[Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
since this is the first time i actually built an image with any of the downloadable TI stuff, can i just sanity check what i did? it seems to have worked, but i'd rather have someone else verify i did it correctly. first step in downloading the contents into my dl dir: $ ls -l ti_* -rw-r--r-- 1 rpjday rpjday 77373969 Jun 2 02:44 ti_cgt_c6000_6.1.14_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday 77383796 Jun 1 18:45 ti_cgt_c6000_6.1.9_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday0 Jun 2 04:40 ti_cgt_c6000_6.1.9_setup_linux_x86.bin.lock $ it wasn't immediately clear whether or not i should grab the 6.1.14 binary, but as there was a recipe for it, i took it, just in case. next, i failed to notice this in recipes/ti/ti-cgt6x.inc: SRC_URI = http://install.source.dir.local/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin; obviously, the local fetch is going to fail because of that URL, so i just changed it to: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin i would think that either a comment can be added so that users know to do that, or why not just change that line? actually, it can be tightened up a bit more, in that this: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin could be replaced with this, could it not? BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin SRC_URI = file://${DL_DIR}/${BINFILE};name=cgt6xbin just seems cleaner to avoid the duplication. finally, the bitbake of ti-cgt6x seems to have worked, ending in: NOTE: Running task 720 of 723 (ID: 15, /home/rpjday/oe/openembedded/recipes/ti/ti-cgt6x_6.1.9.bb, do_package_stage_all) NOTE: Running task 721 of 723 (ID: 6, /home/rpjday/oe/openembedded/recipes/ti/ti-cgt6x_6.1.9.bb, do_build) NOTE: Running task 722 of 723 (ID: 17, /home/rpjday/oe/openembedded/recipes/ti/ti-cgt6x_6.1.9.bb, do_rm_work) NOTE: Running task 723 of 723 (ID: 4, /home/rpjday/oe/openembedded/recipes/ti/ti-cgt6x_6.1.9.bb, do_rm_work_all) NOTE: Tasks Summary: Attempted 723 tasks of which 704 didn't need to be rerun and 0 failed. but during the bake, i saw output like: Where do you want to install C6000 Code Generation Tools? [/opt/TI/TI_CGT_C6000_6.1.9] Install Error: You do not have permission to write to /opt. Where do you want to install C6000 Code Generation Tools? [/opt/TI/TI_CGT_C6000_6.1.9] Installing C6000 Code Generation Tools... ... snip ... NOTE: Couldn't find shared library provider for libstdc++-libc6.2-2.so.3 but since it appears to have completed, i'm going to assume it was successful. rday -- Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
On Wed, Jun 02, 2010 at 06:08:22AM -0400, Robert P. J. Day wrote: since this is the first time i actually built an image with any of the downloadable TI stuff, can i just sanity check what i did? it seems to have worked, but i'd rather have someone else verify i did it correctly. first step in downloading the contents into my dl dir: $ ls -l ti_* -rw-r--r-- 1 rpjday rpjday 77373969 Jun 2 02:44 ti_cgt_c6000_6.1.14_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday 77383796 Jun 1 18:45 ti_cgt_c6000_6.1.9_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday0 Jun 2 04:40 ti_cgt_c6000_6.1.9_setup_linux_x86.bin.lock $ it wasn't immediately clear whether or not i should grab the 6.1.14 binary, but as there was a recipe for it, i took it, just in case. next, i failed to notice this in recipes/ti/ti-cgt6x.inc: SRC_URI = http://install.source.dir.local/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin; obviously, the local fetch is going to fail because of that URL, so i just changed it to: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin i would think that either a comment can be added so that users know to do that, or why not just change that line? actually, it can be tightened up a bit more, in that this: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin could be replaced with this, could it not? BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin SRC_URI = file://${DL_DIR}/${BINFILE};name=cgt6xbin just