bug-autoconf
[Top][All Lists]
Advanced

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

Re: tracing output [was: 2.62 AT_SETUP limitations]


From: Eric Blake
Subject: Re: tracing output [was: 2.62 AT_SETUP limitations]
Date: Thu, 24 Apr 2008 23:18:52 +0000 (UTC)
User-agent: Loom/3.14 (http://gmane.org/)

Eric Blake <ebb9 <at> byu.net> writes:

> 
> Perhaps we could add _m4_bpatsubst as a synonym for m4_bpatsubst, and use
> _m4_bpatsubst anywhere that tracing is dangerous (_m4_split) and
> m4_bpatsubst everywhere else.  User code should not be calling
> _m4_bpatsubst, as an undocumented internal macro; the end result would be
> that tracing m4_bpatsubst is less likely to fail, because it no longer
> traces during changed quotes.

I'm thinking about patching along these lines.  Before, I get:

$ autoconf --trace m4_bpatsubst --debug
/usr/local/bin/m4:/wintmp/am4t1mDdqr/traces.m4:669: AT_m4_bpatsubst: end of 
file in string
...
$ echo $?
0

Ouch - the trace failed, but the exit status doesn't reflect that.  Someone 
will need to investigate that.

After the patch, I get a valid trace.  But as is, this patch then breaks 
tracing of m4_builtin, which worked before, so it just moves the problem 
around.  Also, it slows down the macro m4_split, which is frequently called, in 
order to protect the m4_bpatsubst traces.  So, any opinions on whether this is 
worth pursuing further?

diff --git a/lib/m4sugar/m4sugar.m4 b/lib/m4sugar/m4sugar.m4
index 4f86546..f19d2a0 100644
--- a/lib/m4sugar/m4sugar.m4
+++ b/lib/m4sugar/m4sugar.m4
@@ -1745,6 +1745,13 @@ m4_define([m4_toupper],
 # that the list contains single space separators, and a common case is
 # splitting a single-element list.  This macro is called frequently,
 # so avoid unnecessary dnl inside the definition.
+#
+# We want to support autom4te --trace m4_bpatsubst, so that a user can
+# track all regular expressions used in their input.  However, _m4_split
+# invokes the patsubst builtin with different quoting rules in effect, and
+# the autom4te post-processing falls over when it fails to find matching [].
+# So we use m4_builtin for the real work, which will not be traced,
+# while providing a no-op stub m4_bpatsubst for tracing purposes.
 m4_define([m4_split],
 [m4_if([$1], [], [],
        [$2], [ ], [m4_if(m4_index([$1], [ ]), [-1], [[[$1]]], [_$0($@)])],
@@ -1752,9 +1759,11 @@ m4_define([m4_split],
        [_$0($@)])])
 
 m4_define([_m4_split],
+[m4_pushdef([m4_bpatsubst])m4_bpatsubst([$1], [$2], [@:>@, @<:@])]dnl
 [m4_changequote([-=<{(],[)}>=-])]dnl
-[[m4_bpatsubst(-=<{(-=<{($1)}>=-)}>=-, -=<{($2)}>=-,
-              -=<{(], [)}>=-)]m4_changequote([, ])])
+[[m4_builtin(-=<{(patsubst)}>=-, -=<{(-=<{($1)}>=-)}>=-, -=<{($2)}>=-,
+            -=<{(], [)}>=-)]m4_changequote([, ])]dnl
+[m4_builtin([popdef], [m4_bpatsubst])])
 
 
-- 
Eric Blake







reply via email to

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