pcp
[Top] [All Lists]

Re: [pcp] pcp updates

To: nathans@xxxxxxxxxx
Subject: Re: [pcp] pcp updates
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Fri, 10 Dec 2010 11:07:13 +1100
Cc: pcp@xxxxxxxxxxx
In-reply-to: <1291878180.3214.415.camel@xxxxxxxxxxxxxxxx>
References: <852930361.254181291863051105.JavaMail.root@xxxxxxxxxxxxxxxxxxxxxx> <1291878180.3214.415.camel@xxxxxxxxxxxxxxxx>
Reply-to: kenj@xxxxxxxxxxxxxxxx
On Thu, 2010-12-09 at 18:03 +1100, Ken McDonell wrote:
> On Thu, 2010-12-09 at 13:50 +1100, nathans@xxxxxxxxxx wrote:
> ...
> 
> toolchain problem solved ... -Wall means -Wsfa unless -On is present for
> n >= 1 ... sigh.  Now to see if I can pass QA with -O1 binaries, as I
> know -O2 generates bad code for me.

Well with binaries made with -O1 I get these QA failures (I did have
100% passing the day before):

038 061 072 091 134 135 151 178 179 184 186 202 203 210 214 218 252 261
266 270 280 287 291 292 302 303 311 328 337 338 352 369 418 419 450 460
530 532 650

Remade pmlogger with no -O and these pass: 252 280 418 419

Remade pmdumplog with no -O and these pass: 038 061 072 134 135 179 184
210 311 352 369 650

Remade libpcp with no -O and these pass: 091 287

Remade pmlogextract with no -O and these pass: 151 178 186 202 203 214
218 261 266 270 291 292 302 303 328 337 338 450 460 530 532

So I can have compiler warnings or bad executables ... hmm!


<Prev in Thread] Current Thread [Next in Thread>