[02:56] <CIA-17> nbd * r4104 /branches/buildroot-ng/openwrt/include/verbose.mk: always set NO_TRACE_MAKE
[13:13] <florian___> nbd: I think that this solves what we were talking about yesterday : http://pastebin.ca/76959
[13:13] <florian___> at least it seems to work for me (tm)
[13:27] <nbd> uhm... that still doesn't allow the user to put the sitedefs into the Build/Configure/Sitedefs template
[13:27] <florian___> why ?
[13:28] <florian___> in a package makefile if you $(call Build/Confifgure/Sitedefes,your option) it should work ?
[13:28] <nbd> i wanted to have it look like this: http://pastebin.ca/76964
[13:29] <nbd> that's why i referred to preinst/postinst/etc.
[13:29] <florian___> ok
[13:32] <florian___> I don't know how to export variables defined like that
[13:32] <nbd> have you looked at the handling of preinst and postinst?
[13:33] <florian___> yes
[13:33] <florian___> you have to call a function anyway right ?
[13:34] <nbd> you do it like this:
[13:34] <nbd> if the template is defined, put it in a make variable
[13:34] <nbd> then export this variable to the environment
[13:34] <nbd> and write shell code that appends the contents of that variable to the site.def inside the PKG_BUILD_DIR
[13:35] <florian___> that's tricky
[13:35] <nbd> that's how it works with preinst
[13:35] <florian___> is this what you are doing with : $(1)_COMMANDS += if [ -n "$$$$$$$$pkg_$(subst .,_,$(subst -,_,$(1)))_$(2)" ]; then echo "$$$$$$$$pkg_$(subst .,_,$(subst -,_,$(1)))_$(2)" > $(2); fi;
[13:35] <nbd> ignore that
[13:35] <nbd> at least ignore the subst part
[13:36] <nbd> it's only necessary if you need to export several versions for different packages in one makefile
[13:36] <florian___> ok
[13:36] <nbd> and the lots of $$$$$$$ are only necessary because i'm using an extra template called from a template
[13:36] <florian___> how do I check if a template is defined ?
[13:37] <nbd> you only need half of them if you do it in the Configure/default template directly
[13:37] <nbd> in this case, use shell code to see if the variable is empty
[13:37] <nbd> you can't use make's ifdef on multi-line commands
[15:39] <CIA-17> nico * r4105 / (2 files in 2 dirs): move libpthread in since it's part of uclibc, standardize Makefile.
[16:03] <nbd> .
[16:57] <CIA-17> florian * r4106 /branches/buildroot-ng/openwrt/toolchain/gcc/Config.version: Avoid menuconfig warnings
[17:03] <florian___> nbd: how is the squashfs+jffs2 partitioning done for bcm boards ?
[17:03] <florian___> because on alchemy, I can have a squashfs partition, but the remaining space is not jffs2, at least for the moment
[17:10] <CIA-17> florian * r4107 /branches/buildroot-ng/openwrt/Config.in: Disable squashfs for au1000 target since it does not allow the user to use the whole flash space
[17:12] <CIA-17> florian * r4108 /branches/buildroot-ng/openwrt/target/ (image/au1000/Makefile linux/au1000-2.6/Makefile): Disable the use of lzma since the alchemy images are zlib compressed and there is a lot of space. Fix makefile error when "touching" zImage.flash.srec
[17:50] <CIA-17> florian * r4109 /branches/buildroot-ng/openwrt/target/Config.in: Fancy name for au1000 target
[19:27] <florian___> nbd: ping ?
[19:28] <florian___> {Nico}: have you tested the alchemy port ?
[19:29] <db90h> what's this reference to alchemy? I though that was a Sveasoft firmware ;p
[19:30] <Kaloz> alchemy = amd mips cpu
[19:30] <db90h> ah
[19:30] Action: db90h eyes are opened
[19:31] <Kaloz> db90h: http://www.amd.com/us-en/ConnectivitySolutions/ProductInformation/0,,50_2330_6625,00.html
[19:33] <db90h> cool, thanks
[20:04] <florian___> a really good box
[20:05] <Kaloz> florian___: stil lcan't understand why do you use zlib when lzma is working
[20:05] <Kaloz> sounds stupid
[20:05] <florian___> because I don't get lzma working actually
[20:09] <nbd> florian___: i might get access to a meshcube soon
[20:09] <nbd> then i can test the alchemy port
[20:53] <florian___> nbd: great, a friend of mine was so impressed that he is thinking of building them :)
[20:54] <florian___> Kaloz: if you think you can help me fixing the lzma stuff, I take it ;)
[21:40] <florian___> Kaloz: it's ok, looks like I got the lzma-loader working :)
[22:26] <CIA-17> nico * r4110 /packages/utils/ (6 files in 3 dirs): port strace to -ng.
[22:26] <CIA-17> florian * r4111 /: Remove libpthread from packages since it has now moved to buildroot-ng
[22:27] <florian___> {Nico}: ping ?
