mbox series

[0/2] Resolve conflict between /usr/bin/hello from lmbench and mbedtls

Message ID 20230225231717.3887902-1-denis@denix.org
Headers show
Series Resolve conflict between /usr/bin/hello from lmbench and mbedtls | expand

Message

Denys Dmytriyenko Feb. 25, 2023, 11:17 p.m. UTC 
Both lmbench and mbedtls packages provide generically-named /usr/bin/hello
binary, which causes a conflict when trying to install them both into a
rootfs or SDK. Set up alternative providers to resolve the conflict.

Denys Dmytriyenko (2):
  lmbench: set up /usr/bin/hello as alternative
  mbedtls: set up /usr/bin/hello as alternative

 .../recipes-connectivity/mbedtls/mbedtls_2.28.2.bb           | 5 ++++-
 meta-oe/recipes-benchmark/lmbench/lmbench_3.0-a9.bb          | 4 ++--
 2 files changed, 6 insertions(+), 3 deletions(-)

Comments

Ross Burton Feb. 27, 2023, 11:05 a.m. UTC | #1 
On 25 Feb 2023, at 23:17, Denys Dmytriyenko via lists.openembedded.org <denis=denix.org@lists.openembedded.org> wrote:
> 
> Both lmbench and mbedtls packages provide generically-named /usr/bin/hello
> binary, which causes a conflict when trying to install them both into a
> rootfs or SDK. Set up alternative providers to resolve the conflict.

Are either of these binaries actually useful?  The, well, alternative is to just delete them stupidly named binaries (or rename).

Ross
Denys Dmytriyenko Feb. 28, 2023, 11:32 p.m. UTC | #2 
On Mon, Feb 27, 2023 at 11:05:36AM +0000, Ross Burton wrote:
> On 25 Feb 2023, at 23:17, Denys Dmytriyenko via lists.openembedded.org <denis=denix.org@lists.openembedded.org> wrote:
> > 
> > Both lmbench and mbedtls packages provide generically-named /usr/bin/hello
> > binary, which causes a conflict when trying to install them both into a
> > rootfs or SDK. Set up alternative providers to resolve the conflict.
> 
> Are either of these binaries actually useful?  The, well, alternative is to 
> just delete them stupidly named binaries (or rename).

My assumption is that at least lmbench is using this binary for some testing 
or benchmarking. But I haven't looked very deep into it. Plus I wanted to get 
this backported to Kirkstone, so the least invasive approach made more sense.