacks | 17:15 |
| lekernel | large bits of FHDL could then be written by mixing that new language with python, similar to what e.g. bison does with C | 17:15 |
| mumptai | actually the most annoying part of the experience was that, every repository besides the milkymist-ng on seemed to rely on a different version of migen/mibuild | 17:15 |
| davidc__ | lekernel: there's a cool python-macro hack | 17:17 |
| davidc__ | lekernel: that uses import hooks to re-write the AST | 17:17 |
| lekernel | well, rhino-gateware is also regularly tested and maintained. others... well, their owners need to catch up ;) | 17:18 |
| davidc__ | lekernel: so, it still needs to pass the tokenizer / AST building; but no need for a preprocessor | 17:18 |
| lekernel | davidc__, but that supposes that the python parser managed to turn the code into a python ast in the first place, no? | 17:20 |
| mumptai | acutally telling the revision it was known to work with would help, but i'm totally aware that this is not a migen/mibuild issue ;) | 17:20 |
| davidc__ | lekernel: yes; it does - but the python AST is pretty expressive; hopefully expressive enough for a FHDL DHSL | 17:20 |
| davidc__ | *DSL | 17:20 |
| lekernel | mumptai, you can look at the last commit date and then rollback migen/mibuild to that date | 17:24 |
| mumptai | i know ;) | 17:24 |
| mumptai | and the first commit date, and check everything with relevant changes in between | 17:26 |
| GitHub180 | [mibuild] sbourdeauducq pushed 1 new commit to master: http://git.io/FWadRA | 22:10 |
| GitHub180 | mibuild/master 9d531ba user: Fix missing string replace. Added support for 32-bit ISE if 64-bit version is missing on 64-bit system. | 22:10 |
| GitHub22 | [migen] sbourdeauducq pushed 5 new commits to master: http://git.io/Oiz0AA | 22:18 |
| GitHub22 | migen/master 7a24317 Sebastien Bourdeauducq: fhdl/namer: new namer with explicit tree | 22:18 |
| GitHub22 | migen/master ceddd8a Sebastien Bourdeauducq: treeviz: improve layout of unbalanced trees | 22:18 |
| GitHub22 | migen/master 0e36931 Sebastien Bourdeauducq: treeviz: support multiline labels | 22:18 |
| GitHub154 | [mibuild] sbourdeauducq pushed 1 new commit to master: http://git.io/l-hkJw | 22:18 |
| GitHub154 | mibuild/master 88611be Sebastien Bourdeauducq: xilinx_ise: cleanup | 22:18 |
| --- Thu Aug 8 2013 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!