unit/regression test failures on osx 10.9.3 with hdf5 1.8.13

[Matthew Gidden]

Hi all,

It looks like we have some failures with this configuration. Unit test failures: https://gist.github.com/gidden/3143f27cd876c77a2357. Regression test failures: https://gist.github.com/gidden/18ad78ff657c59a369bd.

--

Matthew Gidden
Ph.D. Candidate, Nuclear Engineering

The University of Wisconsin -- Madison
Ph. 225.892.3192

[Matthew Gidden]

Cycamore regression tests similarly fail, FYI.

[Matthew Gidden]

Hmm... I get the same failures using 1.8.12. Note that our nightlies tests osx 10.8 and hdf5 1.8.4. Before going down that particular rabbit hole, does anyone else have a osx 10.9 partition? If so, what version of hdf5 are you using to pass tests?

Also, this is a blocking issue for my investigation of the nightly failures.

Thanks,

[Matthew Gidden]

I've also reproduced them with 1.8.11 and 1.8.4. I'm interested to hear what others (Katy?) might have experienced. FWIW, this is my cyclus --version:

matthews-air:cyclus gidden$ cyclus --version
Cyclus Core 1.0.0 (impl-1524-gcb06259)
Dependencies:
   Boost    1_55
   Coin-Cbc 2.8.7
   Hdf5     1.8.4-patch1
   Sqlite3  3.8.4.3
   xml2     2.9.0

[Anthony Scopatz]

Hello Matt, 

I think that you and Katy are the only ones with Mavericks installed.  I wish batlab had this as a target so that we could have CI on it.

If you are running with HDF5 1.8.13 and there are these errors, I doubt that regressing the HDF5 version will fix the problem and probably isn't a useful line of inquiry relative to how expensive it is.  

What happens when you run just the HDF5 tests?

What happens when you run "$ cyclus_unit_tests --gtest_repeat=1"?

Be Well

Anthony

--
You received this message because you are subscribed to the Google Groups "Cyclus Developers discussion group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cyclus-dev+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Matthew Gidden]

On Mon, Aug 4, 2014 at 10:45 AM, Anthony Scopatz <sco...@gmail.com> wrote:

Hello Matt, 

I think that you and Katy are the only ones with Mavericks installed.  I wish batlab had this as a target so that we could have CI on it.

If you are running with HDF5 1.8.13 and there are these errors, I doubt that regressing the HDF5 version will fix the problem and probably isn't a useful line of inquiry relative to how expensive it is.   

What happens when you run just the HDF5 tests?

same failures 

What happens when you run "$ cyclus_unit_tests --gtest_repeat=1"?

same failures 

[Anthony Scopatz]

Hi Matt, 

What happens when you run just one of the failing tests on its own? Also is the int size on MacOS 10.9 64bit on your machine?

Be Well

Anthony

[Matthew Gidden]

On Mon, Aug 4, 2014 at 11:02 AM, Anthony Scopatz <sco...@gmail.com> wrote:

Hi Matt, 

What happens when you run just one of the failing tests on its own?

Failure:

dyn-72-33-19-253:cyclus gidden$ cyclus_unit_tests --gtest_filter=Hdf5BackTests.ReadWriteVLMapVLStringVLString
Note: Google Test filter = Hdf5BackTests.ReadWriteVLMapVLStringVLString
[==========] Running 1 test from 1 test case.
[----------] Global test environment set-up.
[----------] 1 test from Hdf5BackTests
[ RUN      ] Hdf5BackTests.ReadWriteVLMapVLStringVLString
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 358 in H5Dopen2(): not found
    major: Dataset
    minor: Object not found
  #001: H5Gloc.c line 430 in H5G_loc_find(): can't find object
    major: Symbol table
    minor: Object not found
  #002: H5Gtraverse.c line 861 in H5G_traverse(): internal path traversal failed
    major: Symbol table
    minor: Object not found
  #003: H5Gtraverse.c line 641 in H5G_traverse_real(): traversal operator failed
    major: Symbol table
    minor: Callback failed
  #004: H5Gloc.c line 385 in H5G_loc_find_cb(): object '0M???' doesn't exist
    major: Symbol table
    minor: Object not found
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5A.c line 244 in H5Acreate2(): not a location
    major: Invalid arguments to routine
    minor: Inappropriate type
  #001: H5Gloc.c line 253 in H5G_loc(): invalid object ID
    major: Invalid arguments to routine
    minor: Bad value
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5A.c line 929 in H5Awrite(): not an attribute
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5A.c line 2288 in H5Aclose(): not an attribute
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 415 in H5Dclose(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 358 in H5Dopen2(): not found
    major: Dataset
    minor: Object not found
  #001: H5Gloc.c line 430 in H5G_loc_find(): can't find object
    major: Symbol table
    minor: Object not found
  #002: H5Gtraverse.c line 861 in H5G_traverse(): internal path traversal failed
    major: Symbol table
    minor: Object not found
  #003: H5Gtraverse.c line 641 in H5G_traverse_real(): traversal operator failed
    major: Symbol table
    minor: Callback failed
  #004: H5Gloc.c line 385 in H5G_loc_find_cb(): object 'vl_map_vl_string_vl_string' doesn't exist
    major: Symbol table
    minor: Object not found
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 548 in H5Dget_type(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 358 in H5Dopen2(): not found
    major: Dataset
    minor: Object not found
  #001: H5Gloc.c line 430 in H5G_loc_find(): can't find object
    major: Symbol table
    minor: Object not found
  #002: H5Gtraverse.c line 861 in H5G_traverse(): internal path traversal failed
    major: Symbol table
    minor: Object not found
  #003: H5Gtraverse.c line 641 in H5G_traverse_real(): traversal operator failed
    major: Symbol table
    minor: Callback failed
  #004: H5Gloc.c line 385 in H5G_loc_find_cb(): object 'vl_map_vl_string_vl_string' doesn't exist
    major: Symbol table
    minor: Object not found
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 1161 in H5Dset_extent(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 461 in H5Dget_space(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5Shyper.c line 6593 in H5Sselect_hyperslab(): not a data space
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5Dio.c line 225 in H5Dwrite(): can't prepare for writing data
    major: Dataset
    minor: Write failed
  #001: H5Dio.c line 257 in H5D__pre_write(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
unknown file: Failure
C++ exception with description "Failed to write to the HDF5 table:
  file      db.h5
  table     vl_map_vl_string_vl_string
  num. rows 2
  rowsize   36
    # Column 0
      dbtype: 7
      size:   16
      offset: 0
    # Column 1
      dbtype: 123
      size:   20
      offset: 16
" thrown in the test body.
[  FAILED  ] Hdf5BackTests.ReadWriteVLMapVLStringVLString (37 ms)
[----------] 1 test from Hdf5BackTests (37 ms total)
[----------] Global test environment tear-down
[==========] 1 test from 1 test case ran. (47 ms total)
[  PASSED  ] 0 tests.
[  FAILED  ] 1 test, listed below:
[  FAILED  ] Hdf5BackTests.ReadWriteVLMapVLStringVLString
 1 FAILED TEST

 

Also is the int size on MacOS 10.9 64bit on your machine?

 

The following program outputs '4':

#include <iostream>

int main() {
  std::cout << sizeof(1) << std::endl;
};

[Anthony Scopatz]

On Mon, Aug 4, 2014 at 9:15 AM, Matthew Gidden <gid...@wisc.edu> wrote:

On Mon, Aug 4, 2014 at 11:02 AM, Anthony Scopatz <sco...@gmail.com> wrote:

Hi Matt, 

What happens when you run just one of the failing tests on its own?

Failure:

Good to know.

Hmm this isn't super helpful because the 1 may be cast to short.  Could you re-rum please with sizeof(int)? Thanks!

Be Well

Anthony

[Matthew Gidden]

sizeof(int) == 4 

[Anthony Scopatz]

OK. Hmm. All of the ones that fail come from nested VLMaps and VLStrings.  The error comes from this line, which doesn't make a lot of sense in this context.  This indicates that either the VLMap dataset can't be opened or the VLString one cannot be.  Without diving into gdb or putting in some print statements it is hard to know which.

