Amar Takhar commented on a discussion: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/merge_requests/582#note_126513


It was discussed during several meetings and I was told to submit an MR with 
the HAL files.  The issue had been around for a while as well.

The HAL files are currently a mess: Files are renamed, changes are made that 
aren't even noted in commit messages.  Some pull from multiple repositories.  
No instructions for updating _and_ no reasoning why the version was picked.

This is a _vast_ improvement over the current situation.  Looking at the other 
3rd party imports:

* Most don't record what version or hash it came from
* No record of where the sources were pulled.
* Some are 10+ years out of date which is fine but we don't have any record as 
to why we have stayed on older versions
* No instructions on how to update
* No rules for import: every single import was done differently.  Yes, every 
single one.
* None of them with the exception of zlib is a pure import they have changes 
that were brought in from the original source we have no way of even detecting 
those changes unless you figure out what the original source was.. which of 
course, we don't record anywhere not in docs or in the repo as a separate 
commit.

The maintenance burden with what we have now is incredibly high as they're all 
done differently.  Space is extremely cheap.  With this the update procedure is:

* \`r`` m -rf *` ``
* `cp -R <path to repo>/* .`
* Commit changes
* Tag pure import
* Re-apply changes on top.

FreeBSD has been using this method for 40+ years and it's incredible to look 
back on the history of changes it's trivial to see:

* What has changed between upstream versions -- pure import tags
* What changes FreeBSD is maintaining to see maintenance overhead
* Upstream as many changes as possible to reduce that overhead or make changes 
to avoid modifying the code

Every single one of our imports was already impossible to review as the RTEMS 
changes came in with the import so I'm confused as to how you see this MR which 
has very clear distinctions of the changes is _difficult_ to review when the 
original had all the changes hidden.

-- 
View it on GitLab: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/merge_requests/582#note_126513
You're receiving this email because of your account on gitlab.rtems.org.


_______________________________________________
bugs mailing list
bugs@rtems.org
http://lists.rtems.org/mailman/listinfo/bugs

Reply via email to