678845 - testIndexToString.cpp:46:5: warning: this decimal constant is unsigned only in ISO C90

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Daniel Holbert [:dholbert]]

Filing bug on this GCC build warnings:
{
js/src/jsapi-tests/testIndexToString.cpp:46:5: warning: this decimal constant is unsigned only in ISO C90
js/src/jsapi-tests/testIndexToString.cpp:47:5: warning: this decimal constant is unsigned only in ISO C90
js/src/jsapi-tests/testIndexToString.cpp:48:5: warning: this decimal constant is unsigned only in ISO C90
js/src/jsapi-tests/testIndexToString.cpp:49:5: warning: this decimal constant is unsigned only in ISO C90
}

This is from this chunk:
> static const struct TestPair {
>     uint32 num;
>     const char *expected;
> } tests[] = {
[----------SNIP----------]
>     { 2147483648, "2147483648" },
>     { 2147483649, "2147483649" },
>     { 4294967294, "4294967294" },
>     { 4294967295, "4294967295" },
> };

The compiler is complaining about those integer literals - like all integer literals, they'll be treated as signed values in the compiled code -- but in this case, they're larger than INT_MAX, so they'll end up being represented with negative signed values.

Since we're trying to store them in a uint32 variable, we presumably want to consider them unsigned.  We can address this easily by just adding a "u" suffix and taking the implicit signed-int representation out of the loop.

Comment 1

[Daniel Holbert [:dholbert]]

Attachment: fix: add 'u' suffix

Comment 2

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

Why not do that with the others as well?

Comment 3

[Daniel Holbert [:dholbert]]

'cause it's not necessary? If Waldo prefers, I'd be happy to make that change.

(Of course, it's technically "not necessary" with the ones touched by the patch, too -- it's only necessary to avoid buildspam. :))

Comment 5

[Jeff Walden [:Waldo]]

Comment on attachment 553000 [details] [diff] [review]
fix: add 'u' suffix

Review of attachment 553000 [details] [diff] [review]:
-----------------------------------------------------------------

My policy on tests is that it has to be clear what's happening, and the formatting/naming/etc. can't be misleading.  But beyond that I don't care about nitpickeries of absolute consistency, adherence to style guidelines, and so on.  :-)  So this is fine, and to be honest even if I did have reason to care about style I'm not sure I'd find this wrong.

Comment 6

[Daniel Holbert [:dholbert]]

Landed as-is:
 http://hg.mozilla.org/integration/mozilla-inbound/rev/edd7aab15b8b

Comment 7

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

http://hg.mozilla.org/mozilla-central/rev/edd7aab15b8b