[Katy Huff]

FWIW I vaguely suspect it's the string. So far, all of the mavericks issues have been related to char* vs. string behavior. 

http://katyhuff.github.com

[Anthony Scopatz]

Hi Katy, 

I agree with you that it has something to do with strings.  The error is with respect to the name of the data set that is being opened.  That said, Matt, what happens if in the VLDataset() function you replace all of the lines lines that look like 

name = "String";

with 

name = std::string("String");

I recall Katy did something like this to fix some string bugs on Mavericks previously.

Be Well

Anthony

[Katy Huff]

yep.

[Matthew Gidden]

On Mon, Aug 4, 2014 at 12:32 PM, Anthony Scopatz <sco...@gmail.com> wrote:

Hi Katy, 

I agree with you that it has something to do with strings.  The error is with respect to the name of the data set that is being opened.  That said, Matt, what happens if in the VLDataset() function you replace all of the lines lines that look like 

name = "String";

with 

name = std::string("String");

Same failure. I also tried updating `name = std::string("MapStringString");`.

[Anthony Scopatz]

Hi Matt, 

Can you email us the db.h5?

Be Well

Anthony

[Matthew Gidden]

This was generated by commenting out the file removal lines in the test's TearDown and running

dyn-72-33-19-253:cyclus gidden$ ./install.py -j 4 && cyclus_unit_tests --gtest_filter=Hdf5BackTests.ReadWriteVLMapVLStringDouble

[Anthony Scopatz]

Ok try this, 

On tests/hdf5_back_tests.cc:693, change the line 

TestBasic<map<string, double> >("vl_map_vl_string_double", x, y);

to 

TestBasic<map<string, double> >(string("vl_map_vl_string_double"), x, y);

The reason for this is if you look at h5ls, 

$ h5ls db.h5

MapStringDoubleKeys      Dataset {2/Inf}

MapStringDoubleVals      Dataset {4294967295, 4294967295, 4294967295, 4294967295, 4294967295}

StringKeys               Dataset {3/Inf}

StringVals               Dataset {4294967295, 4294967295, 4294967295, 4294967295, 4294967295}

\212\001                 Dataset {0/Inf}

There is definitely a messed up dataset name and the only one it could be is /vl_map_vl_string_double.

Be Well

Anthony

[Matthew Gidden]

Same error using the following diff:

dyn-72-33-19-253:cyclus gidden$ git diff
diff --git a/tests/hdf5_back_tests.cc b/tests/hdf5_back_tests.cc
index c9385d8..3f3e92f 100644
--- a/tests/hdf5_back_tests.cc
+++ b/tests/hdf5_back_tests.cc
@@ -678,7 +678,7 @@ TEST_F(Hdf5BackTests, ReadWriteMapVLStringDouble) {
   y["wondering"] = 42.0;
   y["why"] = 43.0;
   y["I'm here!"] = 44.0;
-  TestBasic<map<string, double> >("map_vl_string_double", x, y);
+  TestBasic<map<string, double> >(std::string("map_vl_string_double"), x, y);
 }
 

[Anthony Scopatz]

Hi Matt, 

You are on the wrong line.  You need to be on line 693 in the ReadWriteVLMapVLStringDouble test case, not on line 678 in the ReadWriteMapVLStringDouble case.

Be Well

Anthony

[Matthew Gidden]

Realized that just after I sent it. Still have the error. Relevant output is below:

diff --git a/tests/hdf5_back_tests.cc b/tests/hdf5_back_tests.cc
index c9385d8..264ffa9 100644
--- a/tests/hdf5_back_tests.cc
+++ b/tests/hdf5_back_tests.cc
 
