Timeline
07/03/13:
07/02/13:
- 17:57 Changeset [16945] by
- 10:22 Changeset [16944] by
- Fixed FITS conformance issue for .fz files when numrows=0
06/30/13:
06/28/13:
- 14:32 Changeset [16941] by
- Moved the conversion from dac to volt to its own function - for later easier replacement of the calibration.
- 14:10 Changeset [16940] by
- Added a comment.
- 11:57 Changeset [16939] by
- got rid of over 1024 start cell values by adding modulo
06/27/13:
- 23:27 Changeset [16938] by
- Added begining and end of line to output.
- 22:21 Changeset [16937] by
- Some fixes for the display of command completion results; fixed the removal of double entries from the history; added Home and End to key bindings output
- 20:11 Changeset [16936] by
- Added a very simple command completion for java.
- 13:10 Changeset [16935] by
- god rid of ambiguous cast error
- 04:13 Changeset [16934] by
- 03:14 Changeset [16933] by
- 03:01 Changeset [16932] by
- 03:00 Changeset [16931] by
- 02:59 Changeset [16930] by
- 02:57 Changeset [16929] by
- 02:55 Changeset [16928] by
- 02:55 Changeset [16927] by
- 02:33 Changeset [16926] by
- 02:29 Changeset [16925] by
- 02:15 Changeset [16924] by
- 02:14 Changeset [16923] by
- 02:13 Changeset [16922] by
- 02:06 Changeset [16921] by
- 02:05 Changeset [16920] by
- 02:03 Changeset [16919] by
06/26/13:
- 20:30 Changeset [16918] by
- 18:05 Changeset [16917] by
- Some consistency... dorce update is the strongest call we have to update the prompt as far as I understand.
- 18:00 Changeset [16916] by
- Not only when the screen geometry changes, but also when the prompt change we have to go back to the beginning of the line and re-display the prompt, this is because readline cannot determine the size of the prompt well, because of the ascii control condes
- 03:08 Changeset [16915] by
- 03:04 Changeset [16914] by
- 03:00 Changeset [16913] by
- 02:55 Changeset [16912] by
06/25/13:
- 15:38 Changeset [16911] by
- 15:33 Changeset [16910] by
- 15:11 Changeset [16909] by
- 15:08 Changeset [16908] by
- 15:05 Changeset [16907] by
- 15:04 Changeset [16906] by
- 14:41 Changeset [16905] by
- 14:32 Changeset [16904] by
- 14:28 Changeset [16903] by
- 14:26 Changeset [16902] by
- 14:20 Changeset [16901] by
- 14:18 Changeset [16900] by
- 02:46 Changeset [16899] by
06/24/13:
- 16:04 Changeset [16898] by
- Propagated the new naming of enums to other files....
- 15:57 Changeset [16897] by
- Removed bytes swapping for RAWSUM
- 15:16 Changeset [16896] by
- Added RAWSUM calculation, merged fits.h
06/23/13:
- 09:57 Changeset [16895] by
- 09:54 Changeset [16894] by
- 00:10 Changeset [16893] by
- Fixed a typo, added the missing cast and removed also the MRawEvtData from the list to avoid confusion in MReadFitsFile when the size of the array is dtermined.
06/22/13:
- 16:48 Changeset [16892] by
- 16:47 Changeset [16891] by
- Added a check for underflows, for the case the gains are read as offsets... only for checking purpose.
- 16:28 Changeset [16890] by
- Consistency of member names
- 15:31 Changeset [16889] by
- Some cleaning to easier understand the code.
- 15:30 Changeset [16888] by
- Given proper names to the enums... we are not programming Fortran style anymore.
- 11:45 Changeset [16887] by
- Added some smoothing to the deviation responsible for the OnTrack/Tracking; allow a second auto resume after five seconds.
- 10:07 Changeset [16886] by
- Make sure no consietncy check is done for the last file.
06/21/13:
- 22:30 Changeset [16885] by
- Slightly modified appearance of log-stream.
- 22:29 Changeset [16884] by
- Fixed two compiler warnings.
- 22:29 Changeset [16883] by
- Fixed two typos.
- 17:08 Changeset [16882] by
- added correct checksum calculation for compressed fits reader (zfits)
- 16:05 Changeset [16881] by
- Replaced some loops with cout.write for better performance
- 16:02 Changeset [16880] by
- Redisplay has to check if this is an update due to change of terminal size. If this is the case the cursor must be moved back to the beginning of the line. For convenience we also clean the line (to remove the last charcter in the line, which (maybe fpor safty reason to avoid newlines) readline doesn't remove.
- 16:01 Changeset [16879] by
- Removed the newline before the prompt. That's were the problems with the display came from.
