#426 TestPrinter1 fails

[Manon Linder]

Hi,

With GDCM 2.6.8, under visual studio 2013 win64, the test 155 - TestPrinter1 fails and returns:

Problem with : 012345.002.050.dcm a5b2cb7d27bf0565296ae0391c07dd54 vs 37f2807ca5397f121aa71abe8675498f
Problem with : AMIInvalidPrivateDefinedLengthSQasUN.dcm 6c18a8fdd766660fbb475c05d0611634 vs 49959a36e65708b5251388ba4410cfa7
Problem with : BugGDCM2_UndefItemWrongVL.dcm 9ddd03e5fd29f92bd02c3d3812b4e93e vs 9efdfdc3d54113dca430c2234654aeb0
Problem with : D_CLUNIE_CT1_J2KI.dcm a46ca467a0df5e1c71c8017b6b6768e7 vs e107459658f1576d259292dc58ae8b96
Problem with : D_CLUNIE_CT1_J2KR.dcm d2698545b07ad721faa71fdc1cf4a35d vs ce515c915aefbf6d20480a98ff303186
Problem with : D_CLUNIE_CT1_JLSL.dcm 8984d6319cd913d0f88fdb61989412f2 vs 15cfbd6ff1ef2887d84d83ad2067cc36
Problem with : D_CLUNIE_CT1_JLSN.dcm bb52d9f7a9116c87103066dd18a60e68 vs e11e88445d1b6bf83677c51da24618f5
Problem with : D_CLUNIE_CT1_JPLL.dcm 78f619a764a94a2c2bdc084d53ad2577 vs 7697a6e190270aebf76d38e79a9ce5ef
Problem with : D_CLUNIE_CT1_RLE.dcm 4aaa4a602d7a1fd80c023acba36d6a84 vs e46fd985b0406906cb01855df8ff8b5b
Problem with : D_CLUNIE_MR3_JPLL.dcm 4ac28edcc53e3c5a2b14cee29ac4b350 vs 66cd02e3a58a8c741c7d0f37c1b77232
Problem with : D_CLUNIE_MR3_JPLY.dcm 1279c4aaacd1a058fc4b101cbfebec54 vs 53a1c72c555be09407940d7e2e9737a3
Problem with : D_CLUNIE_MR3_RLE.dcm d6ca154c96d206995023f7349f6ca861 vs 1c3fc115fb55d24252fe9fb88922d859
Problem with : GE_CT_With_Private_compressed-icon.dcm 9e9f42e825db2951519320c2e907d936 vs a53c7847808cf3b31cf005f7b81957ee
Problem with : GE_GENESIS-16-MONO2-Uncompressed-UnusualVR.dcm 7920bf7f1269ba54dd8d0e0d375dc4b9 vs 27c889fca90bbdbff833d64686d1636c
Problem with : GE_LOGIQBook-8-RGB-HugePreview.dcm bb5d500a0391a399b035968496b6fd5d vs fd394f2324090e3a0aa48c22705b81d1
Problem with : GE_MR_0025xx1bProtocolDataBlock.dcm 1ad963ca1e91932bb1ac58f3dddd981f vs 747582d09f1a2ad8ead035efb3eb45f9
Problem with : GE_RHAPSODE-16-MONO2-JPEG-Fragments.dcm d5efa34d8091e1ad04683eefb41f33c7 vs 846d79f208510a197e2b6155d3e0a2bc
Problem with : MR_GE_with_Private_Compressed_Icon_0009_1110.dcm 7a61574251f8788a11cfe06ddc109bd9 vs 047ab9a463da6d824f70e3420c783235
Problem with : MR_Philips_Intera_PrivateSequenceImplicitVR.dcm 3d24233b19788d349f64b7b267723186 vs 1feb8aecac038cbb2f2d8c54113a9c91
Problem with : MR_Philips_Intera_SwitchIndianess_noLgtSQItem_in_trueLgtSeq.dcm 1693399bc1edb7cdc494095a27201eab vs 5e20012df11e4b1871a49cbdf79ab204
Problem with : MR_Spectroscopy_SIEMENS_OF.dcm acbf80a7c610b44caebd52bc22de74f9 vs 1fd7b67f7451e70bff6520cbedbe9261
Problem with : PhilipsInteraSeqTermInvLen.dcm fdd8a53915712a924ac74c26ec4d1051 vs 2c954b0adfc3edc2840011858d6fe0d1
Problem with : PHILIPS_Intera-16-MONO2-Uncompress.dcm 1f198bdb09f5e52dc3f9ad1dfced8a4d vs 721e260ea162b27bf26e3626f46da3b5
Problem with : TheralysGDCM120Bug.dcm aa708d3b15e451c8367e7909257d9346 vs 74412b4c2b52089caa676bf5f59f9fd5
Problem with : US-GE-4AICL142.dcm 787ca80c8dd1aa619d5f85610862380b vs f789279c7cd54bac7991ab8ce342eacb

Do you have the same issue ?
Is the printmd5 that is wrong from TestPrinter1.cxx ? Is the dicom images that changed ? Or the gdcm::Testing::ComputeMD5() function ?

