Unicode string weirdness post-1.0.13 #6
Comments
|
Wasn't ably to take look today. I will take a look by Monday. Is the fix urgent? |
|
No, not urgent, I'm just pinning to 1.0.13 for now. |
|
I found a test case for you to try when you have time: key = ('test','test','test')
client.remove(key)
client.put(key, {'test': [unicode('a'), unicode('b')]})
print(client.get(key)[2]['test']) #you should see [u'a', u'b']
client.remove(key)
client.put(key, {'test': [unicode('aa'), unicode('bb')]})
print(client.get(key)[2]['test']) #on my machine, I get [u'bb', u'bb'] |
|
Sorry, didn't mean to bump this. I did just pull down 1.0.19 and confirm that the problem still exists. |
|
Sorry, I have not gotten to this one yet. I will check it out today... but may not have a fix until tomorrow. |
|
I just tested the code. Looks like the the data is clobbered on the put() call, because it is stored that way in the database. Taking a deeper look. |
|
Released v1.0.20, with a fix. It is up on GitHub and PyPi Commit: |
dwelch-spike
added a commit
that referenced
this issue
Oct 24, 2019
* fixed OverflowError: Python int too large to convert to C long * added large job id test * fixed double linked list corruption error * added query apply set detection tests * updated test_query_apply * refactored double free and python overflow fixes * added test for non list udf arguments * update c client * check for illegal udf arguments in dictionaries * added bytes udf test * added py_to_val coversion for python3 bytes * update illegal udf args exception, add stream udf tuple test * iterative udf arg check * removed unnecessary assignments, consistent string encoding in tests, added udf arg documentation to query.apply * fail on 4XX response codes https://curl.haxx.se/mail/archive-2006-02/0004.html the exit status of curl on a 4XX response code is zero without the -f flag. * Remove transaction queue and thread-per-queue config options These config options were removed in server version 4.7.0 https://www.aerospike.com/download/server/notes.html#4.7.0.2 * Bitwise operations (#6) * Initial bitwise file * Start of bitwise ops * Tests for bitwise beginning * Basic bitwise testing * Tests for resize * bitwise remove and set * set_bit tests, documentation, set_bit Byte_Offset fixed * implemented bit_count * bit_count tests * implemented bit_add and bit_add tests * implemented bit_and, bit_and tests * implemented bit_get * bit_get tests * implemented bit_get_int and bit_get_int tests * implemented bit_insert * bit_insert tests * implemented bit_lscan and bit_lscan tests * implemented bit_lshift and lshift tests * implemented bit_not and bit_not tests * implemented bit_or and bit_or tests * implemented bit_rscan and bit_rshift with tests * implemented bit_subtract and tests * implemented bit_xor and tests * bit_wise documentation and tests * update c client, fix travis config thread queue option, cleanup bit operate * cleanup bitwise documentation, change isbitop to use range * add bitwise helper get_uint32t_from_pyargs, refactor bitwise code and documentation * Nested_cdt support for map and list ops. (#7) * nested cdt groundwork and tests * nested cdt constants * added support for as_val based nested cdt ops, list_op nested cdt ops, and tests * nested cdt for list ops * nested cdt_ctx logic for all list ops, more tests for nested cdt list ops * nested cdt ctx groundwork, nested cdt for map_get_by_key * Remove transaction queue and thread-per-queue config options These config options were removed in server version 4.7.0 https://www.aerospike.com/download/server/notes.html#4.7.0.2 * nested map cdt get by val and index, fixed nested cdt list tests and python cdt_ctx converter * nested cdt map ops for all remove by rank, map size, map policy, increment, put, put items and tests * nested cdt ops and tests for all map ops except relative ops * nested cdt ops and tests for maps * fix ctx error mem leak * move ctx_operations out of operations * sort nested_map by key_order
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Thanks for the awesome work! I'm running into a problem when storing unicode values (the bins are strings). Values from one bin might show up in other bins. I'll try and figure out a dataset that consistently replicates the bug, but I CAN tell you it works great in .12 and .13, but starts breaking in .14
The first time I saw it, I had a list of unique values, which when stored became a list of duplicates, hope that helps.
The text was updated successfully, but these errors were encountered: