Zsh Mailing List Archive
Messages sorted by: Reverse Date, Date, Thread, Author

legend for match_str (was: Re: file completion(?) erases word typed)

Bart Schaefer wrote on Tue, Aug 30, 2016 at 00:03:31 -0700:
> ainfo->line->olen appears ultimately to come from add_match_part()
> called from match_str() at line 739 of compmatch.c; it's merged from
> there into ainfo by join_clines() called from add_match_data().  At
> this point, though, I get lost trying to keep track of what's going
> on with l + loff and llen (which comes from mp->llen) and ll and all
> of Sven's other usefully named variables.

Some variable name documentation attached.

I haven't renamed them for various reasons, e.g., to keep 'blame' usable
and to maintain similarity to other functions which may also use the
same short names.

I'm planning to commit this with distinct X-Seq numbers such as
39309/{0001..0011} — didn't want to send 11 separate emails for such
trivial changes.

Cheers,

Daniel

From f2fd645003518e60a8045b3c9c42dcf4488932ae Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:34 +0000
Subject: [PATCH 01/10] internals: match_str: Document some local variables. 
 See 39123.

---
 Src/Zle/compmatch.c | 54 +++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 52 insertions(+), 2 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index 0e41ac3..a920c8a 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -593,9 +593,59 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		    continue;
 
 		if (mp->wlen < 0) {
-		    int both, loff, aoff, llen, alen, zoff, moff, ct, ict, aol;
+		    /* 
+		     * 1 iff the anchor and the word are on the same side of
+		     * the line pattern; that is: if either
+		     * - the anchor is on the left and we are matching a prefix; or
+		     * - the anchor is on the right and we are matching a suffix.
+		     */
+		    int both;
+		    /*
+		     * Offset from the line pattern pointer ('l') to the start of the
+		     * line pattern.
+		     */
+		    int loff;
+		    /*
+		     * Offset from the line pattern pointer ('l') to the start of the
+		     * anchor.
+		     */
+		    int aoff;
+		    /*
+		     * The length of the line pattern.
+		     */
+		    int llen;
+		    /*
+		     * The length of the anchor.
+		     *
+		     * SEE: ap; aol, aop
+		     */
+		    int alen;
+		    /*
+		     * ### These two are related: they're set symmetrically.
+		     */
+		    int zoff, moff;
+		    /*
+		     * ### These two are related.
+		     */
+		    int ct, ict;
+		    /*
+		     * The length of the OTHER anchor: the left anchor when we're anchored
+		     * on the right, and of the right anchor when we're anchored on the left.
+		     */
+		    int aol;
+		    /*
+		     * LOST: Documentation comment.  Last seen 10 years ago in the temporal lobe.
+		     * Reward promised for its safe return.  Contact zsh-workers@xxxxxxx.
+		     */
 		    char *tp, savl = '\0', savw;
-		    Cpattern ap, aop;
+		    /*
+		     * The anchor on this end.
+		     */
+		    Cpattern ap;
+		    /*
+		     * The anchor on the other end.
+		     */
+		    Cpattern aop;
 
 		    /* This is for `*' patterns, first initialise some
 		     * local variables. */

From 15b342777a9fe5b8d20c2f5119c0d704091d9900 Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:34 +0000
Subject: [PATCH 02/10] internals: match_str: Simplify by removing 'zoff'.

'zoff' was only used within 'if (sfx)' blocks, in which case it was initialized
to 'alen', so simply s/zoff/alen/g.  'alen' is not const but it first changes
on line 794, after the last use of 'zoff'.
---
 Src/Zle/compmatch.c | 20 ++++++++++----------
 1 file changed, 10 insertions(+), 10 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index a920c8a..1e5f551 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -621,9 +621,9 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		     */
 		    int alen;
 		    /*
-		     * ### These two are related: they're set symmetrically.
+		     * ### Related to 'zoff', which was removed in 2016.
 		     */
-		    int zoff, moff;
+		    int moff;
 		    /*
 		     * ### These two are related.
 		     */
@@ -661,14 +661,14 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 			continue;
 