seems cleaner to avoid the duplication. Without testing it myself, here is what I expect: You don't have to change this if you download it to ${DL_DIR} yourself as asked in recipe. If there is archive already in download dir, bitbake won't try to download it from http link and will use that already downloaded archive and check it's checksums (which is good especially as you can download a bit different archive when downloading manually). After replacing it with file:// it won't be checked. Only one part I'm not so sure about is that all my downloads in DL_DIR has also file.md5 (ie ti_cgt_c6000_6.1.9_setup_linux_x86.bin.md5), you can easily create it with md5sum, but IIRC I've seen (at least with some bitbake version) that downloaded archive was ignored when it didn't have .md5 neighbour. If that's the case, then the recipe download instrcutions should be improved. but since it appears to have completed, i'm going to assume it was successful. I never tried to build those.. so I cannot comment on this issue. Regards, -- uin:136542059jid:martin.ja...@gmail.com Jansa Martin sip:jama...@voip.wengo.fr JaMa ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
On Wed, 2 Jun 2010, Martin Jansa wrote: On Wed, Jun 02, 2010 at 06:08:22AM -0400, Robert P. J. Day wrote: since this is the first time i actually built an image with any of the downloadable TI stuff, can i just sanity check what i did? it seems to have worked, but i'd rather have someone else verify i did it correctly. first step in downloading the contents into my dl dir: $ ls -l ti_* -rw-r--r-- 1 rpjday rpjday 77373969 Jun 2 02:44 ti_cgt_c6000_6.1.14_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday 77383796 Jun 1 18:45 ti_cgt_c6000_6.1.9_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday0 Jun 2 04:40 ti_cgt_c6000_6.1.9_setup_linux_x86.bin.lock $ it wasn't immediately clear whether or not i should grab the 6.1.14 binary, but as there was a recipe for it, i took it, just in case. next, i failed to notice this in recipes/ti/ti-cgt6x.inc: SRC_URI = http://install.source.dir.local/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin; obviously, the local fetch is going to fail because of that URL, so i just changed it to: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin i would think that either a comment can be added so that users know to do that, or why not just change that line? actually, it can be tightened up a bit more, in that this: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin could be replaced with this, could it not? BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin SRC_URI = file://${DL_DIR}/${BINFILE};name=cgt6xbin just seems cleaner to avoid the duplication. Without testing it myself, here is what I expect: You don't have to change this if you download it to ${DL_DIR} yourself as asked in recipe. If there is archive already in download dir, bitbake won't try to download it from http link and will use that already downloaded archive and check it's checksums (which is good especially as you can download a bit different archive when downloading manually). After replacing it with file:// it won't be checked. Only one part I'm not so sure about is that all my downloads in DL_DIR has also file.md5 (ie ti_cgt_c6000_6.1.9_setup_linux_x86.bin.md5), you can easily create it with md5sum, but IIRC I've seen (at least with some bitbake version) that downloaded archive was ignored when it didn't have .md5 neighbour. If that's the case, then the recipe download instrcutions should be improved. yup, that's the issue i must have been having. with the original recipe file, i was getting a fetch error, but i overlooked creating the .md5 checksum file. i can throw together a patch to clarify that in the recipe file unless you've already done it. rday -- Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
