Front page | perl.perl5.porters |
Postings from July 2017
Kent Fredric wrote:
>I've seen *reams* of test suites that have this problem simply because
>of the fun with needing to deal with marshalling your strings through
>Test::More.
Do you mean things like the got/expected output from a failed is() test?
It periodically annoys me that this outputs control characters found
in the data. In similar vein, that it tries to output non-Latin-1
characters direct to a stream that it doesn't know can handle it seems
like a bug. The fact that, as a result, the encoding of the diagnostics
is inconsistent within a single test script run certainly is a bug.
This is a bug that has been allowed to arise and persist precisely
because of this broken core behaviour.
The testing infrastructure should really be using something like
Data::Dumper to represent unprintable and non-ASCII strings. Even in
a fully Unicode-capable environment, sticking to ASCII has value for
testing output. It has value to the immediate user in making "\x{391}"
distinct from "\x{41}" and "\x{e9}" distinct from "\x{65}\x{301}".
And it has value in making the test output robust for communicating it
via the myriad mechanisms we have that can't be relied on to get the
encoding right at every stage.
-zefram
Thread Previous
|
Thread Next