Re: [Bug-ed] Ed scripts failing in versions after ~1.10

[David Given]

Try this:

$ wget https://raw.githubusercontent.com/davidgiven/ack/default/h/em_table

$ wget https://raw.githubusercontent.com/davidgiven/ack/default/modules/src/em_code/make.em.gen

$ sh make.em.gen em_table ; echo $?

Out of the box, the output (ignoring the actual script output) ends with:

?

2

(--verbose shows this is the warning, as expected.)

But changing the script so Q is the last command just produces:

1

So, no diagnostic, but still a non-zero exit status. This is GNU ed 1.14.2 from Debian.

Adding -l does produce an exit status of 0 --- but will this still cause genuine failures if, for example, there's a syntax error in the script?

On Tue, 22 May 2018 at 17:48 Antonio Diaz Diaz <address@hidden> wrote:

Hello David,

David Given wrote:
> Since about ed 1.11, the ACK fails to build because some of the ed scripts
> won't run any more. Here is an example of one:
>
> #!/bin/sh
> echo "/* this part is generated from $1 at: " `date` "*/"
> ed -s $1<<'EOI'
> 1,/^$/d
> 1,/^$/d

I can't find the problem without a sample of the file processed by this
script. I suggest you to add a --verbose option to the invocation of ed
above in order to get a diagnostic beyond the '?' shown by ed by default:

ed -v -s $1 <<'EOI'

After this change I get this result (using a random text file):
$ ./make.em.gen.new tmp.txt ; echo $?
/* this part is generated from tmp.txt at:  Tue May 22 17:35:13 CEST 2018 */
?
script, line 3: No match
1


> Does anyone know what's happening here, and if there's a way to persuade ed
> to quit with an exit status of 0? (I'd rather not discard ed's exit status
> as that would hide processing errors in the scripts.)

You may try adding -l to the invocation of ed:
'-l'
'--loose-exit-status'
      Don't exit with bad status if a command happens to "fail" (for
      example if a substitution command finds nothing to replace). This
      can be useful when 'ed' is invoked as the editor for crontab.