From mboxrd@z Thu Jan 1 00:00:00 1970 From: Martin Kelly Date: Wed, 18 May 2016 15:13:04 -0700 Subject: [Buildroot] [PATCH v2] Config.in: add -Og option In-Reply-To: <2cb76318-9eb7-eacc-cdc3-e014604d2628@mind.be> References: <1463442955-8178-1-git-send-email-martin@surround.io> <7e07c970-b00e-f605-7192-3152d53885e1@mind.be> <573CCB3C.1040703@surround.io> <2cb76318-9eb7-eacc-cdc3-e014604d2628@mind.be> Message-ID: <573CE8F0.80200@surround.io> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On 05/18/2016 03:00 PM, Arnout Vandecappelle wrote: > On 05/18/16 22:06, Martin Kelly wrote: >> On 05/18/2016 12:52 PM, Arnout Vandecappelle wrote: >>> On 05/17/16 01:55, Martin Kelly wrote: >>>> -Og (introduced in GCC 4.8) lets you optimize for debugging experience, >>>> which can be useful for when you want optimized code that is >>>> nonetheless >>>> debuggable. >>>> >>>> Signed-off-by: Martin Kelly >>>> --- >>>> Changes based on feedback: >>>> - select --> depends on >>>> - Reworded help text >>>> - Wrapped text to 72 lines >>> >>> Well, actually you didn't: you just copied my text, which I >>> incorrectly wrapped at 78 columns instead of 72... >>> >> >> You may have wrapped to 78, but I rewrapped it to 72 columns, so I >> think the >> patch I sent is correctly wrapped. > > By my count, the line 'reasonable level of optimization while > maintaining fast compilation' is 68 characters long. with the tab + 2 > spaces that becomes 78. > I was working under the assumption that tabs count as 1 character. If tabs count as 8 characters instead, then the rest of the file is already miswrapped; there are many lines containing a tab, 2 spaces, and 70 characters after. Under BR_OPTIMIZE_2, the line starting with "the performance of the generated code" is an example of that. In addition, most text editors seem to count a tab as 1 character when displaying width (Vim certainly does). If the intention is to count a tab as 8 characters, I'd be happy to do so, but then we should also rewrap the rest of the file for consistency. >> >>>> --- >>>> >>>> Config.in | 10 ++++++++++ >>>> package/Makefile.in | 3 +++ >>>> 2 files changed, 13 insertions(+) >>>> >>>> diff --git a/Config.in b/Config.in >>>> index 9bc8e51..3fe6b7a 100644 >>>> --- a/Config.in >>>> +++ b/Config.in >>>> @@ -510,6 +510,16 @@ config BR2_OPTIMIZE_3 >>>> and also turns on the -finline-functions, -funswitch-loops and >>>> -fgcse-after-reload options. >>>> >>>> +config BR2_OPTIMIZE_g >>> >>> I didn't notice this the first time: config options should be all >>> capitals, like BR2_OPTIMIZE_S (for the -Os option). >>> >> >> I will change this and send a revised patch. Note that there are >> currently >> several config options that are not all capital (e.g. >> BR2_STRIP_strip), but >> perhaps those should change too. > > Historical accident. It's not important enough to change it. > Agreed.