 TEST_F(Hdf5BackTests, ReadWriteVLMapVLStringDouble) {
@@ -686,11 +686,11 @@ TEST_F(Hdf5BackTests, ReadWriteVLMapVLStringDouble) {
   using std::string;
   shape.clear();
   map<string, double> x;
-  x["I am lonesome"] = 6.0;
-  x["Cowboy Burt"] = 28.0;
+  x[std::string("I am lonesome")] = 6.0;
+  x[std::string("Cowboy Burt")] = 28.0;
   map<string, double> y;
-  y["Come smell my friend's shirt"] = 42.0;
-  TestBasic<map<string, double> >("vl_map_vl_string_double", x, y);
+  y[std::string("Come smell my friend's shirt")] = 42.0;
+  TestBasic<map<string, double> >(std::string("vl_map_vl_string_double"), x, y);
 }

and 

dyn-72-33-19-253:cyclus gidden$ cyclus_unit_tests --gtest_filter=Hdf5BackTests.ReadWriteVLMapVLStringDouble
Note: Google Test filter = Hdf5BackTests.ReadWriteVLMapVLStringDouble

[==========] Running 1 test from 1 test case.
[----------] Global test environment set-up.
[----------] 1 test from Hdf5BackTests

[ RUN      ] Hdf5BackTests.ReadWriteVLMapVLStringDouble

HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 358 in H5Dopen2(): not found
    major: Dataset
    minor: Object not found
  #001: H5Gloc.c line 430 in H5G_loc_find(): can't find object
    major: Symbol table
    minor: Object not found
  #002: H5Gtraverse.c line 861 in H5G_traverse(): internal path traversal failed
    major: Symbol table
    minor: Object not found
  #003: H5Gtraverse.c line 641 in H5G_traverse_real(): traversal operator failed
    major: Symbol table
    minor: Callback failed

  #004: H5Gloc.c line 385 in H5G_loc_find_cb(): object '?w???' doesn't exist

  #004: H5Gloc.c line 385 in H5G_loc_find_cb(): object 'vl_map_vl_string_double' doesn't exist

    major: Symbol table
    minor: Object not found
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 548 in H5Dget_type(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 358 in H5Dopen2(): not found
    major: Dataset
    minor: Object not found
  #001: H5Gloc.c line 430 in H5G_loc_find(): can't find object
    major: Symbol table
    minor: Object not found
  #002: H5Gtraverse.c line 861 in H5G_traverse(): internal path traversal failed
    major: Symbol table
    minor: Object not found
  #003: H5Gtraverse.c line 641 in H5G_traverse_real(): traversal operator failed
    major: Symbol table
    minor: Callback failed

  #004: H5Gloc.c line 385 in H5G_loc_find_cb(): object 'vl_map_vl_string_double' doesn't exist

    major: Symbol table
    minor: Object not found
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 1161 in H5Dset_extent(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5D.c line 461 in H5Dget_space(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5Shyper.c line 6593 in H5Sselect_hyperslab(): not a data space
    major: Invalid arguments to routine
    minor: Inappropriate type
HDF5-DIAG: Error detected in HDF5 (1.8.13) thread 0:
  #000: H5Dio.c line 225 in H5Dwrite(): can't prepare for writing data
    major: Dataset
    minor: Write failed
  #001: H5Dio.c line 257 in H5D__pre_write(): not a dataset
    major: Invalid arguments to routine
    minor: Inappropriate type
unknown file: Failure
C++ exception with description "Failed to write to the HDF5 table:
  file      db.h5

  table     vl_map_vl_string_double

  num. rows 2
  rowsize   36
    # Column 0
      dbtype: 7
      size:   16
      offset: 0
    # Column 1

      dbtype: 119

      size:   20
      offset: 16
" thrown in the test body.

[  FAILED  ] Hdf5BackTests.ReadWriteVLMapVLStringDouble (26 ms)
[----------] 1 test from Hdf5BackTests (26 ms total)

[----------] Global test environment tear-down

[==========] 1 test from 1 test case ran. (36 ms total)

[  PASSED  ] 0 tests.
[  FAILED  ] 1 test, listed below:

[  FAILED  ] Hdf5BackTests.ReadWriteVLMapVLStringDouble 

[Anthony Scopatz]

This seems like a new and exciting error, different than the one seen before. Though it does seem similar to another error.

[Matthew Gidden]

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

However, the original intent of this work was to investigate the current cycamore regression test failures on mac. The only other reasonable recourse is to try to use insam; however, it looks like it is no longer up (I get a name or service not known error with ssh)? Steve, is this service still available?

