PATCH: access to names of errors

PATCH: access to names of errors

[Peter Stephenson]

This patches the parameter module to add $sys_errnos, which turns $ERRNO
into the name of the error --- not the error text you would get with
strerror(), the standard name.  This is more useful for programming,
since if you know you can test for

[[ $sys_errnos[$ERRNO] = EINTR ]]

and so on.  It's also more difficult, but I simply copied the code for
signal names.

I'd appreciate comments on
- where this should appear; I'm planning on an interface to the system
  read function and it could go there instead.  This might be
  better since the parameter module usually provides shell rather
  than system information and the new array bloats it somewhat.
  The new module could be a generic zsystem module.
- what it should be called; the current name at least paves the way
  for a move into a different namespace such as $sys.errnos, should
  anybody ever have the time to rewrite the parameter code completely
- what else we should provide; strerror() would be fairly easy, but
  an array would probably have to be generated specially during
  configuration, because the internal system array seems to be
  non-standard.  We can check for _sys_errlist and _sys_nerr first,
  perhaps.

There's no documenation until that gets decided.

Index: zshconfig.ac
===================================================================
RCS file: /cvsroot/zsh/zsh/zshconfig.ac,v
retrieving revision 1.37
diff -u -r1.37 zshconfig.ac
--- zshconfig.ac	25 Apr 2003 11:18:51 -0000	1.37
+++ zshconfig.ac	22 Jul 2003 13:08:04 -0000
@@ -1117,6 +1117,39 @@
 SIGNAL_H=$zsh_cv_path_signal_h
 AC_SUBST(SIGNAL_H)dnl
 
