#milkymist IRC log for Tuesday, 2013-03-12

Action: kiwichris is still looking into gcc-4.5.3 building with newlib-2.0.003:47
kiwichrisBuilt an RTEMS tool set ?.07:56
kiwichris$ ~/development/rtems/4.11/bin/lm32-rtems4.11-gcc --version07:56
kiwichrislm32-rtems4.11-gcc (GCC) 4.5.3 20110428 (RTEMS 4.11-RSB(a77083a3cb1f1ab05aec307d221aba970074ce92-modified)-1,gcc-4.5.3/newlib-2.0.0)07:56
kiwichrisThe issue is gcc-4.5.3 breaks with a jobs option on make07:57
Fallenoucome on ...08:42
Fallenouit was only the -j option ?08:42
kiwichrisYeah that was it09:17
kiwichrisBuilds on MacOS, FreeBSD and Fedora09:19
lekernelhm, works for me09:20
lekernelI just built it yesterday with -j 809:20
lekernelit was lm32-elf though09:21
kiwichrisThe issue is with newlib headers being copied to target location.09:21
kiwichrisThere are specific RTEMS headers needed for the build09:21
kiwichrisTalking to Sebastian in the RTEMS project ? he said it can break.09:22
Fallenoukiwichris: is there a way to forbid use of -j for a specific target ?09:30
Fallenouin order to prevent users from falling into that trap09:31
lekernelbetter fix that bug in the first place, no?09:34
kiwichrisIf building yourself, none that I know of. In the RTEMS Source Builder tool it is disabled on all hosts so should be fine.09:36
kiwichrisOn Windows (MinGW) you need disable it because there are bugs in the MSYS's DLL in handling some of make's pids09:37
kiwichrisI think the bugs have been fixed in later gcc releases09:37
Fallenouyou mean those that don't work anyway on lm32 ? :)09:39
kiwichrisI suspect this -j issue in 4.5.3 is fixed in later gcc versions. As you know I cannot build
kiwichrisMy current focus is on a stable lm32 tool set09:41
Fallenougreat :)09:48
lekernelyay, urjtag patch merged10:32
larscparty! \o/ /o/ \o\ \o/ ;)10:37
Fallenouwe are getting good at getting patch merged !10:52
GitHub49[migen] sbourdeauducq pushed 3 new commits to master: http://git.io/FvMQDg14:46
GitHub49migen/master 12158ce Sebastien Bourdeauducq: fhdl/tracer: recognize LOAD_DEREF opcode14:46
GitHub49migen/master 7e2581b Sebastien Bourdeauducq: fhdl/tracer: recognize CALL_FUNCTION_VAR opcode14:46
GitHub49migen/master 04df076 Sebastien Bourdeauducq: bank: automatic register naming14:46
GitHub28[milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/uHYpBQ14:49
GitHub28milkymist-ng/master a23df42 Sebastien Bourdeauducq: Use automatic register naming14:49
GitHub90[milkymist-ng] sbourdeauducq pushed 1 new commit to master: http://git.io/zedaBw15:14
GitHub90milkymist-ng/master 1e7783a Sebastien Bourdeauducq: build.py: use implicit get_fragment15:14
GitHub87[mibuild] sbourdeauducq pushed 1 new commit to master: http://git.io/6l3p4g15:15
GitHub87mibuild/master c06a821 Sebastien Bourdeauducq: generic_platform: implicit get_fragment15:15
GitHub18[migen] sbourdeauducq pushed 5 new commits to master: http://git.io/gI_qew16:15
GitHub18migen/master ecfe164 Sebastien Bourdeauducq: fhdl/verilog: implicit get_fragment16:15
GitHub18migen/master 907bfa8 Sebastien Bourdeauducq: examples/basic: use new APIs16:15
GitHub18migen/master d92ca43 Sebastien Bourdeauducq: vpi: make it work by default on Arch16:15
larscwondering why the laptop got so hot, some leftover vvp spwaned by migen was still running at 100%, although the migen process was long gone19:48
wpwrakthat's a nice thing about laptops: it's easy to notice when you have a runaway process21:27
azonenbergwpwrak: lol21:38
balrogit sorta feels http://milkymist.org has an excessive amount of 404s23:16
--- Wed Mar 13 201300:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!