Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-4041-

GCC Flags

Name: Anonymous 2009-06-12 11:30

What's your favourite GCC flag?

Name: Anonymous 2009-06-12 11:30

-Wsage

Name: Anonymous 2009-06-12 11:32

-O3

Name: Anonymous 2009-06-12 11:44

-o I'll be die before I have another a.out

Name: Anonymous 2009-06-12 12:01

>>4
lolol, I NEVER use this. ./a.[TAB] <3

Name: Anonymous 2009-06-12 12:06

>>2
-fno-exceptions

Name: Anonymous 2009-06-12 12:07

-funroll-loops

Name: Anonymous 2009-06-12 12:21

-DDEBUG

Name: Anonymous 2009-06-12 12:28

>>6
don't sage if you're replying to the thread. sage if you are saying 'fuck this thread, go to /r/' or whatever. idiot.

Name: Anonymous 2009-06-12 12:36

My favourite flag? Gosh, I don't know. Maybe the flag of the Euroupean Unioun!

Name: Anonymous 2009-06-12 12:40

>>10
don't sage if you're replying to the thread. sage if you are saying 'fuck this thread, go to /r/' or whatever. idiot.

Name: Anonymous 2009-06-12 13:02

-vomit-frame-pointer

Name: Anonymous 2009-06-12 13:05

>>11
You don't really know what sage is then do you

Name: Anonymous 2009-06-12 13:46

>>10
The Queen's language is the one true English language. Anything else is not English.

Name: Anonymous 2009-06-12 13:47

SAGE GOES IN THE COMMENT FIELD

Name: Anonymous 2009-06-12 13:50

>>14,11,9
trolling thread

Name: Anonymous 2009-06-12 16:44

-pedantic -Wall

Name: Anonymous 2009-06-12 16:50

-ffreestanding

Name: Anonymous 2009-06-12 18:33

-fopenmp

Obviously to be used with code that uses OpenMP.

Name: Anonymous 2009-06-12 18:51

-fgrunnur

Name: Anonymous 2009-06-12 18:57

I don't have favorite CFLAGS. I prefer science. I use a genetic algorih to compile every package over 50 generations, profiling the speed of the survivor amongst each generation.

Name: Anonymous 2009-06-12 19:10

>>21
Hello fellow gentoo user :)

Name: Anonymous 2009-06-12 20:41

>8
Oh, I use the same macro in my programs.

Name: Anonymous 2009-06-12 21:28

>>21
Hey, teach!

Name: Anonymous 2009-06-13 4:25

-malicious-double

Name: Anonymous 2009-06-13 5:17

>>7
This. Funrolling loops.

Name: Anonymous 2009-06-13 5:34

>>26
I'll funroll you're loops ;)

Name: Anonymous 2009-06-13 5:46

-Wall

Name: Anonymous 2009-06-13 7:03

Rank: VIPPER  | Subject: Re: GCC flags
Registered:    | |On 2009-06-13 05:34, >>27 said
              |  | I'll funroll you're loops ;)
2005-03-04    |  I'll funroll you are loops.
Posts: 2360    | PROTIP: you're is a contraction of "you are"
               |
 ／l、          |
(ﾟ､ ｡ ７           |
l、 ~ヽ          |_________________________________________________________________________
じしf_, )ノ        | ``Computer science is the black magic of the 20th century'' -- A.C. Cudders

Name: Anonymous 2009-06-13 10:19

>>29
What about MY is a contradiction?

Name: Anonymous 2009-06-13 10:37

-w

Name: Anonymous 2009-06-13 12:58

$ cat defaultMake
CC=gcc
CFLAGS=-ansi -pedantic -Wall -W -Wshadow -Wcast-qual -Wwrite-strings

Name: Anonymous 2009-06-13 13:07

-std=c99

Name: Anonymous 2009-06-13 13:15

-D_GNU_SOURCE

Name: Anonymous 2009-06-13 14:07

>>32
-W
-Wut

Name: Anonymous 2009-06-13 14:14

`-W'  This is a general option similar to `-Wall' which warns about a
      selection of common programming errors, such as functions which
      can return without a value (also known as ``falling off the end of
      the function body''), and comparisons between signed and unsigned
      values. For example, the following function tests whether an un-
      signed integer is negative (which is impossible, of course):
          int
          foo (unsigned int x)
          {
            if (x < 0)
              return 0; /* cannot occur */
            else
              return 1;
          }
      Compiling this function with `-Wall' does not produce a warning...1

1. Gough, Brian. An Introduction to GCC.

Name: Anonymous 2009-06-13 15:24

-笑

Name: Anonymous 2009-06-13 16:15

Static analysis is a piece of shit, -Wall is for faggots. You'll introduce way more bugs trying to silence the retarded warnings than you'll ever solve real problems.

