On Thu, Sep 29, 2022 at 12:55 PM Jose Quaresma <[email protected]> wrote:
> Hi Bruce, > > Bruce Ashfield <[email protected]> escreveu no dia quinta, > 29/09/2022 à(s) 17:42: > >> If this broke another use case, send the details and I'll debug it here. >> But these changes are required for some configurations, so I'm not going to >> simply revert. >> > > I've been trying to figure out why this is happening but unfortunately I > haven't been able to. > I have added some details in the commit message about the differences I > found. > > Anyway I have another patch that fixes the issue but as I don't understand > the reason > behind this I chose to revert. This patch is: > I can see how that would help, I had some similar changes as arrived at that original change. I was breaking one use case to fix another, and finally arrived at that solution that kept both of the common use cases working. Do you have a set of public layers that trigger the error you are seeing ? I still have setups to test the original change, so I'll know if they break. Eventually, I'll pull them into a sanity test, but I haven't had time yet, as the original change took much longer than expected. Apologies if this reply comes through as html, gmail has been making some strange choices on my behalf lately! Bruce > diff --git a/tools/spp b/tools/spp > index f9bb76b..f379205 100755 > --- a/tools/spp > +++ b/tools/spp > @@ -300,8 +300,19 @@ search_include_paths() > done > } > > +expand_if_full_path() > +{ > + for i in $@; do > + if [ "${i:0:1}" = "/" ]; then > + i=$(readlink -f $i) > + fi > + echo $i > + done > +} > + > + > # args are the input files > -infiles=$@ > +infiles=$(expand_if_full_path $@) > > processed_files="" > > >> >> Bruce >> >> >> On Thu, Sep 29, 2022 at 12:37 PM Jose Quaresma <[email protected]> >> wrote: >> >>> This reverts commit 9320c2a1aaa085e94abd15ede0d93ffdab578e9d. >>> >>> The strip_common_prefix function is changed in a way that they will >>> use the expanded path of the input argument. >>> There are some changes in the process_file function for the *.cfg files >>> only and not none for *.scc, *.patch and *defconfig. >>> >>> This doesn't work for cases where we have the defconfig files >>> provided in the some meta-layer and fail with: >>> >>> [ERROR]: processing of file /tmp/tmp.8hcr45TMiO failed >>> >>> The content of the '/tmp/tmp.8hcr45TMiO' file is currently: >>> >>> | # >>> | # spp v0.8 >>> | # processed: Thu Sep 29 04:15:36 PM UTC 2022 >>> | # >>> | # This is a preprocessor output file, do not edit >>> | # >>> | # >>> | prefix >>> | kconf non-hardware >>> /build/conf/../../layers/meta-lmp/meta-lmp-bsp/recipes-kernel/linux/linux-lmp-dev-mfgtool/imx8mq-evk/defconfig >>> | # run time: 0 seconds >>> | # processed files: >>> | # _cfg >>> /build/conf/../../layers/meta-lmp/meta-lmp-bsp/recipes-kernel/linux/linux-lmp-dev-mfgtool/imx8mq-evk/defconfig >>> >>> and with this patch applied (without 9320c2a): >>> >>> | # >>> | # spp v0.8 >>> | # processed: Thu Sep 29 04:18:21 PM UTC 2022 >>> | # >>> | # This is a preprocessor output file, do not edit >>> | # >>> | # >>> | prefix >>> /build/conf/../../layers/meta-lmp/meta-lmp-bsp/recipes-kernel/linux/linux-lmp-dev-mfgtool/imx8mq-evk/defconfig >>> | kconf non-hardware >>> /build/conf/../../layers/meta-lmp/meta-lmp-bsp/recipes-kernel/linux/linux-lmp-dev-mfgtool/imx8mq-evk/defconfig >>> | # run time: 0 seconds >>> | # processed files: >>> | # _cfg >>> /build/conf/../../layers/meta-lmp/meta-lmp-bsp/recipes-kernel/linux/linux-lmp-dev-mfgtool/imx8mq-evk/defconfig >>> >>> so the 9320c2a patch removes the prefix in this use case. >>> >>> Signed-off-by: Jose Quaresma <[email protected]> >>> --- >>> tools/scc-cmds/kconf.cmd | 4 ---- >>> tools/spp | 17 ++--------------- >>> 2 files changed, 2 insertions(+), 19 deletions(-) >>> >>> diff --git a/tools/scc-cmds/kconf.cmd b/tools/scc-cmds/kconf.cmd >>> index e0d4351..a98ae99 100644 >>> --- a/tools/scc-cmds/kconf.cmd >>> +++ b/tools/scc-cmds/kconf.cmd >>> @@ -53,10 +53,6 @@ kconf() { >>> eval echo >>> "configs/${cbranch_name}/${relative_config_dir}/${simple_config_name} \# >>> ${type}" >> "${configqueue}" >>> eval echo "\$text" $outfile_append >>> >>> - if [ -n "${verbose}" ]; then >>> - echo "config: rel: ${relative_config_dir} name: >>> ${simple_config_name} [${cbranch_name}][${prefix}]" >&2 >>> - fi >>> - >>> if [ "${type}" == "hardware" ]; then >>> echo >>> "${outdir}/configs/${cbranch_name}/${relative_config_dir}/${simple_config_name}" >>> >> ${outdir}/hardware_frags.txt >>> fi >>> diff --git a/tools/spp b/tools/spp >>> index f9bb76b..4d3fa10 100755 >>> --- a/tools/spp >>> +++ b/tools/spp >>> @@ -113,11 +113,6 @@ warn() >>> strip_common_prefix() >>> { >>> in_name=$1 >>> - if [ -n "$in_name" ]; then >>> - in_name=$(readlink -f $in_name) >>> - fi >>> - >>> - # >&2 echo "strip_common_prefix: $1 -> $in_name" >>> >>> # this takes an input name and searches all known paths. >>> # the relocation that removes the MOST from the original is >>> @@ -437,7 +432,6 @@ header() >>> one_less_dir=${d%/} >>> # strip last path component >>> one_less_dir=${one_less_dir%/*} >>> - one_less_dir=$(readlink -f $one_less_dir) >>> >>> echo "reloc_dir $one_less_dir" >>> done >>> @@ -516,17 +510,10 @@ process_file() >>> echo "[ERROR]: could not find config fragment: $in" >>> exit 1 >>> fi >>> - # old >>> - # relative_kconf=$(strip_common_prefix $kconf_name) >>> - # containing_dir=${kconf_name/$relative_kconf} >>> - # new >>> - relative_kconf=$(strip_common_prefix $kconf_name_new) >>> - containing_dir=${kconf_name_new/$relative_kconf} >>> + relative_kconf=$(strip_common_prefix $kconf_name) >>> + containing_dir=${kconf_name/$relative_kconf} >>> echo "prefix ${containing_dir}" >>> echo "kconf non-hardware ${kconf_name_new}" >>> - >>> - # >&2 echo "kconf_name: $kconf_name kconf_name_new: >>> $kconf_name_new" >>> - # >&2 echo "relative_kconf: $relative_kconf prefix: >>> $containing_dir" >>> return >>> ;; >>> *defconfig) >>> -- >>> 2.34.1 >>> >>> >>> >>> >>> >> >> -- >> - Thou shalt not follow the NULL pointer, for chaos and madness await >> thee at its end >> - "Use the force Harry" - Gandalf, Star Trek II >> >> > > -- > Best regards, > > José Quaresma > -- - Thou shalt not follow the NULL pointer, for chaos and madness await thee at its end - "Use the force Harry" - Gandalf, Star Trek II
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#11744): https://lists.yoctoproject.org/g/linux-yocto/message/11744 Mute This Topic: https://lists.yoctoproject.org/mt/93998449/21656 Group Owner: [email protected] Unsubscribe: https://lists.yoctoproject.org/g/linux-yocto/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
