doc: extend README explaining how version bumping should be handled #78
Conversation
valeriosetti
changed the title
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
6b28424 to
310c271
Compare
|
@mbolivar I added you as reviewer as well for this PR because during the security meeting today (October 20th) I was told you might have some opinion on this so I wanted to have your feedback as well. |
There was a problem hiding this comment.
I think I like the idea but it's lacking some details. (And there are weird uses of ' and the possessive form.)
Also, some open questions:
- We might still have to follow the convention of calling the default branch
zephyr. I think it wouldn't hurt and would be more clear, simple. - We might want to have more details on the workflow of actually performing an update, some simple steps to make sure everyone is on the same page.
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
310c271 to
9933c46
Compare
|
@tomi-font I've updated the document based on the discussion we had offline and the work I've done in #79. Since most of the document was rewritten not all the comments that you made still apply. |
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
9933c46 to
9b627c6
Compare
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
9b627c6 to
9cb34ab
Compare
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
9cb34ab to
9ec0475
Compare
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
9ec0475 to
c7f1640
Compare
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
c7f1640 to
af49d7e
Compare
valeriosetti
changed the title
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
af49d7e to
d0e9679
Compare
This commit extends README file explaining how version bumping to upstream Mbed TLS' releases should be handled in the Zephyr's fork of Mbed TLS. Signed-off-by: Valerio Setti <vsetti@baylibre.com>
valeriosetti
force-pushed
the
upstream-sync-strategy
branch
from
d0e9679 to
df10be8
Compare
| --------------- | ||
|
|
||
| This section talks about Mbed TLS, but the same guidelines apply to the other | ||
| forks of the TF-M and Mbed TLS projects. |
There was a problem hiding this comment.
I particularly not a fan of creating a policy for another repository in here. I know most people working on TF-M are also working on MBED TLS but they are different projects. For me either it would be done in Zephyr's doc or in both projects. I think we should copy it there as well.
There was a problem hiding this comment.
I think @tomi-font would like to use this document as guide for few other modules (see also https://github.com/zephyrproject-rtos/tf-m-tests/pull/16/files). However since this guideline might not be super generic, he proposed to keep it confined to Mbed TLS and TF-M and not to go in Zephyr documentation.
There was a problem hiding this comment.
The point was exactly to have a single source of truth, and have the other relevant repos point to it. Maintaining copies of it is just extra duplication work and pretty much guaranteed to fall out of sync so I'm really not in favor of that.
And as @valeriosetti said IMO this is very specific to those repos and only affects people contributing to them so here just seems like the natural place for that.
| `[zep fromlist]` and finally `[zep noup]`. | ||
| - Cherry-pick commits without `-x` nor `-s` so as not to inflate the commit | ||
| message over time, unless you had to modify the commit and it's not just a | ||
| clean cherry-pick. |
There was a problem hiding this comment.
Wouldn't it possibly inflate only [zep noup] ? The others will come fresh from the list / tree and wouldn't make sense whoever is cherry picking them sign off ?
There was a problem hiding this comment.
There was a problem hiding this comment.
I proposed that because we have experience with this in NCS and it's just not nice and has no added value when you have a noup with 10 sign-offs and cherry picked from [...] lines. It gets messy and you don't even know anymore who contributed or modified the patch.
4 participants
Extend README file explaining how version bumping to upstream Mbed TLS' releases should be handled.