Bug 13444

Upstream bzip2 has been pretty responsive lately. It's probably worth discussing this with them and seeing if they would be interested in using a more standardized test output format.

Hi,

Here product version is 0.0.0. what does it means?

Rahul, The 0.0.0 is either a poor choice by the reporter or an interaction bug between web browsers and the Yocto Bugzilla version. I've set it to 3.1 now.
Do you want to work on this enhancement? If so feel free to ask questions here, or on IRC.

Hi Randy MacLeod,

I want to work on this enhancement.

I cloned the oe-core and bitbake source code via below link
git clone git://git.openembedded.org/openembedded-core
git clone git://git.openembedded.org/bitbake bitbake

I go through the below link 
https://wiki.yoctoproject.org/wiki/Releases

as product version is 3.1, openembedded-core branch should be Dunfell and bitbake branch should be 1.46.
but But there is no Dunfell branch in oe-core and 1.46 branch in bitbake source code.

can you please guide me on which branch i should work.

Thanks & Regards
Rahul Kumar

Rahul,

The master branch should be used, for that matter you should use the master branch of oe-core also.

Thanks Joshua Watt,

But when i am switching to master branch and enable ptest,below error I am facing 

ERROR: busybox-1.31.1-r0 do_package: Error executing a python function in exec_python_func() autogenerated:

The stack trace of python calls that resulted in this exception/failure was:
File: 'exec_python_func() autogenerated', lineno: 2, function: <module>
     0001:
 *** 0002:ptest_update_alternatives(d)
     0003:
File: '/opt/opensource/openembedded-core.git/meta/classes/ptest.bbclass', lineno: 98, function: ptest_update_alternatives
     0094:        for alt_name, alt_link, alt_target, _ in alternatives:
     0095:            # Some alternatives are for man pages,
     0096:            # check if the alternative is in PATH
     0097:            if os.path.dirname(alt_link) in bin_paths:
 *** 0098:                os.symlink(alt_target, os.path.join(ptest_bindir, alt_name))
     0099:}
     0100:
     0101:do_configure_ptest_base[dirs] = "${B}"
     0102:do_compile_ptest_base[dirs] = "${B}"
Exception: FileExistsError: [Errno 17] File exists: '/bin/busybox.suid' -> '/opt/opensource/openembedded-core.git/build-master/tmp-glibc/work/core2-64-oe-linux/busybox/1.31.1-r0/package/usr/lib/busybox/ptest/bin/login'

ERROR: Logfile of failure stored in: /opt/opensource/openembedded-core.git/build-master/tmp-glibc/work/core2-64-oe-linux/busybox/1.31.1-r0/temp/log.do_package.8901
ERROR: Task (/opt/opensource/openembedded-core.git/meta/recipes-core/busybox/busybox_1.31.1.bb:do_package) failed with exit code '1'

Does anyone have idea on this?

Created attachment 4652 [details]
run-test.sh script result

Hi,

I modified run-tests.sh script to write PASS/FAIL as expected by the ptest infrastructure.
run-test.sh script output file attached here.

basically run-test.sh file is doing two kind of tests
1) First test the good bz2 files (Testing decompression and recompression)
   Output looks like below
	Testing decompression and recompression...

	Processing ./go/regexp/re2-exhaustive.txt.bz2
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 Decompress
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 md5sum Matched
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 Recompress 
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 Redecompress
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 md5sum Matched
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 Decompress (small)
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 Md5sum Matched
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 Recompress (small)
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 Redecompress (small)
	PASS: ./go/regexp/re2-exhaustive.txt.bz2 md5sum Matched

2) Then test all bad inputs
   Output looks like below
	Testing detection of bad input data...

	bzip2: Data integrity error when decompressing.
	PASS: Correctly found data integrity errors in ./go/compress/fail-issue5747.bz2.bad during decompress.

	bzip2: Data integrity error when decompressing.
	PASS: Correctly found data integrity errors in ./go/compress/fail-issue5747.bz2.bad during decompress (small).bzip2

can any one review this attached output file and guide me if any further modification required in output.

If no one replies, I either Trevor (who I added) or I can look at this tomorrow.
Since you are working on the bug, I've assigned it to you. Thanks!

(In reply to comment #8)
> If no one replies, I either Trevor (who I added) or I can look at this
> tomorrow.
> Since you are working on the bug, I've assigned it to you. Thanks!

Hi,

I sent the patch for this enhancement.

https://patchwork.openembedded.org/patch/171539/

Kindly take a look.

Hi,

As per your suggestion I submitted bzip2 patch to bzip2 mailing list
https://sourceware.org/pipermail/bzip2-devel/2020q2/000102.html
But i did not getting any response.


I have sent Patch Version 2.
https://patchwork.openembedded.org/patch/172134/

Kindly take a look .

Hi Randy,

As per your comments i made changes and submitted Patch Version 2 to openembedded-core mailing list.
https://patchwork.openembedded.org/patch/172134/

Kindly take a look and comment if any thing remaining for this bug

Thanks & Regards
Rahul Kumar

Hi Randy,

In mailing list could not conclude whether which option we should choose.
kindly help me on it.

1) LICENSE = \
   "bzip2-1.0.6 & GPLv3+ & Apache-2.0 & MS-PL & BSD-3-Clause & Zlib"

and

2) LICENSE = "GPLv3+"


One more thing,bzip2 community has been accepted my patch and add some additional changes on top of it .
so now our changes is part of bzip2-tests.git source code.
You can check here
https://sourceware.org/git/?p=bzip2-tests.git;a=commit;h=f9061c030a25de5b6829e1abf373057309c734c0

So I need your suggestion with respect to Yocto.

Since required changes for ptest infrastructure is integrated to the bzip2-tests.git source code.
So now we can just change the SRCREV with the latest commit instead of applying the patch (0001-bzip2-modify-run-tests-script.patch) to the bzip2-tests.git source code.

If we are going with second option.
Then i think no need to add any License for it because we are just integrating the bzip2-tests source to the bzip2 recipe and not apply any patch to the bzip2-tests.git source code. 

Kindly correct me if i am wrong at any point.

Thanks and Regards
Rahul Kumar

I have written to the bzip-devel list to discuss the license terms.
   https://sourceware.org/pipermail/bzip2-devel/2020q2/000109.html

Meanwhile, we can merge the recipe with:

   LICENSE = \
   "bzip2-1.0.6 & GPLv3+ & Apache-2.0 & MS-PL & BSD-3-Clause & Zlib"
since that reflects the source terms as they are written now.

And yes, just change the SRCREV with the latest commit instead of applying the patch (0001-bzip2-modify-run-tests-script.patch) to the bzip2-tests.git source code.

(In reply to comment #13)
> I have written to the bzip-devel list to discuss the license terms.
>    https://sourceware.org/pipermail/bzip2-devel/2020q2/000109.html
> 
> Meanwhile, we can merge the recipe with:
> 
>    LICENSE = \
>    "bzip2-1.0.6 & GPLv3+ & Apache-2.0 & MS-PL & BSD-3-Clause & Zlib"
> since that reflects the source terms as they are written now.
> 
> And yes, just change the SRCREV with the latest commit instead of applying
> the patch (0001-bzip2-modify-run-tests-script.patch) to the bzip2-tests.git
> source code.

Hi Randy,

Thanks for your confirmation.
I have sent Patch v3 to openembedded-core mailing list.
https://patchwork.openembedded.org/patch/173009/

Thanks & Regards
Rahul kumar

Hi

I have sent Patch v4 to openembedded-core mailing list.
https://patchwork.openembedded.org/patch/173912/

Thanks & Regards
Rahul kumar

Rahul,

and it was merged:

https://git.openembedded.org/openembedded-core/commit/?id=968529b484265b577e8b8130caef448a8a9d8b68

commit 968529b484265b577e8b8130caef448a8a9d8b68
Author: Rahul Kumar <rahulk@mvista.com>
Date:   Thu Jun 25 09:11:27 2020

    bzip2: Add test suite for bzip2

so I'm closing the defect.

Care to work on another, it's almost certainly going to be less hassle than this one was for you. Thanks for being so patient and working though all the issues.