[22:27] <{Nico}> florian___: pong
[22:28] <florian___> {Nico}: your strace Makefile can be simplified
[22:28] <florian___> in define Build/Configure
[22:28] <florian___> just do a $(call Build/Configure/Default)
[22:28] <florian___> it will do what you have written
[22:30] <{Nico}> florian___: right
[22:31] <florian___> by the way I am fixing the lzma decompression for alchemy, so it should now work out of the box
[22:31] <{Nico}> florian___: i'll try it later
[22:32] <{Nico}> florian___: btw, when should your talk take place?
[22:32] <florian___> on thursday
[22:33] <CIA-17> nico * r4112 /packages/net/tcpdump/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port tcpdump to -ng.
[22:33] <florian___> will you be there ?
[22:34] <{Nico}> florian___: i'll try, where is it?
[22:35] <florian___> vandoeuvre les nancy, rmll 2006 http://www.rmll.info/theme_49
[22:36] <Kaloz> i paste this one here, too http://basshunter.m0o.eu/anna/index_en.php
[22:36] <Kaloz> :P
[22:36] <Kaloz> it kicks hard
[22:36] <{Nico}> florian___: what time on thursday?
[22:37] <florian___> 10h45 to 12h15
[22:39] <{Nico}> florian___: i'll check my schedule for next week and ping you back tomorrow
[22:39] <florian___> great
[22:39] <{Nico}> i have to gow now, cu
[22:39] <florian___> see you
[22:55] <jr-> kaloz. there's "#ifndef IXDPG425" in IxEthMii.c (part of intel ethernet driver). i just can't find out where IXDPG425 gets defined so I could change that line so that part of code is also skipped for FGS3. any hints?
[22:55] <Kaloz> no idea
[23:03] <florian___> jr-: is not this the contraction of CONFIG_IXDPG425 ?
[23:09] <jr-> florian. i don't know. i'll try it later.
[23:10] <florian___> jr-: what gives you a grep -R "IXDPG425" ?
[23:15] <jr-> florian. plenty of stuff, but always CONFIG_ARCH_IXDPG425 and MACH_IXDPG425 not plain IXDPG425. Only match with plain IXDPG425 is 'bool "IXDPG425"' in arch/arm/mach-ixp4xx/Kconfig.
[23:18] <florian___> ok, from what I could have experience, CONFIG_ARCH_IXDPG425 is contracted to IXDPG425 in drivers/ dir
[23:26] <jr-> ok
[00:00] --- Mon Jul 3 2006
[13:13] <florian___> nbd: I think that this solves what we were talking about yesterday : http://pastebin.ca/76959
[13:13] <florian___> at least it seems to work for me (tm)
[13:27] <nbd> uhm... that still doesn't allow the user to put the sitedefs into the Build/Configure/Sitedefs template
[13:27] <florian___> why ?
[13:28] <florian___> in a package makefile if you $(call Build/Confifgure/Sitedefes,your option) it should work ?
[13:28] <nbd> i wanted to have it look like this: http://pastebin.ca/76964
[13:29] <nbd> that's why i referred to preinst/postinst/etc.
[13:29] <florian___> ok
[13:32] <florian___> I don't know how to export variables defined like that
[13:32] <nbd> have you looked at the handling of preinst and postinst?
[13:33] <florian___> yes
[13:33] <florian___> you have to call a function anyway right ?
[13:34] <nbd> you do it like this:
[13:34] <nbd> if the template is defined, put it in a make variable
[13:34] <nbd> then export this variable to the environment
[13:34] <nbd> and write shell code that appends the contents of that variable to the site.def inside the PKG_BUILD_DIR
[13:35] <florian___> that's tricky
[13:35] <nbd> that's how it works with preinst
[13:35] <florian___> is this what you are doing with : $(1)_COMMANDS += if [ -n "$$$$$$$$pkg_$(subst .,_,$(subst -,_,$(1)))_$(2)" ]; then echo "$$$$$$$$pkg_$(subst .,_,$(subst -,_,$(1)))_$(2)" > $(2); fi;
[13:35] <nbd> ignore that
[13:35] <nbd> at least ignore the subst part
[13:36] <nbd> it's only necessary if you need to export several versions for different packages in one makefile
[13:36] <florian___> ok
[13:36] <nbd> and the lots of $$$$$$$ are only necessary because i'm using an extra template called from a template
[13:36] <florian___> how do I check if a template is defined ?
[13:37] <nbd> you only need half of them if you do it in the Configure/default template directly
[13:37] <nbd> in this case, use shell code to see if the variable is empty
[13:37] <nbd> you can't use make's ifdef on multi-line commands
[15:39] <CIA-17> nico * r4105 / (2 files in 2 dirs): move libpthread in since it's part of uclibc, standardize Makefile.
[16:03] <nbd> .
