| 1 |
mocchiutti |
1.1 |
v10r17 : used for processing at NA and MEPHI |
| 2 |
|
|
v10r18 : BUG FIXED: crash (ToF, probably also CALO and ORB) when reprocessing full file (probably any reprocessing) without extended tracks |
| 3 |
mocchiut |
1.2 |
New methods for tracks retrieving added in CaloLevel2.h, ToFLevel2.h and OrbitalInfo.h |
| 4 |
mocchiut |
1.3 |
v10r19 : testing purpose (ExtTrack processing only) |
| 5 |
|
|
v10r20 : BUG FIXED: crash due to new default in MySQL >=5.5.5, |
| 6 |
|
|
The query that is failing is something like: |
| 7 |
|
|
select ID_ROOT_L0, FROM_TIME, TO_TIME, EV_ROOT from GL_CALO_CALIB where SECTION=0 and VALIDATION=1 ORDER BY ABS(1322051607-FROM_TIME) asc limit 1 |
| 8 |
|
|
this issue and it seems that is related to the "NO_UNSIGNED_SUBTRACTION" mode of MySQL. |
| 9 |
|
|
"As of MySQL 5.5.5, overflow during numeric expression evaluation results in |
| 10 |
|
|
an error. For example, the largest signed BIGINT value is |
| 11 |
|
|
9223372036854775807, so the following expression produces an error." |
| 12 |
|
|
You can find other infos about this here |
| 13 |
|
|
http://dev.mysql.com/doc/refman/5.5/en/server-sql-mode.html |
| 14 |
|
|
essentialy now, by default, subtraction between integer operands produces an |
| 15 |
|
|
UNSIGNED result if any operand isUNSIGNED. This is something that I think is |
| 16 |
|
|
weird and in the code there is not such an assumption of course, I expect |
| 17 |
mocchiut |
1.4 |
that the subtraction between two unsigned integer COULD result in a negative number... |
| 18 |
|
|
v10r21 : BUG FIXED: sync lost between CAL/TOF in TrkCore when skipping events outside time window |
| 19 |
mocchiut |
1.6 |
v10r22 : Bug fixed: mysql api does not support multiple queries (introduce to fix MySQL >=5.5.5 bug, see v10r20). |
| 20 |
|
|
v10r23 : Bug fixed: NaN in IGRF F77 routine in rare situations |
Parent Directory
| 
Revision Log