- 16:00 Changeset [16878] by
- Removed the obsolete check for 'use striut'
- 14:41 Changeset [16877] by
- Removed a stray debug output.
- 13:38 Changeset [16876] by
- removed wrong error tags in log files
06/20/13:
- 21:54 Changeset [16875] by
- Fixed an issue with the detection of the 'camera' argument.
- 21:54 Changeset [16874] by
- Direct the output in case of wrong states to the dim log.
- 05:02 Changeset [16873] by
06/19/13:
- 17:14 Changeset [16872] by
- Updated EventHook to match virtual function definition.
- 17:10 Changeset [16871] by
- Use a to_string instead of an ostringstream for conversion
- 17:09 Changeset [16870] by
- When we output something ourselves before the event hook is called in readline, we have to signal readline that we have started a new row, so that readline can properly update the prompt. I hope this finally solves the flickering issue on my laptop and/or the problematic display in the fadctrl.
- 17:09 Changeset [16869] by
- When we output something ourselves before the event hook is called in readline, we have to signal readline that we have started a new row, so that readline can properly update the prompt. I hope this finally solves the flickering issue on my laptop and/or the problematic display in the fadctrl.
- 13:14 Changeset [16868] by
- Added 'home' and 'end' key to fixed key bindings, this makes it work in screen. Updated the EventHook again. This version still gives some flicker on my laptop, but it works on La Palma and at least the line buffer doesn't jump.
- 12:47 Changeset [16867] by
- got rid of defines
- 11:21 Changeset [16866] by
- Removed the output of the ratescans to an ascii file. Sending it to the datalogger has proven to be stable.
- 11:20 Changeset [16865] by
- Increased the time to wait for the data of a ftm reply to a macroscopic time. There is no need to be picky about the receiving time and it can happen that other events in the queue are processed in between.
06/18/13:
- 15:58 Changeset [16864] by
- Fixed the twilight check
- 03:50 Changeset [16863] by
- bugfix for case histogram in star does not exist, update code for nightly files
06/17/13:
- 16:33 Changeset [16862] by
- Deleteing the line does not seem necessary anymore.
- 14:16 Changeset [16861] by
- Indentation cosmetics.
- 13:58 Changeset [16860] by
- Added 'use strict'
- 13:57 Changeset [16859] by
- Added 'use strict'
- 13:57 Changeset [16858] by
- Fixed a typo: Should be var, not int
- 13:57 Changeset [16857] by
- Added the 'use strict'
06/15/13:
- 19:14 Changeset [16856] by
- Forgot this one...
- 19:10 Changeset [16855] by
- Fixed a couple of issues including missing setting of badbit, broken returns and compilation with CINT, added some const qualifiers
- 18:18 Changeset [16854] by
- The length of the bytes released must be converted from short to bytes.
- 18:11 Changeset [16853] by
- Replaced some console.out by dim.log; take only 1min runs during twilight conditions; added some code to properly print the next observation, also when it is changed in the db; improved some output.
- 18:08 Changeset [16852] by
- Some changes to the output, replaced some console.out by dim.log
- 18:08 Changeset [16851] by
- Changed underflow level from 650 to 600; replaced some console.out by dim.log
- 04:45 Changeset [16850] by
- fixed bug in getting source name from drive file
06/14/13:
- 17:13 Changeset [16849] by
- Added ln_get_lunar_earth_dist
- 14:56 Changeset [16848] by
- Fixed reconstructed file filename
- 14:53 Changeset [16847] by
- fixed -again- checksum calculation while reading... hopefully
- 14:38 Changeset [16846] by
- reverted the change: this did not work...
- 14:21 Changeset [16845] by
- Added guard handling of case where startCells not found and Drs requested
- 14:20 Changeset [16844] by
- fixed checksum calculation while reading... hopefully
- 11:27 Changeset [16843] by
- Changed the way how the returned maximum is calculated to get a better estimate of what a good shower is to be displayed (the old version was actually also buggy)
- 11:16 Changeset [16842] by
- Updated some output.
- 11:15 Changeset [16841] by
- Added the possibility to interrupt the script peacefully; changed to the output; added an additonal check for sun-rise if a long operation took place like the drs calib; added a STOP before Tracking position is changed (this is not ideal, but currently the only way to ensure that we can easily detect the change to OnTrack and are not confused by the current OnTrack)
- 11:15 Changeset [16840] by
- bug fix and emplace_back
- 11:12 Changeset [16839] by
- Some updates on the output; removed an obsolete wait for MCP Idle.
- 11:11 Changeset [16838] by
- Try a CLOSE, if the LID is in the state 'Unknown' - might still need some interaction of the shifter.
- 11:06 Changeset [16837] by
- Added 'Overcurrent' as possible status.
