* jerry gay <[EMAIL PROTECTED]> [2006-09-30 01:30]: > On 9/29/06, Ovid <[EMAIL PROTECTED]> wrote: > > got: this is line 1 > > this is line 2 > > this is line 3 > > expected: this is line 1 > > this is line b > > this is line 3 > > i prefer this syntax, as it's easier to parse in languages > without builtin heredocs. however, it bothers me that with this > syntax it's difficult to tell if and how many leading spaces > the first line of got or expected output contains.
You quoted too selectively. There are more lines of diagnostic info before the got/expected stuff. You can pick up the width of the indent from there. I like this format just fine, exactly as proposed. Heredocs seem full of downsides to me. If the `END` token is a fixed string, then this raises the question of what happenes when the data contains that sequence. If it’s instead variable and is generated in such a way as to not collide with data, then it gets more complex to write both TAP emitters and parsers. All of that for a format that is harder to read for humans anyway. Regards, -- Aristotle Pagaltzis // <http://plasmasturm.org/>
