| | 31 | * Greg busy working on Wynton problems |
| | 32 | - Making new accounts broken, 50 users waiting to get new accounts. |
| | 33 | - Working on restoring 15 Tbytes, and millions of files where file names are known but directory structure was lost. |
| | 34 | |
| | 35 | * mmCIF out of memory crash |
| | 36 | - Eric says 1 or 2 mmcif crash reports coming in every day. |
| | 37 | - Greg has valgrind giving the out of memory crash. |
| | 38 | - Tried not using memory mapping and not doing template fetches and valgrind still gets out of memory. |
| | 39 | - valgrind does not get out of memory opening pdb file. |
| | 40 | - next step is put in print statements and see how far it gets under valgrind. |
| | 41 | |
| | 42 | * Trac login from Windows Chrome broken |
| | 43 | - Greg says ucsf web server scan says we are not offering TLS or SSL, believed to be false result. |
| | 44 | - Possibly Chrome login failure is related. |
| | 45 | - Scooter will need to look at this problem after he is back tomorrow. |
| | 46 | |
| | 47 | * Mac Intel daily builds failed to start |
| | 48 | - PyOpenGL could not find OpenGL.framework, only on Intel, not ARM. |
| | 49 | - Daily build does start on macOS 10.14 build machine essex2. |
| | 50 | - Also starts using TomG's development build on new macOS 13.3 |
| | 51 | - Seems to be related to building on old macOS 10.14. |
| | 52 | - Using development build ChimeraX launcher executable fixes problem. |
| | 53 | - Zach patched PyOpenGL to work around problem. |
| | 54 | |
| | 55 | * Update Intel Mac build machine to macOS 10.15? |
| | 56 | - Chimera is also built on this machine, essex2. |
| | 57 | - Might run into difficulties building chimera on newer XCode. |
| | 58 | - Would want to try chimera build on 10.15 but don't have a test machine. |
| | 59 | - Decided to just keep the builds on 10.14 since it is working. |
| | 60 | - Zach said he made a bundle builder change when updating to python 3.11 because 10.14 machine did not know how to deal with Python universal build. |
| | 61 | |
| | 62 | * Chimera 1.17.2 release |
| | 63 | - Chimera 1.17.1 was missing ffmpeg so no movies can be recorded. |
| | 64 | - Only one bug report so far. Why so few? |
| | 65 | - Error message does not give "report a bug" button, so probably users don't know how to report it. |
| | 66 | - Decided to make 1.17.2 release on all platforms. |
| | 67 | - Ubuntu gets ffmpeg from RPM dependency so probably already works. |
| | 68 | - Will just promote current daily build to 1.17.2. |
| | 69 | - Chimera will notify users of new release. |
| | 70 | - Don't plan on announcing new release. |
| | 71 | - Eric will update release notes. |
| | 72 | - Release planned or this Thursday. |
| | 73 | |
| | 74 | * PyQt 6.5.1 update |
| | 75 | - yes we will update to PyQt 6.5.1 |
| | 76 | - Also time to update all the other prereqs. |
| | 77 | - Greg suggests adding required hash values to requirements.txt file because some PyPi hacking has been in the news. |
| | 78 | |
| | 79 | * Python 3.11.4 |
| | 80 | - ChimeraX is at Python 3.11.2 |
| | 81 | - A change in 3.11.4 breaks PyQt handling of enums. |
| | 82 | - Wait for python 3.11.5 to see how it gets resolved, or whether we will need to fix our PyQt enum usage. |
| | 83 | |
| | 84 | * Python 3.11 speed up |
| | 85 | - addh took 23.0 seconds on ChimeraX 1.6 / Python 3.9 and 20.7 seconds on daily build with Python 3.11, 10% speed-up. Tom only tried it once, so needs more testing to verify. |
| | 86 | - mmcif read speeds vary fro 9-11 seconds in weekly benchmark so not clear if read speed changed. Not expected to change since mostly C++. |
| | 87 | |
| | 88 | * Toolshed prevent bundle downgrades |
| | 89 | - Eric asks how to avoid users downgrading to the old addh bundle on Toolshed. |
| | 90 | - Greg says put in that ChimeraX core version must be older than date that had that addh version. |