search

OCR-D / ocrd_tesserocr

Run tesseract with the tesserocr bindings with @OCR-D's interfaces
MIT License
39 stars 11 forks source link

Missing v0.5.0 tag #85

Closed mikegerber closed 5 years ago

mikegerber commented 5 years ago

PyPI has 0.5.0, while this repo has 0.4.0 as the latest version. (Found this while investigating a problem with 0.5.0)

Related: The releases page in this repo says that 0.2.2 is the newest version while also offering to show newer tags – with 0.3.0 and 0.4.0 🙈 I have no idea what GitHub is doing here.

bertsky commented 5 years ago

I can explain:

bertsky commented 5 years ago 
* [533d83c](https://github.com/OCR-D/ocrd_tesserocr/commit/533d83c049ba93c86e0d77e40ada7ef8e122dc87) should have become PyPI release 0.4.1, but was forgotten

I am not sure whether we should still release the current master as 0.4.1. It makes little sense now that we already have 0.5.0. But maybe @kba has another opinion?

* Github can only make releases for things that have a version tag. This has been neglected for a while. I will fix this now...

Actually the tags are there, but not automatically turned into GH releases. I don't what reasons @kba has for publishing some but not all tags as GH releases.

Meanwhile, I added a fix to current master which ensures that ocrd does not come from core:edge, so at least we have both branches working in CI.

mikegerber commented 5 years ago

PyPI release 0.5.0 is a pre-release

It's not a pre-release on PyPI :) (For some reason, in one of my installations, it does not pull in ocrd 2.0.0a1, and so I get non-functional ocrd-tesserocr-* tools – this was the cause for me to look for 0.5.0 here ← Seems to be fixed now)

bertsky commented 5 years ago

0.5.0 is by definition a pre-release, as is everything below 1.0. So the problem is that we cannot prevent installers from using the newest version. We should have made a "stable" 1.0 out of the current master, and release edge as 1.1a1 or so.

I guess we have to wait until @kba gets back for a strategy how to proceed with pilots.

kba commented 5 years ago

I am not sure whether we should still release the current master as 0.4.1. It makes little sense now that we already have 0.5.0. But maybe @kba has another opinion?

No, that's why I went ahead with 0.5.0. As @bertsky said, everythin below 1.0 is pre-release and by definition unstable. We want the early adopters to use the newest version of tesserocr (and cis_ocrd BTW) which use the newest pre-release of core.

I would strongly prefer to get rid of the edge branches and ..a1 releases ASAP.

For now, I'll update the tags (probably just forgot to push v.0.5.0) and mark it as a pre-release and we'll hopefully have a clear path forward with the outstanding PR in core by the end of the week.

kba commented 5 years ago

Fixed now, releases released here as well.