search

OCR-D / ocrd_all

Master repository which includes most other OCR-D repositories as submodules
MIT License
72 stars 17 forks source link

update ocrd_kraken #288

Closed bertsky closed 2 years ago

bertsky commented 2 years ago

Even if still PR, the newer ocrd_kraken is already usable and therefore should be re-activated.

(Note: will probably work even better after applying the recent suggestions and updating here again.)

bertsky commented 2 years ago

Note: CI stops short due to a CircleCI syntax error (min instead of m for no_output_timeout) – which I have already fixed in master, but strangely it does not ever apply.

kba commented 2 years ago

Note: CI stops short due to a CircleCI syntax error (min instead of m for no_output_timeout) – which I have already fixed in master, but strangely it does not ever apply.

Did you merge the fix in master back to this branch? I did for your PR https://github.com/OCR-D/ocrd_all/pull/289 and it's running again.

bertsky commented 2 years ago

Did you merge the fix in master back to this branch? I did for your PR #289 and it's running again.

I didn't know CI config is taken from the very branch it applies to.

CI failure: see here

stweil commented 2 years ago

What about https://github.com/OCR-D/ocrd_kraken/pull/33? I think we should require kraken 3.0.9 or even 4.0.

stweil commented 2 years ago

Maybe PR #297 which also touches the Makefile rules for ocrd_kraken can be integrated here, too (or applied after this PR to avoid a merge conflict here).

bertsky commented 2 years ago

@stweil,

What about OCR-D/ocrd_kraken#33? I think we should require kraken 3.0.9 or even 4.0.

I don't understand your question. This PR sets ocrd_kraken to PR 33, which already requries kraken>=3.0 in ocrd_all.

Maybe PR #297 which also touches the Makefile rules for ocrd_kraken can be integrated here, too (or applied after this PR to avoid a merge conflict here).

No, it's the other way round. This PR is minimal and has been hanging here for one month already.

stweil commented 2 years ago

No, it's the other way round. This PR is minimal and has been hanging here for one month already.

To bring it to the point: this PR breaks some build configurations which no longer work when it is applied. I only suggested to fix that before the PR is merged instead of first breaking and then fixing.

bertsky commented 2 years ago

To bring it to the point: this PR breaks some build configurations which no longer work when it is applied. I only suggested to fix that before the PR is merged instead of first breaking and then fixing.

You mean Python 3.10? That can wait. Users should finally have ocrd_kraken back. Merging is on your side.