From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <zsh-workers-return-34980-mason-zsh=primenet.com.au@zsh.org>
Received: (qmail 8291 invoked by alias); 28 Apr 2015 00:46:05 -0000
Mailing-List: contact zsh-workers-help@zsh.org; run by ezmlm
Precedence: bulk
X-No-Archive: yes
List-Id: Zsh Workers List <zsh-workers.zsh.org>
List-Post: <mailto:zsh-workers@zsh.org>
List-Help: <mailto:zsh-workers-help@zsh.org>
X-Seq: 34980
Received: (qmail 27884 invoked from network); 28 Apr 2015 00:46:03 -0000
X-Spam-Checker-Version: SpamAssassin 3.3.2 (2011-06-06) on f.primenet.com.au
X-Spam-Level: 
X-Spam-Status: No, score=-2.7 required=5.0 tests=BAYES_00,DKIM_SIGNED,
	DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,HTML_MESSAGE,RCVD_IN_DNSWL_LOW
	autolearn=ham version=3.3.2
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
        d=gmail.com; s=20120113;
        h=mime-version:date:message-id:subject:from:to:content-type;
        bh=mK4Mh4eGSx3hX22VURW79lCbze2WmhvcIdfTKWUmsi8=;
        b=XY4feFGPT4nCaExwLNAA5Gr/xWiVqd1Z1s64BRCe4XqJYnq0s0VeIUxIuLf+Q3902X
         qAUUUJMPFSJpOI9uTa6u2fkEZivDuN3+gbXMqOfPfipRreomn25xoi9ZrwTjFsjyaWsp
         faSInjzwaa7uJTd3B/SOa/D4vP0mTomN1FVIXXDcQClUwG1CuZsZBhMfCx99DlgJuD1N
         CrYfnbOf0OoAJU8kkXoE4JnDJRjhBPNMfordx9NevnVjyTR1C0nezGHwq6TOtNaNAJzi
         QZA4frsXNFM4Wf5sL2HfGFzwSo9yBJRSaJt8Nt08A93vdH92yAZ6TJG37jSqTOawoZm1
         Jeiw==
MIME-Version: 1.0
X-Received: by 10.67.15.102 with SMTP id fn6mr27602341pad.120.1430181955879;
 Mon, 27 Apr 2015 17:45:55 -0700 (PDT)
Date: Mon, 27 Apr 2015 20:45:55 -0400
Message-ID: <CA+80RYdqaKGNXE9J9gnT7uNcbmPmqoUHjw+w=N_+jwC+gm_7Lg@mail.gmail.com>
Subject: Proposed patches for make completion bugs
From: Jared Ahern <jared.ahern@gmail.com>
To: zsh-workers@zsh.org
Content-Type: multipart/alternative; boundary=001a11343eca71c9310514be2f87

--001a11343eca71c9310514be2f87
Content-Type: text/plain; charset=UTF-8

Hello,

I've been using zsh for years and it is excellent.  One thing that I
realized recently was that it was not completing the "make" command
correctly, so I investigated, using version 5.0.7 as a base.  As far as I
can tell, my problems resulted from two issues.  The first was that zsh did
not recognize the ?= or ::= forms of variable assignment in at least GNU
make.  This should be corrected with this patch:

diff -rup a/Completion/Unix/Command/_make b/Completion/Unix/Command/_make
--- a/Completion/Unix/Command/_make    2013-11-27 14:00:19.000000000 -0500
+++ b/Completion/Unix/Command/_make    2015-04-27 20:17:05.000000000 -0400
@@ -64,18 +64,18 @@ _make-parseMakefile () {
   while read input
   do
     case "$input " in
-      # VARIABLE = value
-      ([[:alnum:]][[:alnum:]_]#[ $TAB]#=*)
-      var=${input%%[ $TAB]#=*}
+      # VARIABLE = value OR VARIABLE ?= value
+      ([[:alnum:]][[:alnum:]_]#[ $TAB]#(\?|)=*)
+      var=${input%%[ $TAB]#(\?|)=*}
       val=${input#*=}
       val=${val##[ $TAB]#}
       VARIABLES[$var]=$val
       ;;

-      # VARIABLE := value
+      # VARIABLE := value OR VARIABLE ::= value
       # Evaluated immediately
-      ([[:alnum:]][[:alnum:]_]#[ $TAB]#:=*)
-      var=${input%%[ $TAB]#:=*}
+      ([[:alnum:]][[:alnum:]_]#[ $TAB]#:(:|)=*)
+      var=${input%%[ $TAB]#:(:|)=*}
       val=${input#*=}
       val=${val##[ $TAB]#}
       val=$(_make-expandVars $val)

The second issue is that as far as I can tell, expansion of "make"
variables was not happening at all.  This should be addressed by this
separate second patch, although I may have induced regressions of which I
am unaware.  I wasn't 100% sure what the original intent was for the
_make-expandVars function, but my cases appear to be fixed.

diff -rup a/Completion/Unix/Command/_make c/Completion/Unix/Command/_make
--- a/Completion/Unix/Command/_make    2013-11-27 14:00:19.000000000 -0500
+++ c/Completion/Unix/Command/_make    2015-04-27 20:17:17.000000000 -0400
@@ -7,51 +7,52 @@ _make-expandVars() {
   local open close var val front ret tmp=$1

   front=${tmp%%\$*}
+  tmp=${tmp#$front}
   case $tmp in
-    (\(*) # Variable of the form $(foobar)
+    (\$\(*) # Variable of the form $(foobar)
     open='('
     close=')'
     ;;

-    ({*) # ${foobar}
+    (\${*) # ${foobar}
     open='{'
     close='}'
     ;;

-    ([[:alpha:]]*) # $foobar. This is exactly $(f)oobar.
+    (\$[[:alpha:]]*) # $foobar. This is exactly $(f)oobar.
     open=''
     close=''
-    var=${(s::)var[1]}
+    var=${(s::)tmp[2]}
     ;;

-    (\$*) # Escaped $.
+    (\$\$*) # Escaped $.
     print -- "${front}\$$(_make-expandVars ${tmp#\$})"
     return
     ;;

     (*) # Nothing left to substitute.
-    print -- $tmp
+    print -- ${front}${tmp}
     return
     ;;
   esac

   if [[ -n $open ]]
   then
-    var=${tmp#$open}
+    var=${tmp#\$$open}
     var=${var%%$close*}
   fi

   case $var in
     ([[:alnum:]_]#)
     val=${VARIABLES[$var]}
-    ret=${ret//\$$open$var$close/$val}
+    ret=${tmp//\$$open$var$close/$val}
     ;;

     (*)
     # Improper variable name. No replacement.
     # I'm not sure if this is desired behavior.
     front+="\$$open$var$close"
-    ret=${ret/\$$open$var$close/}
+    ret=${tmp/\$$open$var$close/}
     ;;
   esac

The sourceforge site for zsh indicated that proposed patches should be sent
to this email address, but please email me if I should do anything
different, or if you need more info.  Hopefully I've included all the
correct patch text. With luck, these two changes will make it into a future
version of zsh!

Thanks,
Jared

--001a11343eca71c9310514be2f87--