 		    if (mp->flags & CMF_LEFT) {
-			ap = mp->left; zoff = 0; moff = alen; aop = mp->right;
+			ap = mp->left; moff = alen; aop = mp->right;
 			if (sfx) {
 			    both = 0; loff = -llen; aoff = -(llen + alen);
 			} else {
 			    both = 1; loff = alen; aoff = 0;
 			}
 		    } else {
-			ap = mp->right; zoff = alen; moff = 0; aop = mp->left;
+			ap = mp->right; moff = 0; aop = mp->left;
 			if (sfx) {
 			    both = 1; loff = -(llen + alen); aoff = -alen;
 			} else {
@@ -694,8 +694,8 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 
 		    /* Fine, now we call ourselves recursively to find the
 		     * string matched by the `*'. */
-		    if (sfx && (savl = l[-(llen + zoff)]))
-			l[-(llen + zoff)] = '\0';
+		    if (sfx && (savl = l[-(llen + alen)]))
+			l[-(llen + alen)] = '\0';
 		    for (t = 0, tp = w, ct = 0, ict = lw - alen + 1;
 			 ict;
 			 tp += add, ct++, ict--) {
@@ -717,12 +717,12 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 				!match_parts(l + aoff , tp - moff, alen, part))
 				break;
 			    if (sfx) {
-				if ((savw = tp[-zoff]))
-				    tp[-zoff] = '\0';
+				if ((savw = tp[-alen]))
+				    tp[-alen] = '\0';
 				t = match_str(l - ll, w - lw,
 					      NULL, 0, NULL, 1, 2, part);
 				if (savw)
-				    tp[-zoff] = savw;
+				    tp[-alen] = savw;
 			    } else
 				t = match_str(l + llen + moff, tp + moff,
 					      NULL, 0, NULL, 0, 1, part);
@@ -732,7 +732,7 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		    }
 		    ict = ct;
 		    if (sfx && savl)
-			l[-(llen + zoff)] = savl;
+			l[-(llen + alen)] = savl;
 
 		    /* Have we found a position in w where the rest of l
 		     * matches? */

From 6333669d7dc9525eb8103d993d6bf90c002c97d0 Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:34 +0000
Subject: [PATCH 03/10] internals: match_str: Document 'savw'.  Avoid magic
 number.

All callees checked to ensure that they only check that parameter for nonzeroness.
---
 Src/Zle/compmatch.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index 1e5f551..654eec4 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -637,7 +637,7 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		     * LOST: Documentation comment.  Last seen 10 years ago in the temporal lobe.
 		     * Reward promised for its safe return.  Contact zsh-workers@xxxxxxx.
 		     */
-		    char *tp, savl = '\0', savw;
+		    char *tp, savl = '\0';
 		    /*
 		     * The anchor on this end.
 		     */
@@ -717,15 +717,18 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 				!match_parts(l + aoff , tp - moff, alen, part))
 				break;
 			    if (sfx) {
+				/* Call ourselves recursively with the
+				 * anchor removed. */
+				char savw;
 				if ((savw = tp[-alen]))
 				    tp[-alen] = '\0';
 				t = match_str(l - ll, w - lw,
-					      NULL, 0, NULL, 1, 2, part);
+					      NULL, 0, NULL, sfx, 2, part);
 				if (savw)
 				    tp[-alen] = savw;
 			    } else
 				t = match_str(l + llen + moff, tp + moff,
-					      NULL, 0, NULL, 0, 1, part);
+					      NULL, 0, NULL, sfx, 1, part);
 			    if (t || (mp->wlen == -1 && !both))
 				break;
 			}

From 8f7e6f89ba493f47939adc093b1708597fba9705 Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:34 +0000
Subject: [PATCH 04/10] internals: match_str: Document 'savl'.

---
 Src/Zle/compmatch.c | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index 654eec4..001a166 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -637,7 +637,22 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		     * LOST: Documentation comment.  Last seen 10 years ago in the temporal lobe.
 		     * Reward promised for its safe return.  Contact zsh-workers@xxxxxxx.
 		     */
-		    char *tp, savl = '\0';
+		    char *tp;
+		    /* 
+		     * Temporary variable.  Used as temporary storage for a
+		     *
+		     *     {
+		     *         () {
+		     *           local foo="$foo"
+		     *           foo[1]=bar
+		     *           ...
+		     *         }
+		     *         (use original $foo here)
+		     *     }
+		     *
+		     * operation.  Similar to savw.
+		     */
+		    char savl;
 		    /*
 		     * The anchor on this end.
 		     */

From 4fb33a729171beba2309cae3cf3ec5866950babe Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:34 +0000
Subject: [PATCH 05/10] internals: match_str: Constify some local variables.

