Philippe Canal

The following is the most unexpected of all and may indicate either that valgrind is not ready for the platform or that there is a serious memory corruption somewhere. Since...

> Did ROOT get an ability to find out if a local file is on (CERN) EOS, and in which case it prepends the file path with root://eoshome-m.cern.ch/ (or similar)...

If you know it is a local file and want to stay local, you use `new TFile` instead of `TFile::Open`

The module map used by ROOT was renamed recently. Do you also see the problem with v6.32.00?

> TStreamerInfoActions:255 This is the inner most part of the I/O and the most commonly uses code. Consequently this will (should) have a very severe impact on performance. Furthermore, that...

>or you'll have a leak.(edit: depending on exactly how the looping and naming is done, but it can easily happen) In other words a portion of the new/delete problem has...

No sure what is wrong (errors are new/related) but one important missing feature is the invalidation of (some of) the cache upon transaction/file unloading.

This reminds me of https://sft.its.cern.ch/jira/browse/ROOT-8544 and https://sft.its.cern.ch/jira/browse/ROOT-8523 which is the end where fixed by a7b0b3e647409c7510b38198b08ff94fd079f857. Most likely the way the interpreter sets up the stack frames does not match the...

> even LLVM upstream is unable to handle exceptions during JIT I am not too surprised. it is the same code (+/- a few things ;)). > So maybe it's...

Changing the onfile normalized name might not be forward compatible (old version of ROOT not being able to read new files).