-
Notifications
You must be signed in to change notification settings - Fork 18
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Move compression and decompression routines into dxtbx #228
Conversation
Selecting draft appears to reset target ... |
Question to reviewers: would it aid readability to express e.g.
in hex? |
Attempts to fix SConscript in keeping with Will revisit Monday |
Wait, what? This fixed it 🤔 |
OK, something weird here. The tests were failing for me, passing in here. I thought was my build being skewed so I have just made a clean bootstrap and ... still fails. Lots of
why is this not failing centrally? |
covered by tests skipped here because dials_regression? |
I have a clean build with no dials regression, so that is quite odd. Suspect it would fail in a clean build, but maybe the route for bootstrapping to where it got to with incremental builds is such that it has something built & linked. Guess no-op changes to force rebuild could flush out some errors? |
You can force flushed cache builds by resetting the date in line 2 of |
This branch was branched before the Azure builds were set up, so you would only get PR builds, and not push builds. Updating branch to enable push builds.
Codecov Report
@@ Coverage Diff @@
## master #228 +/- ##
==========================================
- Coverage 45.36% 45.36% -0.01%
==========================================
Files 229 229
Lines 19200 19199 -1
Branches 2720 2720
==========================================
- Hits 8711 8710 -1
Misses 9975 9975
Partials 514 514 |
to pick up the new pytest dispatcher
)" Caused downstream build errors in dials. This reverts commit 8824453.
Move CBF compression and decompression routines back into dxtbx once more rehash of #228
The byte-offset compression introduced by #238, #228 was incorrect, causing problems for anything _writing_ cbf files... so dlsnsx2cbf (and probably dials.merge_cbf). - 0xf777 -> 0x7fff as one greater than -0x8000 - values with too large a negative value were being incorrectly encoded - treatment of the fall through / last case in compression for similarity with decompression layout - more thorough testing in particular tests with negative delta of magnitude greater than 0x8000 - test added which verifies that the HDF5 and CBF representation of the data from dlsnxs2cbf are the same
3rd time lucky...