Patch #25861
closed
CSV Importer - handle UndefinedConversionErrors
0%
Description
The CSV import already handles a couple of Exceptions that may happen when the user-selected encoding does not match the real encoding of the CSV (or it contains otherwise illegal byte sequences). We recently encountered Encoding::UndefinedConversionError, which is not yet caught by that error handling. The attached patches add a test case and a fix for that.
Files
Updated by Go MAEDA over 7 years ago
- Target version set to 3.2.7
We can reproduce the problem with the test in 0001-adds-failing-test-case-for-import-with-wrong-encodin.patch. Setting target version to 3.2.7.
By the way, the encoding of the file invalid-Shift_JIS.csv is valid CP932. Shift_JIS and CP932 are almost identical, but CP932 includes more characters.
$ iconv -f cp932 -t utf8 test/fixtures/files/invalid-Shift_JIS.csv ①
Updated by Jean-Philippe Lang over 7 years ago
- Status changed from New to Needs feedback
Go MAEDA wrote:
We can reproduce the problem with the test in 0001-adds-failing-test-case-for-import-with-wrong-encodin.patch. Setting target version to 3.2.7.
The test passes for me with ruby 2.0 (I would get a Encoding::InvalidByteSequenceError). Maybe we could simply rescue EncodingError which is the base class for encoding errors?
Updated by Go MAEDA over 7 years ago
Jean-Philippe Lang wrote:
Maybe we could simply rescue
EncodingErrorwhich is the base class for encoding errors?
I agree, caching EncodingError is better. We can simplify the code and catch all errors caused by iconv.
Updated by Jean-Philippe Lang over 7 years ago
- Status changed from Needs feedback to Resolved
- Assignee set to Jean-Philippe Lang
Updated by Jean-Philippe Lang over 7 years ago
- Status changed from Resolved to Closed