- 10:56 Changeset [16836] by
- Use make_shared where possible.
- 10:55 Changeset [16835] by
- Moved the Write command back to the main thread, to ensure fast output on the console - the file access now has its own queue to decouple disk load and waiting time.
- 10:54 Changeset [16834] by
- Changed the time for an update of the EVENTS service to a number whihc is not a mutiple of 1s, to avoid coincidences with interleaved events (just cosmetics); added a fix which ensures that the replacement of the calibration in runClose is not affecting events which are currenly processed in another thread; use make_shared where possible.
- 10:52 Changeset [16833] by
- The disposed handle fInterruptCallback must also be cleared, otherwise it might still be accessible. Removed a stray debug output.
- 10:51 Changeset [16832] by
- Using a queue, to be able to decouple writing to the log-files from the disk load, but at the same time ensure that the data is flushed immediately and not buffered for a long time.
- 10:36 Changeset [16831] by
- Do not print any thresholds when the ratecontrol is not in operation.
- 10:36 Changeset [16830] by
- Print a warning if an interrupt request is received but no script is running.
06/13/13:
- 20:52 Changeset [16829] by
- Updated.
- 20:51 Changeset [16828] by
- In Configured2 we do just wait.
- 20:51 Changeset [16827] by
- Make sure all boards are connected.
- 17:11 Changeset [16826] by
- added 0 length files handling
- 16:04 Changeset [16825] by
- Added heap consistency check. Not yet enabled
- 13:28 Changeset [16824] by
- Updated usage information.
- 13:27 Changeset [16823] by
- Added libTool to showlog libraries.
- 13:26 Changeset [16822] by
- Removed the day-tag. It looks a bit odd, because we have night-wise files anyway. Added the possibility that begin is after end (e.g. when displaying data from 23:00 to 01:00. Added a conversion of numbers to paths for easy access of log files in standard locations.
- 13:10 Changeset [16821] by
- Switch off database use for showlog as a default.
- 08:55 Changeset [16820] by
- Modified block header's sequence scheme
06/12/13:
- 22:14 Changeset [16819] by
- Added new program showlog
- 22:13 Changeset [16818] by
- Added new program showlog
- 22:11 Changeset [16817] by
- 18:32 Changeset [16816] by
- 15:56 Changeset [16815] by
- replaced compression enum handling
- 15:56 Changeset [16814] by
- added block headers and changed compression mechanism
- 11:43 Changeset [16813] by
- Implemented some debug out to a log-file when the sql query fails.
- 11:42 Changeset [16812] by
- When a user was logged in, the setting of a collaborator failed.. this is fixed now.
- 10:59 Changeset [16811] by
- oops, got rid of cout
- 10:58 Changeset [16810] by
- Added tile header to zfits
- 10:30 Changeset [16809] by
- Added tile num rows tag
06/11/13:
- 18:25 Changeset [16808] by
- kSM_KeepState must also be accounted for Events.
- 18:20 Changeset [16807] by
- Second argument was missing in all SetCurrentState (although I hope they will disappear soon)
- 16:44 Changeset [16806] by
- Fixed HEAP size calculation bug
- 16:37 Changeset [16805] by
- Added TILE marker and size between tile data
06/10/13:
- 17:25 Changeset [16804] by
- Updated the description of setInterruptHandler and added triggerInterrupt
- 17:23 Changeset [16803] by
- If the first argument does contain an equal sign consider it to be the first argument rather than the filename.
- 17:22 Changeset [16802] by
- Allow arguments to be given to the interrupt handler and added a possibility to trigger an interrupt also from within the script; some modification to the way code is executed so that return values are only printed really when some code is finished from the command line, but not for include or when the main program is run.
- 17:20 Changeset [16801] by
- Replaced the new invalid FTM::State::kConfigured with Configured1 and Configured2
- 12:38 Changeset [16800] by
- Changed 90V to 80V
- 12:33 Changeset [16799] by
- fixed fitsCompressor's copy verification issue. I was closing the file before the destructor, hence padding not always added. Fixed now.
- 12:04 Changeset [16798] by
- Reoragnized the logging output, to get closer to a final version sending everything to the log stream; ftmctrl now in Valid not in Idle; added check for the clock conditioner.
- 12:03 Changeset [16797] by
- Improved output when a command is sent.
- 12:02 Changeset [16796] by
- Wait for ftmctrl to be in Valid not Idle; changed the level to detect an underflow.
- 12:02 Changeset [16795] by
- Added Configuring3 state; improved output whena command is sent:
- 12:01 Changeset [16794] by
- output some information so that it is aparent which FTU gives the problem.
