threads / patch / 3831

patchxdiff/xdiffi.c: fix warnings about possibly uninitialized variables

Subject: [PATCH] xdiff/xdiffi.c: fix warnings about possibly uninitialized variables

## tl;dr

2 messages between Apr 8, 2006 and Apr 8, 2006. Diffs are folded; open one to read it.

replies: 1people: 2as markdown or json

Marco Roeland· Apr 8, 2006, 15:27 UTC · lore
Compiling this module gave the following warnings (some double dutch!):

xdiff/xdiffi.c: In functie 'xdl_recs_cmp': xdiff/xdiffi.c:298: let op: 'spl.i1' may be used uninitialized in this function xdiff/xdiffi.c:298: let op: 'spl.i2' may be used uninitialized in this function xdiff/xdiffi.c:219: let op: 'fbest1' may be used uninitialized in this function xdiff/xdiffi.c:219: let op: 'bbest1' may be used uninitialized in this function

A superficial tracking of their usage, without deeper knowledge about the algorithm, indeed confirms that there are code paths on which these variables will be used uninitialized. In practice these code paths might never be reached, but then these fixes will not change the algorithm. If these code paths are ever reached we now at least have a predictable outcome. And should the very small performance impact of these initializations be noticeable, then they should at least be replaced by comments why certain code paths will never be reached.

Some extra initializations in this patch now fix the warnings.
---
 xdiff/xdiffi.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)
0b0bf00d67a66b3ef47862cc51b1d37763f4b99b
Show changes to xdiff/xdiffi.c +3 −2
diff --git a/xdiff/xdiffi.c b/xdiff/xdiffi.c
index e81bca6..641362d 100644
--- a/xdiff/xdiffi.c
+++ b/xdiff/xdiffi.c
@@ -218,7 +218,7 @@ static long xdl_split(unsigned long cons
 		if (ec >= xenv->mxcost) {
 			long fbest, fbest1, bbest, bbest1;
 
-			fbest = -1;
+			fbest = fbest1 = -1;
 			for (d = fmax; d >= fmin; d -= 2) {
 				i1 = XDL_MIN(kvdf[d], lim1);
 				i2 = i1 - d;
@@ -230,7 +230,7 @@ static long xdl_split(unsigned long cons
 				}
 			}
 
-			bbest = XDL_LINE_MAX;
+			bbest = bbest1 = XDL_LINE_MAX;
 			for (d = bmax; d >= bmin; d -= 2) {
 				i1 = XDL_MAX(off1, kvdb[d]);
 				i2 = i1 - d;
@@ -296,6 +296,7 @@ int xdl_recs_cmp(diffdata_t *dd1, long o
 	} else {
 		long ec;
 		xdpsplit_t spl;
+		spl.i1 = spl.i2 = 0;
 
 		/*
 		 * Divide ...

-- 
1.3.0.rc3.gad0b
Davide Libenzi· Apr 8, 2006, 17:18 UTC · re: Marco Roeland · lore

Re: [PATCH] xdiff/xdiffi.c: fix warnings about possibly uninitialized variables

On Sat, 8 Apr 2006, Marco Roeland wrote:
Show 16 quoted lines
> Compiling this module gave the following warnings (some double dutch!):
>
> xdiff/xdiffi.c: In functie 'xdl_recs_cmp':
> xdiff/xdiffi.c:298: let op: 'spl.i1' may be used uninitialized in this function
> xdiff/xdiffi.c:298: let op: 'spl.i2' may be used uninitialized in this function
> xdiff/xdiffi.c:219: let op: 'fbest1' may be used uninitialized in this function
> xdiff/xdiffi.c:219: let op: 'bbest1' may be used uninitialized in this function
>
> A superficial tracking of their usage, without deeper knowledge about the
> algorithm, indeed confirms that there are code paths on which these
> variables will be used uninitialized. In practice these code paths might never
> be reached, but then these fixes will not change the algorithm. If these
> code paths are ever reached we now at least have a predictable outcome. And
> should the very small performance impact of these initializations be
> noticeable, then they should at least be replaced by comments why certain
> code paths will never be reached.

These paths are never reached because of the way data is prepared before and passed to the function. Unfortunately the compiler cannot know this. Using them as -1 or XDL_LINE_MAX won't help either, since those are out of domain values. You can leave it there and the algo won't suffer, or you can relax a little the warning level when building the file.

- Davide

← back to recent threads