Opened 11 years ago
Last modified 10 years ago
#332 accepted defect
invert_tool hangs when SO fit range is only one point
| Reported by: | Stig Syndergaard | Owned by: | Stig Syndergaard |
|---|---|---|---|
| Priority: | critical | Milestone: | DMI ROPP developments |
| Component: | ropp_pp | Version: | 6.1 |
| Keywords: | Cc: |
Description
ropp_pp_invert_tool hangs when processing atm20130701_050945_M02_1030493022_N0020_XXXX.nc
SO fit range consists of only 1 point
Attachments (3)
Change history (5)
by , 11 years ago
| Attachment: | atm20130701_050945_M02_1030493022_N0020_XXXX.nc added |
|---|
by , 10 years ago
| Attachment: | atm20131114_001654_M02_1230506601_N0021_XXXX.nc added |
|---|
by , 10 years ago
| Attachment: | atm20131114_210736_M02_1020506688_N0021_XXXX.nc added |
|---|
comment:1 by , 10 years ago
| Status: | new → accepted |
|---|
Note:
See TracTickets
for help on using tickets.
The issue was reproduced with two different atm files. The key problem is that the variable i_ltr (and others like that) in ropp_pp_ionospheric_correction.f90 becomes zero, because it is a short profile. Later, ba_is(i_ltr:nh) is assigned a value, but i_ltr = 0 is outside the array bounds. Code crashes differently for the two atm files.