617913 - FuzzOneInvalidCaseConversion could check result of CaseInsensitiveCompare/CharByCharCompareEqual

Status: RESOLVED FIXED

(Core :: Internationalization, defect)

Description

[timeless]

coverity likes to be able to warn when the result of a function is only sometimes checked. typically when its result isn't checked, that's an accidental omission.

Comment 1

[timeless]

Attachment: check function calls

Comment 2

[Simon Montagu :smontagu]

Comment on attachment 496483 [details] [diff] [review]
check function calls

>-  CaseInsensitiveCompare((char*)aBuf, (char*)bBuf, aLen, bLen);
>-  CharByCharCompareEqual((char*)aBuf, (char*)bBuf, aLen, bLen);
>+  if (CaseInsensitiveCompare((char*)aBuf, (char*)bBuf, aLen, bLen))
>+    printf("\tSurprise, two random strings compared insensitively as equal!\n");

Shouldn't this be |if (!CaseInsensitiveCompare...| ?

>+  if (CharByCharCompareEqual((char*)aBuf, (char*)bBuf, aLen, bLen))
>+    printf("\tSurprise, two random strings compared as exactly equal!\n");

Comment 3

[timeless]

afaict they return true if strings compare as equal. since both strings are random, that shouldn't happen, so i'm checking for true.

Comment 4

[Simon Montagu :smontagu]

Hmm. Afa*I*ct CaseInsensitiveCompare returns strcmp-style result codes so you should be testing for 0.

http://mxr.mozilla.org/mozilla-central/source/intl/unicharutil/util/nsUnicharUtils.cpp#478

Comment 5

[timeless]

Attachment: oh!

right, one's a Compare and one's an Equals, how confusing. thanks.

Comment 6

[Benjamin Smedberg]

Comment on attachment 497253 [details] [diff] [review]
oh!

Test-only, so approval really isn't needed.

Comment 7

[Daniel Holbert [:dholbert]]

http://hg.mozilla.org/mozilla-central/rev/07cc6b9b1596