Conquest 1.5.0 bug list

  • Here bug reports for 1.5.0 will be collected. Download 1.5.0beta3 here for testing: http://ingenium.home.xs4all.nl…r/dicomserver150beta3.zip


    Leftover bugs from earlier versions:


    1) add LittleEndianExplicit to GUI connecting to dgate. Futured. Partly in 1.5.0beta1

    2) Improve safety of sqlite by regular flushing. Futured, but users can change SqLite startup pragma for safer operation in 1.4.19c

    3) Check these changes: 20140608 mvh Added BYTE PCid to CStoreRSP::Write; This change is under review: pcid does not seem to get the right value

    4) brainix sample data fails to load from zip; maybe because of French folder names (works in 1.5.0)

    5) sql server and access database ODBC install only work for 32 bits of dgate (e.g. disable 64 bits or create ODBC entry by hand)

    6) last line of dicom.ini does not read if no CR (IndexDBF = 50) - futured

    7) Publish about Docker release of Lance

    8 ) GUI fails to run as user after install as service on server2012; does run as admin - not confirmed

    9) browse tab does not work if user of server service and server gui are differerent.

    10) The latest postgres no longer allows unencrypted passwords. Double check install code. (Fixed in 1.5.0)

    11) Bomberbug posted jpeg compressed US image 12345.zip that decompresses to green hue. Unfortunately not yet fixed due to lack of time. Fixed in 1.5.0.

    12) The web installer will not start dgate (with dicom.ini) installed with maklinux due to use of relative paths in the config. Maybe add reconfig button. (fixed in 1.5.0beta3).

    13) When opening a zip from web interface running on apache, the filename is not set correctly; it is when running from Ladle. Filename set in 1.5.0beta

    14) use BaseDir on dgate to find all folders (now many use current folder, complicating installer development). Futured. Implemented for web interface and lua in 1.5.0beta, fixes most requirements

    15) default webcodebase from newweb is wrong =http://127.0.0.1/cgi-bin/, also mode and patid2 are wrong, set to DefaultPage (fix in 1.5.0beta3)


    1.5.0-beta bugs (https://github.com/marcelvanherk/Conquest-DICOM-Server):


    16) Remove superfluous files from linux and src/doc folder. In 1.5.0beta

    17) Document source of contributions, in particular the Papaya redistributable

    18 ) web installer should copy html folders as well

    19) web viewers are not properly started when there are spaces in the patient ID

    20) Multiple instance matching incomplete

    21) Add return flag for character encoding of (non worklist) queries. In 1.5.0

    22) Document limitations of strings (e.g. 360 for process by xxx)

    23) The value sent to ImportConverter by %f is different than the actual filename written to the filesystem if a counter is in FilenameSyntax (documented in 1.5.0)

    24) Fedora uses /var/www, and nohup has an issue. Therefore provide specific web install scripts for other linux flavours, e.g. fedora.sh. Once started the same web installer should run. Fedura also requires "sudo setenforce Permissive" for the web server and install to work; also acrnema.map must be copied locally to the cgi-bin/newweb folder rather than pointing to the server folder.

    25) I was also expecting DicomObject:Read('stu\ser\sop') to work with virtual servers but I tested it again and sadly no luck there (unconfirmed)

    26) Use of 'Rows' as field name gives issues on newer SQL servers, modify for new installs. In 1.5.0beta

    27) Ladleport is not correctly configured in server GUI. In 1.5.0beta

    28 ) Several compilation errors in linux related to printf formatting and requirement of uncompcomv in dgatesmall. In 1.5.0beta

    29) starting zbs debugger fails under linux; does not find e.g. lua/socket.lua. In 1.5.0beta

    30) Numbers in tag names confused scripting. In 1.5.0beta

    31) Add (debug) log of (dis)connecting in "forward to". In 1.5.0beta1

    32) Web installer on windows does not offer a 'regen' by itself

    33) Web install adds 'conquest' to exceptions but this is an image - probably not needed (is used in install, accepted in 1.5.0)