[16:57] <CIA-17> florian * r4106 /branches/buildroot-ng/openwrt/toolchain/gcc/Config.version: Avoid menuconfig warnings
[17:03] <florian___> nbd: how is the squashfs+jffs2 partitioning done for bcm boards ?
[17:03] <florian___> because on alchemy, I can have a squashfs partition, but the remaining space is not jffs2, at least for the moment
[17:10] <CIA-17> florian * r4107 /branches/buildroot-ng/openwrt/Config.in: Disable squashfs for au1000 target since it does not allow the user to use the whole flash space
[17:12] <CIA-17> florian * r4108 /branches/buildroot-ng/openwrt/target/ (image/au1000/Makefile linux/au1000-2.6/Makefile): Disable the use of lzma since the alchemy images are zlib compressed and there is a lot of space. Fix makefile error when "touching" zImage.flash.srec
[17:50] <CIA-17> florian * r4109 /branches/buildroot-ng/openwrt/target/Config.in: Fancy name for au1000 target
[19:27] <florian___> nbd: ping ?
[19:28] <florian___> {Nico}: have you tested the alchemy port ?
[19:29] <db90h> what's this reference to alchemy? I though that was a Sveasoft firmware ;p
[19:30] <Kaloz> alchemy = amd mips cpu
[19:30] <db90h> ah
[19:30] Action: db90h eyes are opened
[19:31] <Kaloz> db90h: http://www.amd.com/us-en/ConnectivitySolutions/ProductInformation/0,,50_2330_6625,00.html
[19:33] <db90h> cool, thanks
[20:04] <florian___> a really good box
[20:05] <Kaloz> florian___: stil lcan't understand why do you use zlib when lzma is working
[20:05] <Kaloz> sounds stupid
[20:05] <florian___> because I don't get lzma working actually
[20:09] <nbd> florian___: i might get access to a meshcube soon
[20:09] <nbd> then i can test the alchemy port
[20:53] <florian___> nbd: great, a friend of mine was so impressed that he is thinking of building them :)
[20:54] <florian___> Kaloz: if you think you can help me fixing the lzma stuff, I take it ;)
[21:40] <florian___> Kaloz: it's ok, looks like I got the lzma-loader working :)
[22:26] <CIA-17> nico * r4110 /packages/utils/ (6 files in 3 dirs): port strace to -ng.
[22:26] <CIA-17> florian * r4111 /: Remove libpthread from packages since it has now moved to buildroot-ng
[22:27] <florian___> {Nico}: ping ?
[22:27] <{Nico}> florian___: pong
[22:28] <florian___> {Nico}: your strace Makefile can be simplified
[22:28] <florian___> in define Build/Configure
[22:28] <florian___> just do a $(call Build/Configure/Default)
[22:28] <florian___> it will do what you have written
[22:30] <{Nico}> florian___: right
[22:31] <florian___> by the way I am fixing the lzma decompression for alchemy, so it should now work out of the box
[22:31] <{Nico}> florian___: i'll try it later
[22:32] <{Nico}> florian___: btw, when should your talk take place?
[22:32] <florian___> on thursday
[22:33] <CIA-17> nico * r4112 /packages/net/tcpdump/ (. Config.in Makefile Makefile ipkg/ patches/ patches/): port tcpdump to -ng.
[22:33] <florian___> will you be there ?
[22:34] <{Nico}> florian___: i'll try, where is it?
[22:35] <florian___> vandoeuvre les nancy, rmll 2006 http://www.rmll.info/theme_49
[22:36] <Kaloz> i paste this one here, too http://basshunter.m0o.eu/anna/index_en.php
[22:36] <Kaloz> :P
[22:36] <Kaloz> it kicks hard
[22:36] <{Nico}> florian___: what time on thursday?
[22:37] <florian___> 10h45 to 12h15
[22:39] <{Nico}> florian___: i'll check my schedule for next week and ping you back tomorrow
[22:39] <florian___> great
[22:39] <{Nico}> i have to gow now, cu
[22:39] <florian___> see you
[22:55] <jr-> kaloz. there's "#ifndef IXDPG425" in IxEthMii.c (part of intel ethernet driver). i just can't find out where IXDPG425 gets defined so I could change that line so that part of code is also skipped for FGS3. any hints?
[22:55] <Kaloz> no idea
[23:03] <florian___> jr-: is not this the contraction of CONFIG_IXDPG425 ?
[23:09] <jr-> florian. i don't know. i'll try it later.
[23:10] <florian___> jr-: what gives you a grep -R "IXDPG425" ?
[23:15] <jr-> florian. plenty of stuff, but always CONFIG_ARCH_IXDPG425 and MACH_IXDPG425 not plain IXDPG425. Only match with plain IXDPG425 is 'bool "IXDPG425"' in arch/arm/mach-ixp4xx/Kconfig.
[23:18] <florian___> ok, from what I could have experience, CONFIG_ARCH_IXDPG425 is contracted to IXDPG425 in drivers/ dir
[23:26] <jr-> ok
[00:00] --- Mon Jul 3 2006