[PATCH] Test-suite McCat signed char

Kristof Beyls kristof.beyls at arm.com
Tue Mar 5 01:28:59 PST 2013


Hi Renato,

 

Thanks for working on this.

 

I think that the ifdef code:

 

+#ifdef __arm__

+typedef signed char char_t;

+#else

+typedef char char_t;

+#endif

 

might be better written as just an unconditional "typedef signed char
char_t" for all architectures.
Maybe a one-line comment would be useful to state that the test relies on
chars being signed, and that chars are not signed by default on all
architectures?

 

By dropping the "ifdef __arm__" part, the test would also be fixed on any
other architectures on which chars are unsigned by default.



Thanks,

 

Kristof

 

From: llvm-commits-bounces at cs.uiuc.edu
[mailto:llvm-commits-bounces at cs.uiuc.edu] On Behalf Of Renato Golin
Sent: 05 March 2013 09:01
To: LLVM Commits
Subject: [PATCH] Test-suite McCat signed char

 

Creating a typedef for ARM, since the test assumes signed char.

 

Changing all chars to "signed char" would also work on x86, but I wanted to
make it explicit why we're using it (ifdef __arm__) rather than leaving it
to anyone's guess.

 

Thanks Tim for pointing the issue.

 

Can I commit?

 

cheers,

--renato
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130305/7614c594/attachment.html>


More information about the llvm-commits mailing list