bug-texinfo
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: "[7031] filter_files zero return status" commit not working on OpenB


From: Gavin Smith
Subject: Re: "[7031] filter_files zero return status" commit not working on OpenBSD
Date: Fri, 20 May 2016 16:09:19 +0100

On 19 May 2016 at 05:46, Ingo Feinerer <address@hidden> wrote:
> I am using the ksh included with OpenBSD (which reports a very old
> version number but is maintained/developed in OpenBSD):
>
> $ sysctl kern.version
> kern.version=OpenBSD 6.0-beta (GENERIC.MP) #2058: Mon May 16 22:38:41 MDT 2016
>     address@hidden:/usr/src/sys/arch/amd64/compile/GENERIC.MP
>
> $ echo $KSH_VERSION
> @(#)PD KSH v5.2.14 99/07/13.2
>
>> It could be a problem with "set -e" as was the case before. To confirm
>> this, try deleting this line at the start of texi2dvi and see if the
>> same error happens.
>
> Good catch!
> Commenting out "set -e" and it works :-)
>
>> The challenge then would be to find where exactly
>> in "decide_aux_files_method" the failure is detected.
>
> Is there a specific way to provide further debug output that helps
> investigating this issue?

Maybe try reading the ksh man page ("info ksh" or "man ksh") to see if
there are options that can give more tracing. If I were you I'd
investigate with echo statements, e.g.

--- texi2dvi    (revision 7149)
+++ texi2dvi    (working copy)
@@ -1,4 +1,4 @@
-#! /bin/sh
+#! /bin/ksh
 # texi2dvi --- produce DVI (or PDF) files from Texinfo (or (La)TeX) sources.
 # $Id$
 #
@@ -1566,7 +1566,9 @@
   #
   #  The default behaviour is `nomaybe'.

+  echo got here 1
   test -n "$TEXI2DVI_USE_RECORDER" || TEXI2DVI_USE_RECORDER=nomaybe
+  echo got here 2

   case $TEXI2DVI_USE_RECORDER in
     yes) set_aux_files_from_fls;;
@@ -1596,6 +1598,7 @@
     *) error 1 "Invalid value of TEXI2DVI_USE_RECORDER environment
variable : $TEXI2DVI_USE_RECORDER.";;

   esac
+  echo got here 3
 }

 # remove FILE...

You can tell how far you get by what's printed out.

Likely there's some failing command that should be exempted from "set
-e", but which isn't.
https://lists.gnu.org/archive/html/bug-texinfo/2016-02/msg00141.html



reply via email to

[Prev in Thread] Current Thread [Next in Thread]