Re: sed disappearing last line
On Tue, Jul 10, 2018 at 07:34:48AM +, Andreas Kusalananda Kähäri wrote: > On Tue, Jul 10, 2018 at 09:29:55AM +0200, Andreas Kusalananda Kähäri wrote: > > This is really a question about what the "N" command should be doing if > > there is no further input. GNU sed seems to output the current pattern > > space. > > > > POSIX says: > > > > If no next line of input is available, the N command verb shall > > branch to the end of the script and quit without starting a new > > cycle or copying the pattern space to standard output. > > > > > > Note: "without [...] copying the pattern space to standard output" > > > > So this makes me believe that OpenBSD sed is doing the correct thing > > with regards to the POSIX spec. for sed. Indeed, this matches the OpenBSD behaviour. Thank you. -- db
Re: sed disappearing last line
On Tue, Jul 10, 2018 at 09:29:55AM +0200, Andreas Kusalananda Kähäri wrote: > On Mon, Jul 09, 2018 at 06:23:17PM -0300, Daniel Bolgheroni wrote: > > Working with some input/commands which dealt with multiline pattern > > spaces, noticed a behaviour I was not expecting. Reducing the test to a > > simple case: > > > > $ echo "foo\nbar\nbaz" | sed "1d; N" > > bar > > baz > > $ > > > > This is for 3 lines of input. However, same command but for 4 lines of > > input: > > > > $ echo "foo\nbar\nbaz\nqux" | sed "1d; N" > > bar > > baz <--- qux not here > > $ > > > > For 6 lines of input: > > > > $ echo "foo\nbar\nbaz\nqux\nquux\nquuz" | sed "1d; N" > > bar > > baz > > qux > > quux <--- quuz not here > > $ > > > > (use echo -e if on bash) > > > > When the input has an even number of lines, the last line disappears. My > > understanding is this isn't supposed to happen, but I'm unsure if it has > > something to do with POSIX gray areas of undefined behaviours. For > > instance, GNU sed from ports presents a different behaviour (second test > > case but with gsed): > > > > $ echo "foo\nbar\nbaz\nqux" | gsed "1d; N" > > bar > > baz > > qux <--- qux here > > $ > > > > Before suggesting anything, anyone with a stronger experience on sed > > historical behaviours can shed some light on if this is acceptable? > > > > Thank you. > > > > -- > > db > > > > This is really a question about what the "N" command should be doing if > there is no further input. GNU sed seems to output the current pattern > space. > > POSIX says: > > If no next line of input is available, the N command verb shall > branch to the end of the script and quit without starting a new > cycle or copying the pattern space to standard output. > > > Note: "without [...] copying the pattern space to standard output" > > So this makes me believe that OpenBSD sed is doing the correct thing > with regards to the POSIX spec. for sed. > > Also note that the input to sed should be "text files". In all your > examples, the last line is not properly terminated, which, strictly and > pedantically speaking, violates the "text file" requirement. This does > not affect the outcome of the commands though. If you're using echo, that adds the terminating newline. I forgot about that, sorry, early morning here... I generally use printf for this sort of data though... > > Also, printf may be used to portably print strings with C escape > sequences. > > > Cheers, -- Andreas Kusalananda Kähäri, National Bioinformatics Infrastructure Sweden (NBIS), Uppsala University, Sweden. När du har kontakt med oss på Uppsala universitet med e-post så innebär det att vi behandlar dina personuppgifter. För att läsa mer om hur vi gör det kan du läsa här: http://www.uu.se/om-uu/dataskydd-personuppgifter/ E-mailing Uppsala University means that we will process your personal data. For more information on how this is performed, please read here: http://www.uu.se/om-uu/dataskydd-personuppgifter/
Re: sed disappearing last line
On Mon, Jul 09, 2018 at 06:23:17PM -0300, Daniel Bolgheroni wrote: > Working with some input/commands which dealt with multiline pattern > spaces, noticed a behaviour I was not expecting. Reducing the test to a > simple case: > > $ echo "foo\nbar\nbaz" | sed "1d; N" > bar > baz > $ > > This is for 3 lines of input. However, same command but for 4 lines of > input: > > $ echo "foo\nbar\nbaz\nqux" | sed "1d; N" > bar > baz <--- qux not here > $ > > For 6 lines of input: > > $ echo "foo\nbar\nbaz\nqux\nquux\nquuz" | sed "1d; N" > bar > baz > qux > quux <--- quuz not here > $ > > (use echo -e if on bash) > > When the input has an even number of lines, the last line disappears. My > understanding is this isn't supposed to happen, but I'm unsure if it has > something to do with POSIX gray areas of undefined behaviours. For > instance, GNU sed from ports presents a different behaviour (second test > case but with gsed): > > $ echo "foo\nbar\nbaz\nqux" | gsed "1d; N" > bar > baz > qux <--- qux here > $ > > Before suggesting anything, anyone with a stronger experience on sed > historical behaviours can shed some light on if this is acceptable? > > Thank you. > > -- > db > This is really a question about what the "N" command should be doing if there is no further input. GNU sed seems to output the current pattern space. POSIX says: If no next line of input is available, the N command verb shall branch to the end of the script and quit without starting a new cycle or copying the pattern space to standard output. Note: "without [...] copying the pattern space to standard output" So this makes me believe that OpenBSD sed is doing the correct thing with regards to the POSIX spec. for sed. Also note that the input to sed should be "text files". In all your examples, the last line is not properly terminated, which, strictly and pedantically speaking, violates the "text file" requirement. This does not affect the outcome of the commands though. Also, printf may be used to portably print strings with C escape sequences. Cheers, -- Andreas Kusalananda Kähäri, National Bioinformatics Infrastructure Sweden (NBIS), Uppsala University, Sweden. När du har kontakt med oss på Uppsala universitet med e-post så innebär det att vi behandlar dina personuppgifter. För att läsa mer om hur vi gör det kan du läsa här: http://www.uu.se/om-uu/dataskydd-personuppgifter/ E-mailing Uppsala University means that we will process your personal data. For more information on how this is performed, please read here: http://www.uu.se/om-uu/dataskydd-personuppgifter/
sed disappearing last line
Working with some input/commands which dealt with multiline pattern spaces, noticed a behaviour I was not expecting. Reducing the test to a simple case: $ echo "foo\nbar\nbaz" | sed "1d; N" bar baz $ This is for 3 lines of input. However, same command but for 4 lines of input: $ echo "foo\nbar\nbaz\nqux" | sed "1d; N" bar baz <--- qux not here $ For 6 lines of input: $ echo "foo\nbar\nbaz\nqux\nquux\nquuz" | sed "1d; N" bar baz qux quux <--- quuz not here $ (use echo -e if on bash) When the input has an even number of lines, the last line disappears. My understanding is this isn't supposed to happen, but I'm unsure if it has something to do with POSIX gray areas of undefined behaviours. For instance, GNU sed from ports presents a different behaviour (second test case but with gsed): $ echo "foo\nbar\nbaz\nqux" | gsed "1d; N" bar baz qux <--- qux here $ Before suggesting anything, anyone with a stronger experience on sed historical behaviours can shed some light on if this is acceptable? Thank you. -- db