[Manon Linder]

In release, these tests do not work either.
3 - TestCryptographicMessageSyntax (Failed)
133 - TestImageChangeTransferSyntax1 (Failed)
134 - TestImageChangeTransferSyntax2 (Failed)
135 - TestImageChangeTransferSyntax3 (Failed)
136 - TestImageChangeTransferSyntax4 (Failed)
137 - TestImageChangeTransferSyntax6 (Failed)
138 - TestImageChangeTransferSyntax7 (Failed)
158 - TestImageReader (Failed)
160 - TestImageRegionReader1 (Failed)
161 - TestImageRegionReader2 (Failed)
162 - TestImageRegionReader3 (Failed)
165 - TestImageWriter (Failed)
204 - TestImageChangeTransferSyntax5 (Failed)

[Mathieu Malaterre]

I cannot reproduce the issue with TestPrinter1 on my linux box. Are you sure you updated your git submodule properly ?

Eg do you have:

$ cmake -E md5sum gdcm/Testing/Data/012345.002.050.dcm
c2fdbb35ba2a179939a8608e1320c7ac  gdcm/Testing/Data/012345.002.050.dcm

[Manon Linder]

I took gdcm with git clone on: git://git.code.sf.net/p/gdcm/gdcm
Then I do checkout for the v2.6.8.
For gdcm data : https://git.code.sf.net/p/gdcm/gdcmdata

When I do: cmake -E md5sum 012345.002.050.dcm
I get the same result than you:
c2fdbb35ba2a179939a8608e1320c7ac 012345.002.050.dcm

I perform the test in Win32 and x64.

In Debug:
Win32 | x64
----------------------------------------------------------------- | ----------
48 - TestReaderCanRead (Failed) | 155 - TestPrinter1 (Failed)
155 - TestPrinter1 (Failed) |
163 - TestImageReaderRandomEmpty (Timeout) |

In Release:
Win32 | x64
------------------------------------------------------------- | ----------
3 - TestCryptographicMessageSyntax (Failed) | 3 - TestCryptographicMessageSyntax (Failed)
48 - TestReaderCanRead (Failed) | 133 - TestImageChangeTransferSyntax1 (Failed)
155 - TestPrinter1 (Failed) | 134 - TestImageChangeTransferSyntax2 (Failed)
| 135 - TestImageChangeTransferSyntax3 (Failed)
| 136 - TestImageChangeTransferSyntax4 (Failed)
| 137 - TestImageChangeTransferSyntax6 (Failed)
| 138 - TestImageChangeTransferSyntax7 (Failed)
| 155 - TestPrinter1 (Failed)
| 158 - TestImageReader (Failed)
| 160 - TestImageRegionReader1 (Failed)
| 161 - TestImageRegionReader2 (Failed)
| 162 - TestImageRegionReader3 (Failed)
| 165 - TestImageWriter (Failed)
| 204 - TestImageChangeTransferSyntax5 (Failed)

[Mathieu Malaterre]

Hum, that's odd. Can you upload the following debug.txt file, simply run:

> bin/gdcmMSFFTests.exe TestPrinter1 gdcm/Testing/Data/012345.002.050.dcm > debug.txt

[Manon Linder]

The terminal displays:
Problem with : 012345.002.050.dcm a5b2cb7d27bf0565296ae0391c07dd54 vs 37f2807ca5397f121aa71abe8675498f

[Mathieu Malaterre]

Apparently under some conditions, the formatting for floating point has changed in Visual Studio:

@@ -103,7 +103,7 @@
 (0019,101a) LO [A ]                                               # 2,1 Last scan ras
 (0019,101b) DS [89.400002 ]                                       # 10,1 Last scan loc
 (0019,101e) DS [170.000000]                                       # 10,1 Display field of view
-(0019,105a) FL 1.31431e+09                                        # 4,1 Acquisition Duration
+(0019,105a) FL 1.31431e+009                                       # 4,1 Acquisition Duration
 (0019,107d) DS [0 ]                                               # 2,1 Second echo
 (0019,107e) SS 1                                                  # 2,1 Number of echoes
 (0019,107f) DS [0.000000]                                         # 8,1 Table delta

This means we need to find something smarter for TestPrinter unit test.

[Manon Linder]

OK I remain available to do the tests under visual for the next few weeks/months.

[Vivien Delmon]

The attached patch fix this issue. I was not sure if it would not be better to put the code directly in the Printer::Print function. Fixing just the test has a lower impact perhaps.

[Mathieu Malaterre]

TestPrinter1 seems to be working ok these days.