+dnl Where are error names located?  Needed as input for errnos.awk
+AC_CACHE_CHECK(where error names are located, zsh_cv_path_errno_h,
+[dnl Look at the output from the preprocessor.
+dnl We should get lines of the form `# 1 "/usr/include/errno.h"'
+dnl The following assumes the real definitions are in a file which
+dnl contains the name `err'; we could relax this if necessary,
+dnl but then you can get a rather long list of files to test.
+dnl The backslash substitution is to persuade cygwin to cough up
+dnl slashes rather than doubled backslashes in the path.
+echo "#include <errno.h>" > nametmp.c
+errfile_list="`$CPP nametmp.c |
+sed -n 's/^#[ 	].*\"\(.*\)\"/\1/p' |
+sed 's/\\\\\\\\/\//g' |
+$AWK '{ if (\$1 ~ \"err\") files[[\$1]] = \$1 }
+  END { for (var in files) print var }'`"
+rm -f nametmp.c
+for ERRNO_H in $errfile_list /dev/null
+do
+  dnl Try to make sure it doesn't get confused by files that don't
+  dnl have real error definitions in.  Count definitions to make sure.
+  nerrs=`test -f $ERRNO_H && \
+  grep '#[ 	]*define[ 	][ 	]*E[0-9A-Z]*[ 	]*[0-9][0-9]*' $ERRNO_H | \
+  wc -l | sed 's/[ 	]//g'`
+  test "x$nerrs" != x && test "$nerrs" -ge 7 && break
+done
+if test $ERRNO_H = "/dev/null"; then
+  AC_MSG_ERROR(ERROR MACROS NOT FOUND:  please report to developers)
+fi
+zsh_cv_path_errno_h=$ERRNO_H
+])
+ERRNO_H=$zsh_cv_path_errno_h
+AC_SUBST(ERRNO_H)dnl
+
 dnl -----------------------------------------------------
 dnl Look for the file containing the RLIMIT_* definitions
 dnl -----------------------------------------------------
Index: Src/Modules/errnames1.awk
===================================================================
RCS file: Src/Modules/errnames1.awk
diff -N Src/Modules/errnames1.awk
--- /dev/null	1 Jan 1970 00:00:00 -0000
+++ Src/Modules/errnames1.awk	22 Jul 2003 13:08:06 -0000
@@ -0,0 +1,18 @@
+# Edited version of Src/signames1.awk.
+#
+# This is an awk script which finds out what the possibilities for
+# the error names are, and dumps them out so that cpp can turn them
+# into numbers.  Since we don't need to decide here what the
+# real signals are, we can afford to be generous about definitions,
+# in case the definitions are in terms of other definitions.
+# However, we need to avoid definitions with parentheses, which will
+# mess up the syntax.
+BEGIN { printf "#include <errno.h>\n\n" }
+
+/^[\t ]*#[\t ]*define[\t ]*E[A-Z0-9]*[\t ][\t ]*[^(\t ]/ { 
+    eindex = index($0, "E")
+    etail = substr($0, eindex, 80)
+    split(etail, tmp)
+    enam = substr(tmp[1], 2, 20)
+    printf("XXNAMES XXE%s E%s\n", enam, enam)
+}
Index: Src/Modules/errnames2.awk
===================================================================
RCS file: Src/Modules/errnames2.awk
diff -N Src/Modules/errnames2.awk
--- /dev/null	1 Jan 1970 00:00:00 -0000
+++ Src/Modules/errnames2.awk	22 Jul 2003 13:08:06 -0000
@@ -0,0 +1,42 @@
+# Edited version of Src/signames2.awk.
+#
+# {g,n}awk script to generate errnames.c
+# This version relies on the previous output of the preprocessor
+# on sigtmp.c, sigtmp.out, which is in turn generated by errnames1.awk.
+#
+# NB: On SunOS 4.1.3 - user-functions don't work properly, also \" problems
+# Without 0 + hacks some nawks compare numbers as strings
+#
+/^XXNAMES XXE[A-Z0-9]*[\t ][\t ]*[1-9][0-9]*/ {
+    eindex = index($0, "E")
+    etail = substr($0, 11, 80)
+    split(etail, tmp)
+    enam = tmp[1]
+    enum = tmp[2]
+    if (errname[enum] == "") {
+	errname[enum] = enam
+	if (0 + max < 0 + enum && enum < 1024)
+	    max = enum
+    }
+}
+
+END {
+    ps = "%s"
+    printf "/** errnames.c                                 **/\n"
+    printf "/** architecture-customized errnames.c for zsh **/\n"
+    printf "\n"
+    printf "#define ERRCOUNT\t%d\n", max
+    printf "\n"
+    printf "#include %cparameter.mdh%c\n", 34, 34
+    printf "\n"
+    printf "/**/\n"
+    printf "char *sys_errnames[ERRCOUNT+1] = {\n"
+
+    for (i = 1; i <= 0 + max; i++)
+	if (errname[i] == "")
+	    printf("\t%c%d%c,\n", 34, i, 34)
+	else
+	    printf("\t%c%s%c,\n", 34, errname[i], 34)
+    print "\tNULL"
+    print "};"
+}
Index: Src/Modules/parameter.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Modules/parameter.c,v
retrieving revision 1.23
diff -u -r1.23 parameter.c
--- Src/Modules/parameter.c	7 Mar 2003 12:31:12 -0000	1.23
+++ Src/Modules/parameter.c	22 Jul 2003 13:08:07 -0000
@@ -1852,6 +1854,16 @@
     scanaliases(ht, func, flags, 1, DISABLED);
 }
 
+/* Functions for the sys_errnosgetfn special parameter. */
+
+/**/
+static char **
+sys_errnosgetfn(Param pm)
+{
+    /* arrdup etc. should really take const pointers as arguments */
+    return arrdup((char **)sys_errnames);
+}
+
 /* Table for defined parameters. */
 
 struct pardef {
@@ -1936,6 +1948,9 @@
     { "dis_galiases", 0,
       getpmdisgalias, scanpmdisgaliases, setpmdisgaliases,
       NULL, NULL, stdunsetfn, NULL },
+    { "sys_errnos", PM_ARRAY|PM_SPECIAL|PM_READONLY,
+      NULL, NULL, NULL,
+      arrsetfn, sys_errnosgetfn, stdunsetfn, NULL },
     { NULL, 0, NULL, NULL, NULL, NULL, NULL, NULL, NULL }
 };
 