---
 Src/Zle/compmatch.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index 001a166..c51d849 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -498,14 +498,17 @@ add_match_sub(Cmatcher m, char *l, int ll, char *w, int wl)
 /**/
 int
 match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
-	  int sfx, int test, int part)
+	  const int sfx, int test, int part)
 {
     int ll = strlen(l), lw = strlen(w), oll = ll, olw = lw, exact = 0, wexact = 0;
-    int il = 0, iw = 0, t, ind, add, he = 0, bpc, obc = bc, bslash;
+    int il = 0, iw = 0, t, he = 0, bpc, bslash;
     char *ow;
-    Cmlist ms;
+    Cmlist ms; /* loop variable */
     Cmatcher mp, lm = NULL;
     Brinfo bp = NULL;
+    const int obc = bc;
+    const int ind = (sfx ? -1 : 0);
+    const int add = (sfx ? -1 : 1);
 
     if (!test) {
 	start_match();
@@ -516,9 +519,6 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 
     if (sfx) {
 	l += ll; w += lw;
-	ind = -1; add = -1;
-    } else {
-	ind = 0; add = 1;
     }
     /* ow will always point to the beginning (or end) of that sub-string
      * in w that wasn't put in the match-variables yet. */

From 9151d81b576f0b7d560f828b1a28f02326ccec51 Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:34 +0000
Subject: [PATCH 06/10] internals: match_str: Downscope local variable 'bpc'.

---
 Src/Zle/compmatch.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index c51d849..f436a49 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -501,7 +501,7 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 	  const int sfx, int test, int part)
 {
     int ll = strlen(l), lw = strlen(w), oll = ll, olw = lw, exact = 0, wexact = 0;
-    int il = 0, iw = 0, t, he = 0, bpc, bslash;
+    int il = 0, iw = 0, t, he = 0, bslash;
     char *ow;
     Cmlist ms; /* loop variable */
     Cmatcher mp, lm = NULL;
@@ -820,12 +820,14 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		    bc += llen;
 		    exact = 0;
 
-		    if (!test)
+		    if (!test) {
+			int bpc;
 			while (bp &&
 			       bc >= (bpc = (useqbr ? bp->qpos : bp->pos))) {
 			    bp->curpos = matchbufadded + bpc - bc + obc;
 			    bp = bp->next;
 			}
+		    }
 		    ow = w;
 
 		    if (!llen && !alen) {
@@ -943,12 +945,14 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		    bc += mp->llen;
 		    exact = 0;
 
-		    if (!test)
+		    if (!test) {
+			int bpc;
 			while (bp &&
 			       bc >= (bpc = (useqbr ? bp->qpos : bp->pos))) {
 			    bp->curpos = matchbufadded + bpc - bc + obc;
 			    bp = bp->next;
 			}
+		    }
 		    ow = w;
 		    lm = NULL;
 		    he = 0;

From 7db7a9a55f3897e34300ef7462c45fda4a7fbc91 Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:35 +0000
Subject: [PATCH 07/10] internals: match_str: Rename and constify local
 variables 'oll', 'olw'.

---
 Src/Zle/compmatch.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index f436a49..673347e 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -500,7 +500,7 @@ int
 match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 	  const int sfx, int test, int part)
 {
-    int ll = strlen(l), lw = strlen(w), oll = ll, olw = lw, exact = 0, wexact = 0;
+    int ll = strlen(l), lw = strlen(w), exact = 0, wexact = 0;
     int il = 0, iw = 0, t, he = 0, bslash;
     char *ow;
     Cmlist ms; /* loop variable */
@@ -509,6 +509,7 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
     const int obc = bc;
     const int ind = (sfx ? -1 : 0);
     const int add = (sfx ? -1 : 1);
+    const int original_ll = ll, original_lw = lw;
 
     if (!test) {
 	start_match();
@@ -585,7 +586,7 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 	    for (mp = ms->matcher; mp; mp = mp->next) {
 		t = 1;
 		if ((lm && lm == mp) ||
-		    ((oll == ll || olw == lw) &&
+		    ((original_ll == ll || original_lw == lw) &&
 		     (test == 1 || (test && !mp->left && !mp->right)) &&
 		     mp->wlen < 0))
 		    /* If we were called recursively, don't use `*' patterns

From 62442f34017813d9241a2808f4ddf167e12fb68e Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:35 +0000
Subject: [PATCH 08/10] internals: match_str: Document several local variables.

---
 Src/Zle/compmatch.c | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index 673347e..16705f6 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -500,8 +500,15 @@ int
 match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 	  const int sfx, int test, int part)
 {
-    int ll = strlen(l), lw = strlen(w), exact = 0, wexact = 0;
-    int il = 0, iw = 0, t, he = 0, bslash;
+    /* How many characters from the line string and from the word string are yet to be matched. */
+    int ll = strlen(l), lw = strlen(w);
+    /* Number of characters from the line string and word string matched. */
+    int il = 0, iw = 0;
+    /* How many characters were matched exactly in the line and in the word. */
+    int exact = 0, wexact = 0;
+    int he = 0;
+    int bslash;
+    int t;
     char *ow;
     Cmlist ms; /* loop variable */
     Cmatcher mp, lm = NULL;
@@ -511,6 +518,8 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
     const int add = (sfx ? -1 : 1);
     const int original_ll = ll, original_lw = lw;
 
