Confusing branches
Closed this issue · 8 comments
There are some stale branches like update-readme
, trakh_dominik
, or yforster-travis_for_wip
that somewhat disturb the overall branch structure, i.e. coq-8.12
, coq-8.11
, etc.
I suggest those branches to be either moved to private forks (work via pull requests) or deleted to keep the branch structure clean.
I deleted count_states_haskell
, update-readme
and yforster-travis_for_wip
.
@dominik-kirst @DmxLarchey what's the status of trakh_full
,trakthenbrot_ijcar
and trakh_dominik
? Given the webpage of the project and that the results are merged to the library, are the other branches still needed?
master
can be deleted, do you agree @DmxLarchey @mrhaandi?
What's the status of informative_lift
@DmxLarchey? The corresponding PR #27 is closed. Maybe we can move the branch to your private fork as Andrej suggests? Similarly, my_own_tapes
was just for discussion, maybe we can move it to your fork as well?
Afterwards we'd be down to coq-8.12
, coq-8.11
, coq-8.10
and noopam
(the latter becoming potentially obsolete in the near future once coq/platform takes off).
master
can be deleted
Either that or renamed to coq-8.9
.
@yforster @mrhaandi I did copy the references to my_own_tapes
and informative_lift
to my own fork and then removed them from the main uds-psl
git repo.
The trakhtenbrot_ijcar
branch is hyper-linked twice in the IJCAR paper (beginning of the Discussion section 7) so removing it would somehow be problematic. I realize now that it was not a good idea to refer to a branch
of a living repo. Maybe it would have been better to refer to a tag
instead. I do not know if it is possible to move the branch
name to a tag
name without breaking the below http links:
- https://github.com/uds-psl/coq-library-undecidability/tree/trakhtenbrot_ijcar/theories/TRAKHTENBROT
- https://github.com/uds-psl/coq-library-undecidability/tree/trakhtenbrot_ijcar/theories/Shared/Libs/DLW
If possible I will do it and then the tag
trakhtenbrot_ijcar
will be less visible than the branch
trakhtenbrot_ijcar
.
Otherwise, if would prefer if we keep that branch for some time, because of the IJCAR paper.
Concerning, trakh_full
and trakh_dominik
, I did copy them to my own fork as well. However, since I did work on them with @dominik-kirst, I will wait for his green light to remove them from uds-psl
.
@dominik-kirst, you can either copy the refs to your own fork as well or else we can share those on one of our forks. However, they are mostly stalled now if think.
Green light for removing trakh_full
and trakh_dominik
.
Maybe it would have been better to refer to a
tag
instead.
As a side note, I discussed with Yannick that tag
-based references are a future-proof mechanism in future papers that do not pollute the overall repository structure. This also at some point should be discussed and done uniformly, e.g. in a year-conference-topic
style.
Ok the tag
works, ie the former outside references still work with the tag
instead of the branch
. Hence, I will remove the branch
reference for the official repo.
Btw, tags are automatically converted to releases which is maybe not optimal...
master
can be deleted, do you agree @DmxLarchey @mrhaandi?Either that or renamed to
coq-8.9
.
Now the only branch that might be misunderstood is master
, which oftentimes signifies the main branch.
Is there a reason not to rename it to coq-8.9
to retain the project's origin?
Afterwards, this issue can be closed.
I renamed master
to coq-8.9
. Thanks Andrej for poking us, the repo looks a lot tidier now!