Closed cehbrecht closed 3 years ago
test with emu ... still seems to work.
@tlvu @Zeitsperre Maybe you like to check this PR? We can still improve on pywps-4.4 also after merge ...
@cehbrecht ~Would you like me to merge #573 into pywps-4.4
before this is merged?~
I have opened a new PR: #575. Feel free to merge at your leisure!
@Zeitsperre @fmigneault Thanks for the review. Please give the pywps-4.4 branch a try ... when there are no complains I can make the 4.4.0 release.
@cehbrecht sorry I took a day off yesterday, was not able to look at this sooner.
So basically if we diff this branch with master
, the only differences would be the jobqueue feature? Everything else are backported? We could merge this branch back into master
and it should NOT make a difference?
Here is how master
would change today if pywps-4.4
is merged into master
: https://github.com/geopython/pywps/compare/20e1e254a3f7914e555fa89f363d1f6eb5f3895c...e10c99c2c8472287f9463bab49d85f5e7597cac2
Looks like all the changes from 4.2.4 forward are missing on master
. Should be a good thing to have on master
.
Here is the reverse diff (if master
would merge into pywps-4.4
) https://github.com/geopython/pywps/compare/e10c99c2c8472287f9463bab49d85f5e7597cac2...20e1e254a3f7914e555fa89f363d1f6eb5f3895c.
Looks like mostly the jobqueue but there are other changes that I am not familiar with.
I feel it would be a great idea to have first the resolution of python 2/3 with minor patches in 4.4. Then, another (maybe major?) version for jobqueue and co. features. This will greatly align with OGC-API / owslib eventual work (https://github.com/opengeospatial/joint-ogc-osgeo-asf-sprint-2021/issues/2) that will require the job items.
I have started pointing at pywps-4.4
with this PR https://github.com/crim-ca/weaver/pull/214
And so far it doesn't look like there are any major incompatibilities.
I fear though that the alembic database migration and all other job queue elements would cause a lot of problems on my side, while being unnecessary in weaver
's case as jobs already have their own reporting/handling.
I fear though that the alembic database migration and all other job queue elements would cause a lot of problems on my side, while being unnecessary in
weaver
's case as jobs already have their own reporting/handling.
@fmigneault In the next attempt for the job queue implementation I would prefer to life without alembic database migration. The database is essentially used for the queue and this is just temporary storage.
Here is how
master
would change today ifpywps-4.4
is merged intomaster
: 20e1e25...e10c99cLooks like all the changes from 4.2.4 forward are missing on
master
. Should be a good thing to have onmaster
.
I could create a dev-branch from the current master
and re-integrate the pywps-4.4
branch. The master
should then be the same as pywps-4.4
.
I feel it would be a great idea to have first the resolution of python 2/3 with minor patches in 4.4. Then, another (maybe major?) version for jobqueue and co. features.
I am not suggesting merging master
to pywps-4.4
so no jobqueue on 4.4. I am only suggesting merge pywps-4.4
to master
so master
still contain all the fixes (return to the normal workflow before this divergence).
The 2 compare links I generated is simply for us to audit the changes between master
and pywps-4.4
branch.
Here is how
master
would change today ifpywps-4.4
is merged intomaster
: 20e1e25...e10c99c Looks like all the changes from 4.2.4 forward are missing onmaster
. Should be a good thing to have onmaster
.I could create a dev-branch from the current
master
and re-integrate thepywps-4.4
branch. Themaster
should then be the same aspywps-4.4
.
I think we simply just need to return to the regular workflow. Correct me if I am wrong but we used to merge all branches back to master
so master
always have all the cumulative changes?
So when we create new branches from master
, we ensure the new branch is not missing anything from any previous published releases?
If what I understood is correct, then after the merge of pywps-4.4
to master
, we can fix up the jobqueue on master
to your liking @cehbrecht then cut a new branch from there for new release. This assume on master
the only thing new vs pywps-4.4
is the jobqueue and nothing else (hence the usefulness of the compare links to see the diff between the branches).
I feel it would be a great idea to have first the resolution of python 2/3 with minor patches in 4.4. Then, another (maybe major?) version for jobqueue and co. features.
I am not suggesting merging
master
topywps-4.4
so no jobqueue on 4.4. I am only suggesting mergepywps-4.4
tomaster
somaster
still contain all the fixes (return to the normal workflow before this divergence).The 2 compare links I generated is simply for us to audit the changes between
master
andpywps-4.4
branch.
Right. The job-queue is not on pywps-4.4
... only in the history of master.
If what I understood is correct, then after the merge of
pywps-4.4
tomaster
, we can fix up the jobqueue onmaster
to your liking @cehbrecht then cut a new branch from there for new release.
When me make a new attempt on the job-queue I would do this on a dev branch. We can merge it to master when we are really sure we want to keep it :smile:
The next things that could happen on master (after re-integration of pywps-4.4):
When me make a new attempt on the job-queue I would do this on a dev branch.
Correct. I did not literally meant any commits direct on master
. I mean PR to master
instead of PR to pywps-4.4
.
Sorry if my message was not clear. Nothing should land on master
without a PR.
Overview
This branch backports the changes on master to pywps 4.4.
Related Issue / Discussion
568
Additional Information
List of integrated PRs:
Skipped PRs:
Contribution Agreement
(as per https://github.com/geopython/pywps/blob/master/CONTRIBUTING.rst#contributions-and-licensing)