|
From: | Thiemo Seufer |
Subject: | Re: [Qemu-devel] [PATCH] SH7750/51: add register BCR3, NCR4, PCR, RTCOR, RTCNT, RTCSR, SDMR2, SDMR3 and fix BCR2 support |
Date: | Mon, 15 Dec 2008 02:08:42 +0100 |
User-agent: | Mutt/1.5.18 (2008-05-17) |
Shin-ichiro KAWASAKI wrote: [snip] >>>>>> Compared to 'if' statement, 'switch-case' might be more easy to >>>>>> understand, like as follows. >>>>>> case SH7750_SDMR2 ... SH7750_SDMR2 + SDMR2_REGNB >>>>> ok elvenif it's a gcc ony extension >>>> Ah, gccism policy seems not clear in QEMU project. >>>> This extension is used in many hw/*.c, so I guessed it >>>> as a QEMU's usual implementation style. >>>> I hope comments on it from main developers. >>>> >>> Yes it will be nice >> >> I believe "no new gcc-isms" would be a good policy. > > Thank you. I'll follow it. Be aware that this was only my personal opinion, and not well-stated at that. From the recent discussion on IRC I believe we have now the rough consensus: "Use GCC extensions only when it improves the code significantly". > So, my suggestion about 'if' vs 'switch-case' is not appropriate. For the specific patch I believe the 'if' version is better. My opinion is largely formed by the 'SH7750_SDMR2 + SDMR2_REGNB' term: If this state doesn't deserve its own name, then it is probably not clear enough to be used in a case statement. Thiemo
[Prev in Thread] | Current Thread | [Next in Thread] |