    34) Web installer on windows does not copy provided database libraries (in 1.5.0)

    35) constant MYSQL_PORT no longer exists in odbci.cpp. In 1.5.0beta1

    36) manual shows org and dest in delayed forward script command that do not exist. In 1.5.0beta1

    37) Command object not passed to many ImportConverters, especially RetrieveConverter. In 1.5.0beta1

    38 ) RetrieveConverter was run after reading SliceLimit so it could not change it. In 1.5.0beta1

    39) Provide binary dictionary for web interface - reduces CGI call run time by about 90 ms. In 1.5.0beta1

    40) save jpg/bmp/gif as with passed level & window is unreliable, sometimes skips intercept. Fixed in 1.5.0beta2

    41) Show header in wadoviewer (wado text) open as dgate.exe downloads instead of as text viewer (unreproducible in 1.5.0)

    42) lua.5.1.dll is not copied to newweb (in 1.5.0beta3)

    43) start.lua uses an os. call which is blocked in ladle (in 1.5.0beta3)

    44) to allow using iup under linux must link -llua5.1, put after .cpp and .o files (in 1.5.0beta3)

    45) listpatients.lua fails for dbaseIII (fix in 1.5.0beta3)

    46) iup dlls depend on msvrc80.dll which may not be on system (win2000 32bits test) (in 1.5.0beta3)

    47) linux install now requires sudo apt install lua5.1 and lua.5.1-dev (in 1.5.0beta3)

    48 ) lua files in import/export scripts now load from relative paths (in 1.5.0beta3)

    49) deanonymizer now works under linux (in 1.5.0beta3)

    50) win2000: regen on web install fails, dgateserv.exe fails (dgateserv.exe fix in 1.5.0beta4)

    51) Using 9999,0401 to return big results from servercommand does not allow 2-way transfer, changed to 9999,0403 (in 1.5.0beta3)

    52) Must set tempdir in dicom.ini in newweb for upload to work (in 1.5.0beta3)

    ---- beta3 release

    53) Still must fix dgateserv for different OS and with newer compiler (use old one in 1.5.0beta4)

    54) update version numner and release date in windows exe (in 1.5.0beta4)

    55) enable jpeg support by default (in 1.5.0)

    56) update manual and list of software changes (in 1.5.0)

    57) maklinux must delete dicom.ini when changing database in linux (in 1.5.0beta4)

    58 ) make sure JSON.lua is written as json.lua (in 1.5.0beta4)

    59) used wrong quotes in deanomyse script sql processing for (*), must use ' (in 1.5.0beta4)

    60) Make sure Global.basedir is not followed by ds (in 1.5.0beta4)

    61) use mariadb: update manual (install mariadb, libmariadbclient-dev), fix maklinux (e.g. -lmariadbclient); UIDMODS case mixup (in 1.5.0beta4)

    62) when a move fails, the progress bar does not disappear (in 1.5.0)

    63) uploadsql (many queries in one file) is not compatibile with mariadb (fixed in 1.5.0beta4)

    64) Note: tempfile() should not be used in CGI applications

    65) attachfile does not process zips properly in Linux (fixed in 1.5.0beta4)

    66) need command to flush anoymisation cache after modifying UIDMODS (in 1.5.0)


    Unconfirmed bugs:

    U1) overlapvirtualget=0 breaks the virtual server operation (hardw) - Not confirmed by testing

    U2) color print server does not work. - futured, awaiting more information.

    U3) Forward compressed as j2(?) does not set transfer syntax UID correctly in transmitted image. This is an issue on dcm4che but has not been seen in conquest to conquest transmission.

    U4) enpq reported random failure in 'process by after' mechanism (not confirmed)

    Marcel van Herk is developer of the Conquest DICOM server together with Lambert Zijp.

    Edited 49 times, last by marcelvanherk: Split of unconfirmed bugs Note: should catagorise bugs for overview 1.5.0 release ().

