builder-class series cover letter
Get series with b4
b4 am 1899192.UD1Ciha8BC@msi-manjaroVersions newest first · linked by author and title
- v4 Latest posting 2/20 of 30 patches 0 repliesNew
- v4 posting 1/230 patches 0 repliesSuperseded
- v331 patches 0 repliesSuperseded
Earlier versions · 2
| Version | Posted | Patches | Replies | Status |
|---|---|---|---|---|
| v2 | 27 patches | 0 replies | Superseded | |
| v1 | 27 patches | 0 replies | Superseded |
Patches in v4 posting 2/2 No applies recorded yet
None of the 30 patches the cover letter announces are in this archive.
Reviews trailers per patch
No Reviewed-by, Acked-by or Tested-by trailers for v4 yet.
Cover letter [Buildroot] [PATCH v4 00/30] builder-class series cover letter
Quoted 14 lines · IST Thomas Petazzoni
Hello Atharva,
On Thu, 1 Aug 2019 08:16:13 +0530
Quoted 9 lines · Atharva Lele
Since various functions in the autobuilder script use a lot of common data, we introduce a Builder class to house these variables.
I have also included modified versions of Thomas's commits after adapting them to work with Builder class.
I created a v4 to change the order of one patch, and to delete another patch. No other changes have been made.
Quoted 6 lines
Quoted 3 lines
RFC: Few patches, namely patches 25, 27-30 have not been reviewed yet. I'd like some feedback on them, if any, so that necessary modifications can be done and it can be merged as soon as possible.
First of all, thanks a lot for all this work! Nice refactoring of the autobuild-run script.
Thank you!
Quoted 6 lines
I merged all patches up to patch 27, included. According to Arnout, applying patches 28/29/30 is problematic, because builds done in different output directories are currently not reproducible at all, so we would simply fail *all* reproducible builds if we apply those patches. So I believe we need to fix this before we can apply such patches.
Yes, correct.
Quoted 3 lines
A small detail is that patches 28 and 29 should be squashed together:
patch 28 is only introducing a new variable pointing to the secondary output directory, but it's not used anywhere.
OK, so once I'm done with the reproducibility fixes for different output directories, I'll send the revised version.
Quoted 6 lines
Also, long term, I'd like the
send_results()method to be split into several methods. One calculating the failure reason, one deciding if the build result should be rejected, one preparing the build results for upload, one doing the actual upload. But please don't work on this additional refactoring, I think it's much more important that you focus on the reproducible build topic.
Of course! I'd love to work on it once majority of work on reproducible builds is done.
Quoted 3 lines
Thanks!
Thomas
Thanks for merging!
Regards,
Atharva Lele