Regards,

[Matthew Gidden]

On Tue, Aug 5, 2014 at 1:33 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

I should note that I have cornered at least one error to table creation, as we expected from visual confirmation of the malformed db. 

[Matthew Gidden]

Tangential to this thread, I believe I found the nightly error via a cyclopts unit test. A PR is in with a unit test for cyclus. 

[Anthony Scopatz]

On Tue, Aug 5, 2014 at 12:51 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Tangential to this thread, I believe I found the nightly error via a cyclopts unit test. A PR is in with a unit test for cyclus. 

Hi Matt, 

Here is hoping!

 

On Tue, Aug 5, 2014 at 1:34 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 1:33 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

What matters most is that the nightly tests run successfully. If this helped you find the error, then that is good.  We don't yet purport to fully support Mavericks yet, in part because we do not have CI available for this platform. 

Katy, are you seeing the same unit test failures on mac that Matt is?

 

I should note that I have cornered at least one error to table creation, as we expected from visual confirmation of the malformed db. 

Matt, could you put in a PR for the fix (if you have one) so that your effort here isn't a total wash.  Thanks!

 

However, the original intent of this work was to investigate the current cycamore regression test failures on mac. The only other reasonable recourse is to try to use insam; however, it looks like it is no longer up (I get a name or service not known error with ssh)? Steve, is this service still available?

Steve, if Insam is still available that would be great!

Be Well

Anthony

[Katy Huff]

On Tue, Aug 5, 2014 at 1:10 PM, Anthony Scopatz <sco...@gmail.com> wrote:

On Tue, Aug 5, 2014 at 12:51 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Tangential to this thread, I believe I found the nightly error via a cyclopts unit test. A PR is in with a unit test for cyclus. 

Hi Matt, 

Here is hoping!

 

On Tue, Aug 5, 2014 at 1:34 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 1:33 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

What matters most is that the nightly tests run successfully. If this helped you find the error, then that is good.  We don't yet purport to fully support Mavericks yet, in part because we do not have CI available for this platform. 

Katy, are you seeing the same unit test failures on mac that Matt is?

Yep. These are known. Please see the issue matt referenced. I mentioned them and you commented here:  https://github.com/cyclus/cyclus/issues/943#issuecomment-45687221

--
http://katyhuff.github.com

[Matthew Gidden]

On Tue, Aug 5, 2014 at 3:10 PM, Anthony Scopatz <sco...@gmail.com> wrote:

On Tue, Aug 5, 2014 at 12:51 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Tangential to this thread, I believe I found the nightly error via a cyclopts unit test. A PR is in with a unit test for cyclus. 

Hi Matt, 

Here is hoping!

 

On Tue, Aug 5, 2014 at 1:34 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 1:33 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

What matters most is that the nightly tests run successfully. If this helped you find the error, then that is good.  We don't yet purport to fully support Mavericks yet, in part because we do not have CI available for this platform. 

Katy, are you seeing the same unit test failures on mac that Matt is?

 

I should note that I have cornered at least one error to table creation, as we expected from visual confirmation of the malformed db. 

Matt, could you put in a PR for the fix (if you have one) so that your effort here isn't a total wash.  Thanks!

I do not, I was simply debugging; no fix was found. 

[Anthony Scopatz]

On Tue, Aug 5, 2014 at 1:13 PM, Katy Huff <katy...@gmail.com> wrote:

On Tue, Aug 5, 2014 at 1:10 PM, Anthony Scopatz <sco...@gmail.com> wrote:

On Tue, Aug 5, 2014 at 12:51 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Tangential to this thread, I believe I found the nightly error via a cyclopts unit test. A PR is in with a unit test for cyclus. 

Hi Matt, 

Here is hoping!

 

On Tue, Aug 5, 2014 at 1:34 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 1:33 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

What matters most is that the nightly tests run successfully. If this helped you find the error, then that is good.  We don't yet purport to fully support Mavericks yet, in part because we do not have CI available for this platform. 

Katy, are you seeing the same unit test failures on mac that Matt is?

