bazarr: update to 1.4.0

[smaarn]

Description

• Upgrades bazarr version to 1.4.0
• Cleanup no longer relevant ARMv7L_ARCHS logics for numpy

Checklist

• Build rule all-supported completed successfully
• New installation of package completed successfully
• Package upgrade completed successfully (Manually install the package again)
• Package functionality was tested

Type of change

• Package update

[ngthwi]

Working fine on a DS215j and a DS218j
Thanks a lot

[hgy59]

Working fine on a DS215j and a DS218j Thanks a lot

Those models are not ARMv7L

Only models with hi3535 CPU are ARMv7L, i.e.

• Network Video Recorders (NVR216, NVR1218)
• Video Station Models

Details see https://github.com/SynoCommunity/spksrc/wiki/Synology-and-SynoCommunity-Package-Architectures

EDIT:
update models (Routers are not ARMv7L but Video Station Models are), ...

[smaarn]

Flagging as draft as, even if it does compile and build successfully we don't really have any client for testing it, as of now.

[ngthwi]

This is the bazarr status page on a DS218j...
There's a reference to ARMv7l on the os name...
Would it mean that Synology is using the same os on ARMv7l and ARMv7..?

[smaarn]

@ngthwi I think this ought to be ready for merge & release. Do you think you could have time to test the latest 1.4.0 binaries ? They are available here: https://github.com/SynoCommunity/spksrc/pull/5911/checks

@th0ma7 any objection to merging ? (things I wouldn't have spotted for example...)

All apologies for the delays

[hgy59]

Just to be clear: $(ARMv7L_ARCHS) is not related to armv7l as reported by uname -m.

we have the definitions in spksrc.archs.mk

ARMv7_ARCHS = $(GENERIC_ARMv7_ARCH) $(DSM_ARMv7_ARCHS) $(SRM_ARMv7_ARCHS)
# hi3535 is not supported by generic ARMv7 arch
ARMv7L_ARCHS = hi3535

ARMv7L_ARCHS is used to separate hi3535 cpu from other armv7 cpus.

None of the Diskstation Models have ARMv7L_ARCHS. Only the (Network Video Recorder) Models NVR1218 and NVR216 and the (VideoStation) Model VS360HD have.

So the title of this PR is missleading since armv7 support for DSM was never dropped.

Was there any request to support this package on NVR1218, NVR216 or VS360HD?.
If not, I recommend to keep

UNSUPPORTED_ARCHS = $(ARMv5_ARCHS) $(ARMv7L_ARCHS)

and to remove the <br />2. Add back support for ARMv7L architectures' from the changelog.

If you want to keep the ARMv7L_ARCHS then please update the changelog to
<br />2. Add back support for NVR1218, NVR216 and VS360HD Models'

[smaarn]

Just to be clear: $(ARMv7L_ARCHS) is not related to armv7l as reported by uname -m.

we have the definitions in spksrc.archs.mk

ARMv7_ARCHS = $(GENERIC_ARMv7_ARCH) $(DSM_ARMv7_ARCHS) $(SRM_ARMv7_ARCHS)
# hi3535 is not supported by generic ARMv7 arch
ARMv7L_ARCHS = hi3535

ARMv7L_ARCHS is used to separate hi3535 cpu from other armv7 cpus.

None of the Diskstation Models have ARMv7L_ARCHS. Only the (Network Video Recorder) Models NVR1218 and NVR216 and the (VideoStation) Model VS360HD have.

So the title of this PR is missleading since armv7 support for DSM was never dropped.

Was there any request to support this package on NVR1218, NVR216 or VS360HD?. If not, I recommend to keep

UNSUPPORTED_ARCHS = $(ARMv5_ARCHS) $(ARMv7L_ARCHS)

and to remove the <br />2. Add back support for ARMv7L architectures' from the changelog.

If you want to keep the ARMv7L_ARCHS then please update the changelog to <br />2. Add back support for NVR1218, NVR216 and VS360HD Models'

I think I got confused here with the custom logic for numpy which was related to those ARMv7L_ARCHS architectures.

Will be reverting the change and removing the therefore useless tricks for numpy.

[smaarn]

It seems that ARMv7L_ARCHS are actually part of the unsupported architectures and this line is therefore useless

[hgy59]

LGTM

[ngthwi]

@ngthwi I think this ought to be ready for merge & release. Do you think you could have time to test the latest 1.4.0 binaries ? They are available here: https://github.com/SynoCommunity/spksrc/pull/5911/checks

@th0ma7 any objection to merging ? (things I wouldn't have spotted for example...)

All apologies for the delays

Already tested that...
Working fine

[mreid-tt]

Hello @smaarn, I've successfully published the builds for the merged PR on the website. The qoriq build remains inactive, due to a compatibility issue with OpenSSL 3, as reported in #5905 (comment).

By the way, I've observed that the recent commits you've merged were published by other contributors. While we're more than willing to provide assistance, if you encounter any challenges with your publishing process, feel free to reach out to me on Discord for support.