- 12:01 Changeset [16793] by
- Fixed a problem with the 32bit shift operator when evaluating the 40bit incomplete flag; wait for ftmctrl to be in Valid rather than Idle.
- 12:00 Changeset [16792] by
- There is not much we can do in Configured2 other than waiting for the trigger to becoome enabled.
- 11:59 Changeset [16791] by
- Configured is now Configured1 and Configured2
- 11:58 Changeset [16790] by
- Added a chck for the clock conditioner; ftmctrl should now be in Valid not anymore in Idle.
- 11:57 Changeset [16789] by
- Debugging output seems not needed anymore.
- 11:57 Changeset [16788] by
- 11:57 Changeset [16787] by
- Improved the output whenever a command is sent.
- 11:54 Changeset [16786] by
- Make sure no threshold is set initially below the limit (the limit is set to ThresholdMin, so Thresholdreference is not really the correct reference here); directly use fThreshold for the dim commands, therefore replaced uint16 by int32; only send threshold if they have changed; improved output
- 11:36 Changeset [16785] by
- Replaced kConfigured with kConfigured1 according to the recent change in the ftmctrl.
- 11:36 Changeset [16784] by
- Adapted the state checks according to the recent change in the ftmctrl.
- 11:36 Changeset [16783] by
- Replaced kConfigured by kConfigured1. This is the state in which sending thresholds is allowed. As soon as the trigger is enabled we go to kConfugred2. Since the ftm ignores trigger on while still configuring the FTUs, the trigger on is resent until we get a proper answer.
- 11:34 Changeset [16782] by
- Added MagicLidar and Temperature to status page; replaced some numerical states with their enum counterparts.
06/09/13:
- 13:10 Changeset [16781] by
- Implemented the algorithms needed for the new type current control (there was a bug in the old one which actually stopped when the voltage was 'target voltage minus next step' which is about 1.0V. The new current control is not yet enabled, it is only there as code. It needs to be tested and then all depending values need adaption.
- 13:08 Changeset [16780] by
- Of course the server has to run the central loop... looks like it was broken since months.
- 12:51 Changeset [16779] by
- A less CPU hungry version of the msg queue using C++11 techniques.
06/08/13:
- 12:56 Changeset [16778] by
- Set all thresholds at once after the steps have been calculated using the new SET_SELCTED_THRESHOLDS
- 12:55 Changeset [16777] by
- Added the long awaited customized template for vector<T>
- 12:55 Changeset [16776] by
- Implemented SET_SELECTED_THRESHOLDS to allow setting severalö thresholds during data taking at once... this should have a slightly higher performance than requesting a static data block after each setting. Maybe the difference is small, but nevertheless.
- 12:30 Changeset [16775] by
- Introduced a new state kConfiguring3 which checks if the messages to enable the trigger line were really delivered before switching to kConfigured. This never seems necessary, but maybe only because there is a natural delay until the mcp reacts and enables the trigger; removed the ConfigTimer. It seems obsolete with the recent changes in the output buffer queue
- 12:28 Changeset [16774] by
- After five minutes of inactivity, sleep 10ms instead of 1us if all idle conditions are fullfilled.
- 12:28 Changeset [16773] by
- Introduced the new state kConfiguring3
- 12:27 Changeset [16772] by
- When calibrated by currents after the calibration was started, wait until a static data block with the correct currents has been received before setting GlobalThresholdSet. This is to ensure that the thresholds are set _before_ the trigger is started; for better performance, the new command SET_ALL_THRESHOLDS is used.
- 12:25 Changeset [16771] by
- When compiling a new static buffer to be sent to the FTM, not the last received buffer has to be used but the last sent buffer, this is because a new configuration might be in the queue already, but we have not yet received the answer. To avoid race conditions, the sent buffer is now filled at connection time and used to update the configuration; also a new state was introduced to signal (for debugging mainly) whether the last sent and the last received static data are different (this does not include the thresholds); for convenience in the ratecontrol: added a SET_ALL_THRESHOLDS
- 12:22 Changeset [16770] by
- We don't want to dispatch the event (blokcing) but to post the event (non-blocking).
- 12:22 Changeset [16769] by
- There is no need to make the scheduling of the async write asynchronously, so I schedule the async write now directly; this also removed the send queue, which was primarily introduced to make the tranmission to the FTM serial, but this is not needed anymore; to keep track of the send queue size and for debugging, a counter for the messages in the send buffer has been introduced
- 12:21 Changeset [16768] by
- There is no need to make the scheduling of the async write asynchronously, so I schedule the async write now directly; this also removed the send queue, which was primarily introduced to make the tranmission to the FTM serial, but this is not needed anymore; to keep track of the send queue size and for debugging, a counter for the messages in the send buffer has been introduced
Note:
See TracTimeline
for information about the timeline view.
