2019-06-06 15:59:33 +0000, Austin Group Bug Tracker: [...] > Change:<blockquote>The argument <i>text</i> shall consist of one or more > lines. Each embedded <newline> in the text shall be preceded by a > <backslash>. Other <backslash> characters in text shall be removed, and the > following character shall be treated > literally.</blockquote>to:<blockquote>The argument <i>text</i> shall > consist of one or more lines. A <backslash> in the text can be escaped with > another <backslash>. The application shall ensure that each embedded > <newline> (that is, those other than the terminating <newline> of the last > line) in the text is preceded by an unescaped <backslash>. The behaviour is > unspecified if an unescaped <backslash> is immediately followed by any > character other than <backslash> or <newline>, or by the end of a > <i>script</i>.</blockquote> [...]
Thanks for looking into that. Note that I had used "script_file" instead of "script" in my proposed resolution to allow for the historical implementation behaviour. IIRC, the sed script meant to be made of the concatenation of all the script expressions and script_files. So with a sed -e expr1 -f file1 -e expr2 -f file2 sed is meant to interpret the: expr1 content-of-file1 expr2 content-of-file2 "script". However, that doesn't work in historical implementations where: sed -e 'a\' -e 'text' doesn't work the same way as: sed 'a\ text' Is that to say that the approved resolution considers the historical behaviour as non-conforming (I'm fine with that, I would also consider it as a bug)? If so, is the conformance suite being updated to check for that? -- Stephane