Yep. These are known. Please see the issue matt referenced. I mentioned them and you commented here:  https://github.com/cyclus/cyclus/issues/943#issuecomment-45687221

Ahh Ok, So did #959 fix the segfault and not the unit test failures then?  I am not sure that we should reopen #943 at this point, but perhaps another issue is warranted. 

[Anthony Scopatz]

On Tue, Aug 5, 2014 at 1:19 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 3:10 PM, Anthony Scopatz <sco...@gmail.com> wrote:

On Tue, Aug 5, 2014 at 12:51 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Tangential to this thread, I believe I found the nightly error via a cyclopts unit test. A PR is in with a unit test for cyclus. 

Hi Matt, 

Here is hoping!

 

On Tue, Aug 5, 2014 at 1:34 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 1:33 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

What matters most is that the nightly tests run successfully. If this helped you find the error, then that is good.  We don't yet purport to fully support Mavericks yet, in part because we do not have CI available for this platform. 

Katy, are you seeing the same unit test failures on mac that Matt is?

 

I should note that I have cornered at least one error to table creation, as we expected from visual confirmation of the malformed db. 

Matt, could you put in a PR for the fix (if you have one) so that your effort here isn't a total wash.  Thanks!

I do not, I was simply debugging; no fix was found. 

Could you open an issue describing it then, please?

[Matthew Gidden]

On Tue, Aug 5, 2014 at 3:20 PM, Anthony Scopatz <sco...@gmail.com> wrote:

On Tue, Aug 5, 2014 at 1:19 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 3:10 PM, Anthony Scopatz <sco...@gmail.com> wrote:

On Tue, Aug 5, 2014 at 12:51 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Tangential to this thread, I believe I found the nightly error via a cyclopts unit test. A PR is in with a unit test for cyclus. 

Hi Matt, 

Here is hoping!

 

On Tue, Aug 5, 2014 at 1:34 PM, Matthew Gidden <gid...@wisc.edu> wrote:

On Tue, Aug 5, 2014 at 1:33 PM, Matthew Gidden <gid...@wisc.edu> wrote:

Hi all,

Katy and I chatted today to try and fix there error and were unsuccessful. I've spent a pretty significant amount of time investigating this to no avail. From some historical context, it is likely that the error is related to the specialized any.hpp we use and its integration with the hdf5 backend. At this point I'm stumped and am not convinced  that this is a fruitful use of my time. 

What matters most is that the nightly tests run successfully. If this helped you find the error, then that is good.  We don't yet purport to fully support Mavericks yet, in part because we do not have CI available for this platform. 

Katy, are you seeing the same unit test failures on mac that Matt is?

 

I should note that I have cornered at least one error to table creation, as we expected from visual confirmation of the malformed db. 

Matt, could you put in a PR for the fix (if you have one) so that your effort here isn't a total wash.  Thanks!

I do not, I was simply debugging; no fix was found. 

Could you open an issue describing it then, please?

done - https://github.com/cyclus/cyclus/issues/981 

[Skutnik, Steven Eugene]

Hi Matt,

 

I will check into the status of insam for you and see what we can work out in terms of either getting it back or securing a replacement for OSX CI.

 

-Steve

[Skutnik, Steven Eugene]

Hi Matt,

Turns out insam got turned off over a series of power outages (due to renovations next door) and it simply never got turned back on. Insam should be back up and available for use now.

-Steve

—

Steve Skutnik, Ph.D.

Assistant Professor of Nuclear Engineering
505-A Ferris Hall, University of Tennessee

(865)974-2212

OK. Hmm. All of the ones that fail come from nested VLMaps and VLStrings.  The error comes fromthis line, which doesn't make a lot of sense in this context.  This indicates that either the VLMap dataset can't be opened or the VLString one cannot be.  Without diving into gdb or putting in some print statements it is hard to know which.

 

--

http://katyhuff.github.com

[Anthony Scopatz]

On Thu, Aug 7, 2014 at 9:12 AM, Skutnik, Steven Eugene <ssku...@utk.edu> wrote:

Hi Matt,

Turns out insam got turned off over a series of power outages (due to renovations next door) and it simply never got turned back on. Insam should be back up and available for use now.

Steve, 

That is great news!  Thanks.

Be Well

Anthony