Index: Src/Modules/parameter.mdd
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Modules/parameter.mdd,v
retrieving revision 1.2
diff -u -r1.2 parameter.mdd
--- Src/Modules/parameter.mdd	26 Nov 2000 20:01:03 -0000	1.2
+++ Src/Modules/parameter.mdd	22 Jul 2003 13:08:07 -0000
@@ -4,4 +4,21 @@
 
 autoparams="parameters commands functions dis_functions funcstack builtins dis_builtins reswords dis_reswords options modules dirstack history historywords jobtexts jobdirs jobstates nameddirs userdirs aliases dis_aliases galiases dis_galiases"
 
-objects="parameter.o"
+objects="parameter.o errnames.o"
+
+headers="errcount.h"
+
+:<<\Make
+errnames.c: errnames1.awk errnames2.awk $(dir_top)/config.h @ERRNO_H@
+	   if [ x@ERRNO_H@ = x ]; then \
+		touch errtmp.out; \
+	   else \
+		$(AWK) -f $(sdir)/errnames1.awk @ERRNO_H@ >errtmp.c; \
+		$(CPP) errtmp.c >errtmp.out; \
+	   fi
+	   $(AWK) -f $(sdir)/errnames2.awk errtmp.out > $@
+	   rm -f errtmp.c errtmp.out
+
+errcount.h: errnames.c
+	grep 'define.*ERRCOUNT' errnames.c > $@
+Make

-- 
Peter Stephenson <pws@csr.com>                  Software Engineer
CSR Ltd., Science Park, Milton Road,
Cambridge, CB4 0WH, UK                          Tel: +44 (0)1223 692070


**********************************************************************
The information transmitted is intended only for the person or
entity to which it is addressed and may contain confidential 
and/or privileged material. 
Any review, retransmission, dissemination or other use of, or
taking of any action in reliance upon, this information by 
persons or entities other than the intended recipient is 
prohibited.  
If you received this in error, please contact the sender and 
delete the material from any computer.
**********************************************************************

Re: PATCH: access to names of errors

[Oliver Kiddle]

Peter wrote:
> This patches the parameter module to add $sys_errnos, which turns $ERRNO
> into the name of the error --- not the error text you would get with
> strerror(), the standard name.  This is more useful for programming,
> since if you know you can test for
> 
> [[ $sys_errnos[$ERRNO] = EINTR ]]
> 
> and so on.  It's also more difficult, but I simply copied the code for
> signal names.
> 
> I'd appreciate comments on
> - where this should appear; I'm planning on an interface to the system
>   read function and it could go there instead.  This might be
>   better since the parameter module usually provides shell rather
>   than system information and the new array bloats it somewhat.
>   The new module could be a generic zsystem module.

I'd say it shouldn't be in the parameter module because as you say, that
interfaces to shell rather than system information. I'd just call the
module `system'. The `z' seems a bit redundant especially as it is
effectively zsh/system with the namespace.

> - what it should be called; the current name at least paves the way
>   for a move into a different namespace such as $sys.errnos, should
>   anybody ever have the time to rewrite the parameter code completely

I'd have thought just $errnos would be fine. Especially if it can be in
a system namespace in the long term.

> - what else we should provide; strerror() would be fairly easy, but
>   an array would probably have to be generated specially during
>   configuration, because the internal system array seems to be
>   non-standard.  We can check for _sys_errlist and _sys_nerr first,
>   perhaps.

By during configuration, do you mean by autoconf? When the module is
loaded would be better. If it's in an out of the way module that isn't
loaded by default, I don't think it matters. That long over due parameter
code rewrite will have to allow individual array elements to be generated
on demand.

Oliver

________________________________________________________________________
This email has been scanned for all viruses by the MessageLabs Email
Security System. For more information on a proactive email security
service working around the clock, around the globe, visit
http://www.messagelabs.com
________________________________________________________________________