[meta-oe] p7zip recipe broken with rm_work

[meta-oe] p7zip recipe broken with rm_work

[Jan-Simon Möller]

Hi !

./meta-oe/recipes-extended/p7zip/p7zip_9.20.1.bb

defines:

do_install_class-native() {
    install -d ${D}${bindir}
    install -d ${D}${bindir}/Codecs
    install -m 0755 ${S}/bin/7* ${D}${bindir}
    install -m 0755 ${S}/bin/Codecs/* ${D}${bindir}/Codecs

    # Create a shell script wrapper to execute next to 7z.so
    mv ${D}${bindir}/7z ${D}${bindir}/7z.bin
    echo "#! /bin/sh" > ${D}${bindir}/7z
    echo "exec ${D}${bindir}/7z.bin \"\$@\"" >> ${D}${bindir}/7z
    chmod 0755 ${D}${bindir}/7z
}

This wrapper breaks with "rm_work". Shouldn't it use the sysroot ?

Best,
-- 
--
Jan-Simon Möller
dl9pf@gmx.de

Re: [meta-oe] p7zip recipe broken with rm_work

[Burton, Ross]

On 3 November 2016 at 11:03, Jan-Simon Möller <dl9pf@gmx.de> wrote:

>     # Create a shell script wrapper to execute next to 7z.so
>     mv ${D}${bindir}/7z ${D}${bindir}/7z.bin
>     echo "#! /bin/sh" > ${D}${bindir}/7z
>     echo "exec ${D}${bindir}/7z.bin \"\$@\"" >> ${D}${bindir}/7z
>     chmod 0755 ${D}${bindir}/7z
> }
>
> This wrapper breaks with "rm_work". Shouldn't it use the sysroot ?
>

Yeah the script should just use ${bindir}, but why is there even a wrapper
script?

Ross

Re: [meta-oe] p7zip recipe broken with rm_work

[Khem Raj]

[-- Attachment #1: Type: text/plain, Size: 971 bytes --]


> On Nov 3, 2016, at 7:41 AM, Burton, Ross <ross.burton@intel.com> wrote:
> 
> On 3 November 2016 at 11:03, Jan-Simon Möller <dl9pf@gmx.de> wrote:
> 
>>    # Create a shell script wrapper to execute next to 7z.so
>>    mv ${D}${bindir}/7z ${D}${bindir}/7z.bin
>>    echo "#! /bin/sh" > ${D}${bindir}/7z
>>    echo "exec ${D}${bindir}/7z.bin \"\$@\"" >> ${D}${bindir}/7z
>>    chmod 0755 ${D}${bindir}/7z
>> }
>> 
>> This wrapper breaks with "rm_work". Shouldn't it use the sysroot ?
>> 
> 
> Yeah the script should just use ${bindir}, but why is there even a wrapper
> script?

While agreeing on converting the script into a patch, I fail to understand
how it it different then some of other install calls in the same function

> 
> Ross
> --
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-devel


[-- Attachment #2: Message signed with OpenPGP using GPGMail --]
[-- Type: application/pgp-signature, Size: 211 bytes --]

Re: [meta-oe] p7zip recipe broken with rm_work

[Burton, Ross]

On 3 November 2016 at 19:17, Khem Raj <raj.khem@gmail.com> wrote:

> While agreeing on converting the script into a patch, I fail to understand
> how it it different then some of other install calls in the same function
>

Because the wrapper script will contain ${D} which is inside the work
directory.

I just noticed this has already been fixed in master with this commit:

$ git show 67d6ed40bfeb219490e509f33c27ebd2a1149d30
commit 67d6ed40bfeb219490e509f33c27ebd2a1149d30
Author: Jonathan Liu <net147@gmail.com>
Date:   Fri Sep 23 12:39:24 2016 +1000

    p7zip: do not hardcode path in native wrapper script

    Signed-off-by: Jonathan Liu <net147@gmail.com>
    Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>

So if you're using a branch, submit a backport request.

Ross