+    /* INVARIANT: il+ll == original_ll; iw+lw == original_lw */
+
     if (!test) {
 	start_match();
 	bp = *bpp;
@@ -627,6 +636,10 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		    int moff;
 		    /*
 		     * ### These two are related.
+		     *
+		     * ### They may have a relation similar to that of lw/iw (q.v.), 
+		     * ### at least during the 'for' loop.  They may be overloaded/repurposed
+		     * ### after it.
 		     */
 		    int ct, ict;
 		    /*

From 1304e6473db8d1fdb6d7a8ada741a61a38984b88 Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:35 +0000
Subject: [PATCH 09/10] internals: match_str: Downscope local variable 't'.

Remove needless initialization (it is written to again before it is ever read).

Note there was another 't' variable at the end of the function that shadowed
the int 't'.
---
 Src/Zle/compmatch.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index 16705f6..1d4bef8 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -508,7 +508,6 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
     int exact = 0, wexact = 0;
     int he = 0;
     int bslash;
-    int t;
     char *ow;
     Cmlist ms; /* loop variable */
     Cmatcher mp, lm = NULL;
@@ -593,7 +592,6 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 	/* First try the matchers. Err... see above. */
 	for (mp = NULL, ms = mstack; !mp && ms; ms = ms->next) {
 	    for (mp = ms->matcher; mp; mp = mp->next) {
-		t = 1;
 		if ((lm && lm == mp) ||
 		    ((original_ll == ll || original_lw == lw) &&
 		     (test == 1 || (test && !mp->left && !mp->right)) &&
@@ -603,6 +601,11 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		    continue;
 
 		if (mp->wlen < 0) {
+		    /* `*'-pattern. */
+		    /*
+		     * Similar to the identically-named variable in the 'else' block.
+		     */
+		    int t;
 		    /* 
 		     * 1 iff the anchor and the word are on the same side of
 		     * the line pattern; that is: if either
@@ -856,6 +859,10 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 		    break;
 		} else if (ll >= mp->llen && lw >= mp->wlen) {
 		    /* Non-`*'-pattern. */
+		    /*
+		     * Similar to the identically-named variable in the 'if' block.
+		     */
+		    int t = 1;
 		    char *tl, *tw;
 		    int tll, tlw, til, tiw;
 

From d5b919941b4d20641cafd9ff7d329a80d144126c Mon Sep 17 00:00:00 2001
From: Daniel Shahaf <d.s@xxxxxxxxxxxxxxxxxx>
Date: Wed, 14 Sep 2016 03:38:35 +0000
Subject: [PATCH 10/10] internals: match_str: Simplify expression.

In the first hunk we actually know that ind==0 since sfx==0, but keep it identical
to the last hunk.

Also add a comment (unrelated).
---
 Src/Zle/compmatch.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/Src/Zle/compmatch.c b/Src/Zle/compmatch.c
index 1d4bef8..193b105 100644
--- a/Src/Zle/compmatch.c
+++ b/Src/Zle/compmatch.c
@@ -568,8 +568,7 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 	bslash = 0;
 	if (!sfx && lw && (!part || test) &&
 	    (l[ind] == w[ind] ||
-	     (bslash = (lw > 1 && w[ind] == '\\' &&
-			(ind ? (w[0] == l[0]) : (w[1] == l[0])))))) {
+	     (bslash = (lw > 1 && w[ind] == '\\' && w[ind+1] == l[0])))) {
 	    /* No matcher could be used, but the strings have the same
 	     * character here, skip over it. */
 	    l += add; w += (bslash ? (add + add) : add);
@@ -849,8 +848,10 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 
 		    if (!llen && !alen) {
 			lm = mp;
-			if (he)
+			if (he) {
+			    /* Signal the outer for loop to continue. */
 			    mp = NULL;
+			}
 			else
 			    he = 1;
 		    } else {
@@ -989,8 +990,7 @@ match_str(char *l, char *w, Brinfo *bpp, int bc, int *rwlp,
 	bslash = 0;
 	if ((!test || sfx) && lw &&
 	    (l[ind] == w[ind] ||
-	     (bslash = (lw > 1 && w[ind] == '\\' &&
-			(ind ? (w[0] == l[0]) : (w[1] == l[0])))))) {
+	     (bslash = (lw > 1 && w[ind] == '\\' && w[ind+1] == l[0])))) {
 	    /* No matcher could be used, but the strings have the same
 	     * character here, skip over it. */
 	    l += add; w += (bslash ? (add + add ) : add);
Messages sorted by: Reverse Date, Date, Thread, Author