On Wed, Jun 02, 2010 at 06:48:48AM -0400, Robert P. J. Day wrote: yup, that's the issue i must have been having. with the original recipe file, i was getting a fetch error, but i overlooked creating the .md5 checksum file. i can throw together a patch to clarify that in the recipe file unless you've already done it. Please test it's really it and send patch. Thanks, -- uin:136542059jid:martin.ja...@gmail.com Jansa Martin sip:jama...@voip.wengo.fr JaMa ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
On Wed, 2 Jun 2010, Martin Jansa wrote: On Wed, Jun 02, 2010 at 06:48:48AM -0400, Robert P. J. Day wrote: yup, that's the issue i must have been having. with the original recipe file, i was getting a fetch error, but i overlooked creating the .md5 checksum file. i can throw together a patch to clarify that in the recipe file unless you've already done it. Please test it's really it and send patch. i am doing that as we speak. this is the updated ti-cgt6x.inc file that's being tested: DESCRIPTION = TI DSP Code Generation Tools HOMEPAGE = https://www-a.ti.com/downloads/sds_support/TICodegenerationTools/download.htm; SECTION = devel LICENSE = TI require ti-paths.inc require ti-staging.inc require ti-eula-unpack.inc PR = r4 S = ${WORKDIR}/cgt6x_${PV} # steps involved in setting this up # # 1. Log into the TI page, and download the appropriate binfile into #your download directory. # # 2. In that same directory, create the appropriate MD5 checksum file with #the properly-versioned form of the command: # #$ md5sum ti_cgt_c6000_6.1.9_setup_linux_x86.bin ti_cgt_c6000_6.1.9_setup_linux_x86.bin.md5 BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin SRC_URI = http://install.source.dir.local/${BINFILE};name=cgt6xbin; TI_BIN_UNPK_CMDS=Y:qY:workdir: : TI_BIN_UNPK_WDEXT=/cgt6x_${PV} do_install() { install -d ${D}${CODEGEN_INSTALL_DIR_RECIPE} cp -pPrf ${S}/* ${D}${CODEGEN_INSTALL_DIR_RECIPE} } feel free to suggest aesthetic changes. if this works, i'll submit the patch with whatever tweaks you'd like. rday -- Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 02-06-10 12:48, Robert P. J. Day wrote: On Wed, 2 Jun 2010, Martin Jansa wrote: On Wed, Jun 02, 2010 at 06:08:22AM -0400, Robert P. J. Day wrote: since this is the first time i actually built an image with any of the downloadable TI stuff, can i just sanity check what i did? it seems to have worked, but i'd rather have someone else verify i did it correctly. first step in downloading the contents into my dl dir: $ ls -l ti_* -rw-r--r-- 1 rpjday rpjday 77373969 Jun 2 02:44 ti_cgt_c6000_6.1.14_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday 77383796 Jun 1 18:45 ti_cgt_c6000_6.1.9_setup_linux_x86.bin -rw-r--r-- 1 rpjday rpjday0 Jun 2 04:40 ti_cgt_c6000_6.1.9_setup_linux_x86.bin.lock $ it wasn't immediately clear whether or not i should grab the 6.1.14 binary, but as there was a recipe for it, i took it, just in case. next, i failed to notice this in recipes/ti/ti-cgt6x.inc: SRC_URI = http://install.source.dir.local/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin; obviously, the local fetch is going to fail because of that URL, so i just changed it to: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin i would think that either a comment can be added so that users know to do that, or why not just change that line? actually, it can be tightened up a bit more, in that this: SRC_URI = file://${DL_DIR}/ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin;name=cgt6xbin BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin could be replaced with this, could it not? BINFILE=ti_cgt_c6000_${PVwithdots}_setup_linux_x86.bin SRC_URI = file://${DL_DIR}/${BINFILE};name=cgt6xbin just seems cleaner to avoid the duplication. Without testing it myself, here is what I expect: You don't have to change this if you download it to ${DL_DIR} yourself as asked in recipe. If there is archive already in download dir, bitbake won't try to download it from http link and will use that already downloaded archive and check it's checksums (which is good especially as you can download a bit different archive when downloading manually). After replacing it with file:// it won't be checked. Only one part I'm not so sure about is that all my downloads in DL_DIR has also file.md5 (ie