xmo-odoo

xdo note: would strongly benefit from the direct staging com' between mergebot and runbot: if the mergebot has to push branche, for concurrent stagings it needs multiple branches, but then...

@wsw70 I expect it was removed following > The API reference on the website is not to be used. This is just some leftovers of a previous attempt.

Done, and rebased while at it.

> Could it be that fast has much worse compression than zlib-ng for instance? That might make sense. Although FWIW testing combinations of wbits and compression level using Python's stdlib...

> Could it be that `fast` has much worse compression than `zlib-ng` for instance? FWIW assuming the various zlib features of gix-features impact this, I'm not seeing anything significant when...

Thinking a bit I figured the pack itself might cause issues or interfere so I stripped the script to just generating an `Entry` from the tree and log the size:...

> Maybe with this script you can `git bisect` over the codebase, which should yield the commit that breaks it with certainty. Updating the script to not depend on third...

I guess in that case the solution would be to compress the data on the application side and create the `Entry` by hand instead of using the `from_data` helper? (it's...

> That's great to hear! Thanks also for the reproduction script! > > I had a feeling that this might be the default compression settings which are different in `miniz_oxide`....

Wrong button, submitted too early.