  • Here feature requests will be collected:


    1) install second server when service of first runs gives trouble - do an echo on the install button and fail if found - futured

    2) wish: OperatorConsole.printf ("*** multiplex: connection terminated [%s]\n", CallingApTitle); Futured

    3) Implement web interface for batch anonymisation - futured

    4) Make max PDU size configurable. Futured.

    5) request for time stamp in ms. Futured.

    6) Provide TCP based waiting function to better control the move button and things like regen and such

    7) Finish and document dicomget dicomstore dicomprint, and [lua]printserver. In 1.5.0beta

    8 ) Use conquestdicomserver.lua from gui

    9) Modify lua weasis starter for non-jnlp version

    10) Performance improvement in array class by lsp. In 1.5.0beta2

    11) optionally allow 2 viewers from web client, add selection in web installer. One selection in 1.4.19c1

    12) Use flag to avoid error messages from install controller (*** Not enough rights; ***Error loading)

    13) Provide systemctl and dgateserv service installers in web install

    14) Scriptable listofoldestpatients to allow scripted cleanup. In 1.5.0beta

    15) Add json option as Serialize(true). In 1.5.0beta1

    16) Add multithreading on delayed forwarders to parallelize multiple forwards. In 1.5.0beta1

    17) Add 'split' clause to delayed forwarders to allow parallelize a single forward . In 1.5.0beta1

    18 ) Add rest-like control and query services to newweb interface. In 1.5.0beta1

    19) Load lua5.1 as dll to allow using clibs such as iup. Partly in 1.5.0beta2

    20) console for lua scripting. In 1.5.0beta2

  • General todo list


    Test included mysql and postgres DLLs with recent db versions

    Make sure version numbers are consistent in manuals and components

    Update bug list, splitting bugs into catagories (e.g. installation)

    Document limitations (e.g. codepages)

    Document command line installation of conquest on windows

  • Hi everybody
    Bug on Philips IntelliSpace and GE workstations.

    I checked on different version of 1.5.0


    Stripes appear when trying to transfer images from the server to Philips and GE workstations.


    The images themselves are not damaged (checked by pulling from the archive)


    Efilm and Radiant IntelliSpace and GE with stripes example below.


    On version 1419d1 one small line

  • Ok,


    it is in Linux - we are talking about the same issue just reported by mrJohnBravo. Can you help test? If you resend the image, is the problem exactly the same?


    And can you post an anonymized image for me? I am curious what exactly happens, that maybe allows to track the problem. Lets focus on the 1.5.0 version for now.


    The pixel values look similar so maybe a small block of the image is sent twice. In both versions the error is around 55 pixels. Shifts by about 15 pixels. But these screenshots may be zoomed.


    Marcel

  • Hi,


    I compared the images, and in all cases the errors occur fairly precisely at (3,5,7,9,11....)*4096 bytes from the start of group 8 in the DICOM data stream with 8192(+12?) bytes between errors.


    Would be interested to see a log (conquest side). In particular the PDU size that is used to transmit.


    Marcel

  • Hi,


    As service detailed logs are not recorded. Can you stop the service and then run dgate as


    "./dgate -v"


    and send one image again? I would suggest to use image IMG-0001-00001_anon.dcm for all our testing. I can add debug prints into the code and I know exactly where teh erorrs occur. In many other images there are zeros at the edge where the errors are not detectable.


    Marcel

  • Hi,


    this will require several iterations! Once I know the PDU size I can run the same send at home under Windows and Linux and then we can compare byte for byte. This bug, however, does not appear in all Linux versions.


    Marcel

  • Hi,


    I made the following change to line 509~530 in dicomlib/socket.cxx (added variable a to get the result form send and added the printf):


    I compiled and then captured the output when sending IMG-0001-00001_anon.dcm from the compiled server to another server with all the settings I expect to be active in Linux.


    This is the output, starting at the actual send operation:

    Can you try to do the same when you send to the offending system? I am especially interested in the bit after "Image Loaded from Read Ahead Thread, returning TRUE" for a single image. The image is 560x560 pixels.


    Marcel

Participate now!

Don’t have an account yet? Register yourself now and be a part of our community!