ti_cgt_c6000_6.1.9_setup_linux_x86.bin.md5), you can easily create it with md5sum, but IIRC I've seen (at least with some bitbake version) that downloaded archive was ignored when it didn't have .md5 neighbour. If that's the case, then the recipe download instrcutions should be improved. yup, that's the issue i must have been having. with the original recipe file, i was getting a fetch error, but i overlooked creating the .md5 checksum file. i can throw together a patch to clarify that in the recipe file unless you've already done it. There's a readme in the same dir: http://gitorious.org/angstrom/openembedded/blobs/org.openembedded.dev/recipes/ti/README -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (Darwin) iD8DBQFMBjs1MkyGM64RGpERArueAJ42jrY5fdBX7cObXK97q+2j/sNFUgCeO+0D 4p+v7XNYjqRw8GlFSj8N+30= =D5vI -END PGP SIGNATURE- ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
regarding my missing creating an md5 checksum file for ti-cgt6x: On Wed, 2 Jun 2010, Koen Kooi wrote: There's a readme in the same dir: http://gitorious.org/angstrom/openembedded/blobs/org.openembedded.dev/recipes/ti/README true, but it says nothing about md5 sums. and, anyway, i'll submit a patch that just adds that note to the .inc file since that would make it standard with other .inc files, such as, say, powervr-drivers/libgles-omap3.inc, which explains in detail what to do: # some hints to get you going. as this isnt a totally open library, you need # to do some work. follow the steps below. # # 1. if you don't already have a ti account get one here. it's free: # http://www.ti.com/hdr_my_ti # 2. fetch the gl-es bin (full OMAP Graphics SDK) from here: # http://software-dl.ti.com/dsps/dsps_public_sw/sdo_sb/targetcontent/gfxsdk/latest/index_FDS.html # 3. put the OMAP35x_*.bin file in $DL_DIR, i.e. downloads/ # mv OMAP35x_Graphics_SDK_setuplinux_3_01_00_06.bin ./downloads/ # 4. make the checksum file in ./downloads/ directory: # cd ./downloads/ # md5sum OMAP35x_Graphics_SDK_setuplinux_3_01_00_06.bin OMAP35x_Graphics_SDK_setuplinux_3_01_00_06.bin.md5 # # voila! if you explain it in one place, might as well be consistent. rday -- Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
On Wed, Jun 02, 2010 at 07:01:22AM -0400, Robert P. J. Day wrote: #$ md5sum ti_cgt_c6000_6.1.9_setup_linux_x86.bin ti_cgt_c6000_6.1.9_setup_linux_x86.bin.md5 does it work when md5sum output contains also file name: $ md5sum Packages d41d8cd98f00b204e9800998ecf8427e Packages $ md5sum Packages | cut -d\ -f 1 d41d8cd98f00b204e9800998ecf8427e\n $ md5sum Packages | cut -d\ -f 1 | tr -d \n d41d8cd98f00b204e9800998ecf8427e I haven't checked what reads those .md5 files and how, but all my 9075 .md5 files in DL_DIR are only hash even without \n. Regards, -- uin:136542059jid:martin.ja...@gmail.com Jansa Martin sip:jama...@voip.wengo.fr JaMa ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel
Re: [Angstrom-devel] verifying proper bitbaking of ti-cgt6x_6.1.9 for linuxtag
On Wed, 2 Jun 2010, Martin Jansa wrote: On Wed, Jun 02, 2010 at 07:01:22AM -0400, Robert P. J. Day wrote: #$ md5sum ti_cgt_c6000_6.1.9_setup_linux_x86.bin ti_cgt_c6000_6.1.9_setup_linux_x86.bin.md5 does it work when md5sum output contains also file name: $ md5sum Packages d41d8cd98f00b204e9800998ecf8427e Packages $ md5sum Packages | cut -d\ -f 1 d41d8cd98f00b204e9800998ecf8427e\n $ md5sum Packages | cut -d\ -f 1 | tr -d \n d41d8cd98f00b204e9800998ecf8427e I haven't checked what reads those .md5 files and how, but all my 9075 .md5 files in DL_DIR are only hash even without \n. yesterday, i did the TI download to take care of OMAP35x_Graphics_SDK_setuplinux_3_01_00_06.bin, the md5 file that was generated contains exactly: 7aa37ca72f34011d353e72cc67f8aac1 OMAP35x_Graphics_SDK_setuplinux_3_01_00_06.bin and it seems to validate just fine. rday -- Robert P. J. Day Waterloo, Ontario, CANADA Linux Consulting, Training and Kernel Pedantry. Web page: http://crashcourse.ca Twitter: http://twitter.com/rpjday ___ Angstrom-distro-devel mailing list Angstrom-distro-devel@linuxtogo.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel