...
That whole mess with hadd should be fixed.
PyRoot improvements
Supposedly as of Root 5.12 there is support for VisualStudio 8
Remote file access improvements - which I think Wilko mentioned some months ago.
Provides interface to fftw as of v5.11
For hairy details, one can always view the release notes.
...
? - Not tested
- Doesn't work
GLAST Component | Windows | Linux | Comments | |
ROOT Classes | ||||
calibRootData (v1r7p7) |
|
| Compiled just fine on Linux and Windows | |
commonRootData (v2r13) |
|
| Compiled just fine on Linux and windows. Loaded and used the shared library in CINT. | |
digiRootData (v9r16) |
|
| Compiled and ran test routine on Linux and Windows. Loaded and used the shared library in CINT. | |
mcRootData (v2r20) |
|
| Compiled and ran test routine on Linux and Windows. Loaded and used the shared library in CINT. | |
reconRootData (v9r11) |
|
| Compiled and ran test routine on Linux and Windows. Loaded and used the shared library in CINT. | |
gcrSelectRootData (v2r1) |
|
| Compiled and ran test routine on Linux and Windows. | |
General I/O | ||||
Can read in ntuples from DC2 |
|
| Tested interactively. Worked without problems on Linux | |
Can read in ntuples from DC1 |
|
| Tested interactively. Worked without problems on Linux. | |
Can read in full ROOT tree files generated with 5.10.00 using 5.14.00? |
|
| Tested interactively. Worked on Linux (did receive too few bytes warning for AcdRecon) | |
How is hadd? |
|
|
| |
ROOT I/O Packages | ||||
RootConvert (v1r23p1) |
|
| Compiled and ran test routine on Linux | |
RootIo (v17r28) |
| | Compiled and ran test routine for writing on Linux and Windows which performs writing. Running test with reading produced some warning about reading too few bytes for vector<Double32_t> members in AcdRecon which are empty - part of the read job reads in rootTestData files created using v5.10.00. Fixed** | |
ntupleWriterSvc (v3r15) |
|
| Compiled and ran test routine on Linux and Windows just fine. | |
CHECKOUT Packages | ||||
GlastRelease (v9r20) |
|
| When compiling Gleam on Linux - ran into a small error with GlastClassify's RootTuple.cxx - needed to include iostream for cerr. Ran vanilla test_Gleam.exe on both Linux and Window to create a merit ntuple first - that worked fine. The resulting ntuple was readable in CINT. | |
Analysis | ||||
PyRoot - does it work? |
|
| Big hairy deal to get it to work on Windows - it required modifying the ROOT.py file in ROOTSYS/bin. See: http://root.cern.ch/phpBB2/viewtopic.php?p=17618&sid=294348839e56fdaa7580b6e65253dc61 | |
RootAnalysis () |
|
|
| |
Are there any problems with the standard analysis and plotting tools ? |
|
|
| Ran tutorials/demo.c in CINT on Windows - ran fine. Also compiled and ran $ROOTSYS/test, using Makefile.win32 and nmake. test/stress.exe failed the 8th test (trees split and compression modes) due to an assertion failure - sent message to ROOT forum. The failure was due to compiling in debug mode and all is fine, see RootTalk Forum: http://root.cern.ch/phpBB2/viewtopic.php?p=17519#17519. Ran guitest.exe and that seemed just fine. |
This page will change as fixes are made to some problems, and more tests are done.
**Error in <TBuffer::CheckByteCount>: object of class vector<Double32_t> read too few bytes: 26 instead of 46
Viewing Root Forum discussion: http://root.cern.ch/phpBB2/viewtopic.php?t=1437&highlight=few+bytes+vector
This error message occurs when the AcdRecon m_rowDocaCol has no entries - this is a vector<Double32_t>
David provided ROOT-central a simple example of this problem, and 5.14 has now been patched. We can obtain the source from ROOT's CVS repository with the branch tag: v5.14.00-patches
David has compiled and tested this release on linux. This release has now been built on windows as well.