jpr [Wed, 12 Jan 2005 15:22:23 +0000 (15:22 +0000)]
2005-01-12 Jean-Pierre Roux <jpr@creatis.univ-lyon1.fr>
* FIX : Old quick and dirty 'optimistic' heuristic to deal with
Big Endian Transfer Syntax supposed the group following 0002
begins alawys by element 0000 (element 0000 is *optional*)
To avoid further troubles, let's be pessimistic, and use
Document::HandleOutOfGroup0002() method
malaterre [Tue, 11 Jan 2005 16:44:42 +0000 (16:44 +0000)]
ENH: Untangle the transfer syntax from the Document. The Document can only read a string and can only ask gdcm::TS what are the property of the transfer syntax
jpr [Tue, 11 Jan 2005 11:37:12 +0000 (11:37 +0000)]
2005-01-11 Jean-Pierre Roux <jpr@creatis.univ-lyon1.fr>
* Replace confusing name SwitchSwapToBigEndian name by SwitchByteSwapCode
* Add the Document::HandleOutOfGroup0002(uint16_t group) method
to swap the Swap Code, at parsing time, when we come out of group 0002
and Transfer Syntax is Big Endian
We should be able to read now 'true DICOM' Big Endian coded images
* Add optional run time SetDebugOn (last param) for PrintHeader, PrintFile
* Replace the french 'Transfert Syntax' by the english 'Transfer Syntax'
malaterre [Tue, 11 Jan 2005 00:37:41 +0000 (00:37 +0000)]
ENH: Adding a new method on TS: IsTransferSyntax. Where you pass in a string and it return if this is a valid Transfer Syntax. So far gdcm is broken since 1.2.840.113619.5.2 is not handled
malaterre [Mon, 10 Jan 2005 20:52:39 +0000 (20:52 +0000)]
ENH: Finally remove previous solution. Found a much better (and simplier approach) that -seems to- works in every configuration: Win32 static/shared and MacOSX (within ITK or not)
malaterre [Mon, 10 Jan 2005 19:23:31 +0000 (19:23 +0000)]
BUG: Hopefully fix the bug on MacOSX and static global initialization. This is the only -minor- patch I found that both please MacOSX and VS6. Other patch would be to declare a const std::string & GDCM_UNFOUND() instead, but this would require a lot of change. Another approach wqould be that Global create those string and GDCM_UNFOUND becomes a pointer to a string (which gdcm::Global would destroy at exit).
malaterre [Sat, 8 Jan 2005 23:14:05 +0000 (23:14 +0000)]
Two things at once: start removing hack for old ostrstream which was really bad (and shouldn't work anyway). Plus attempt to fix problem with multithread on Mac where string object are not created properly.
jpr [Sat, 8 Jan 2005 15:03:56 +0000 (15:03 +0000)]
According to Benoit's suggestion, and without any objection from anybody
- methods SetxxxByNumber and GetxxxByNumber renamed as Setxxx and Get xxx
- methods Dict::Print() and Dict::PrintByKey() merged into Dict::Print()
- method gdcmDicomDirObject::GetEntry() renamed as
gdcmDicomDirObject::GetEntryHT() to avoid confusion (and compile error)
regrain [Fri, 7 Jan 2005 16:45:51 +0000 (16:45 +0000)]
* src/gdcmDocEntry.[h|cxx] : now the ReadLength is the length of the datas
to read in the file... and only it ! Length is the efective length of the
datas in the DocEntry structure
-- BeNours
jpr [Fri, 7 Jan 2005 16:34:47 +0000 (16:34 +0000)]
Removal of useless groupNameAbbreviations2004.dic
Add dicomGroupNameAbbreviations.dic
should be used to write a method to replace the former
Dict::GetDictEntryNamesByCategory()
malaterre [Fri, 7 Jan 2005 16:26:11 +0000 (16:26 +0000)]
ENH: Change the gdcmDebug approach. Remov the global static debug 'dbg'. And now use a static function call instead, with a global variable. This is much closer to the VTK approach. Hopefully should be bulletproof and easier to use...hopefully
jpr [Thu, 6 Jan 2005 17:08:05 +0000 (17:08 +0000)]
Doxygenation
+ modif Coding style :
According with Benoit (and Doxygen) we shall use, just like anybody :
T *foo
or
T &foo
and NOT
T* foo
or
T & foo
(that was gdcm-special)