A compiler is like a 4-year-old when it comes to understanding code.

Name: Anonymous 2009-06-13 16:42

You'll introduce way more bugs trying to silence the retarded warnings than you'll ever solve real problems.
[citation needed]

Name: Anonymous 2009-06-13 17:08

>>39
I think his point is that, like with Valgrind and Splint, people have a tendency to bend over backwards to get a compiler to shut up about some perfectly acceptable code.

Name: Anonymous 2009-06-13 17:26

>>39
Off the top of my head, you can read http://www.sqlite.org/testing.html (section 9.0: Static Analysis). I'm sure you can find more if you try, and of course I have my own anecdotal evidence that supports such a claim.

Name: Anonymous 2009-06-13 17:52

>>41
Not >>39, but I'm eager to hear it.

Name: Anonymous 2009-06-13 18:12

>>38
You'll introduce way more bugs trying to silence the retarded warnings than you'll ever solve real problems.

Yeah, fucking gcc telling me my function has an unreachable case. STUPID.

Name: Anonymous 2009-06-13 18:30

>>41
So, basically, if you totally ignore compiler warnings, but write a shitload of tests and have your code widely used for a long time, and THEN start to fix those warnings, two things happens: first, you find that most of the bugs resulting from the suspicios code were already discovered by your users and the aforementioned shitload of tests, second, by rewriting the code, especially unexperienced as you are with the compiler's complaints, you occasionally introduce new bugs. No shit, Sherlock! Yet the conclusion is a non-sequitur, isn't it?

Name: Anonymous 2009-06-13 23:36

>>44
the whole point of sepples is to make it easier to introduce more bugs, because there weren't enough bugs in c programs.

Name: Anonymous 2009-06-14 4:58

>>36
Oh, sorry. It isn't in the man page but it works.

Name: Anonymous 2009-06-14 8:48

>>46
Look under -Wextra.

Name: Anonymous 2010-12-09 8:46

Name: Anonymous 2012-03-02 23:17

I use clang

Name: EXPERT PROGRAMMR 2012-03-03 1:11

-Wall -Wextra -ansi -pedantic bitch!

Name: Anonymous 2012-03-03 5:21

>>49
Enjoy your broken executables.

Name: Anonymous 2012-03-03 10:12

>>51
I alreaSegmentation fault

Name: Anonymous 2012-03-03 11:31

-O4 -s

Name: Anonymous 2012-03-03 11:45

>>38
Are you saying that you're doing a better job emulating the C abstract machine than my implementation is?

Name: Anonymous 2012-03-03 17:49

gcc -get -dubs

Name: Anonymous 2012-03-03 21:58

-g -Wall -Werror -Wextra -pedantic -pedantic-errors -std=c99

MASTER RACE

Name: Anonymous 2012-03-03 22:01

>>56
c99
Nope.

Name: Anonymous 2012-03-03 22:15

>>57
not using c99
Enjoy not legally having stdint.h. That's the only reason why I use c99. The rest of the features are bullshit.

Name: Anonymous 2012-03-03 22:39

>>58
Please do tell me why anyone would need stdint.h, it isn't even implemented on Windows since it's so useless.

Name: Anonymous 2012-03-03 22:41

>>59
Ask frozen void.

Name: Anonymous 2012-03-03 22:42

>>60
Funny joke, FrozenVoid doesn't even use stdint.h, he just assumes that everybody is running the same compiler/architecture/OS as he is and makes his own flawed defines.

Name: Anonymous 2012-03-03 22:45

>>59
You're a mental midget. You want to know why? It's because the standard integral types are defined as 'one byte', 'at least two bytes', 'at least two bytes', 'at least four bytes' and 'at least eight bytes' for char, short, int, long and long long respectively. What if I want the largest possible ((u)intmax_t)? What if I want an exact type ((u)intn_t), say, if I want to work with binary file formats or network protocols without size definitions breaking between architectures? Picture this: I'm reading a file into a struct, and one of the struct's fields needs to be four bytes. What type do I use? Hint: it's not int, because that can be 2 bytes, and it's not long, as on LP64 platforms, it's 8 byted.

Name: Anonymous 2012-03-03 22:48

>>62
YHBT

Name: Anonymous 2012-03-03 22:49

>>63
;_;

Name: Anonymous 2012-03-04 10:03

>>62
I can't believe the day has come,  but I actually agree with Kodak. Not that you couldn't unofficially hack it together in C90 using structs, sizeofs and void pointers, but why bother if the work has already been done and standardized?

Name: Dubs Guy 2012-03-17 15:46

DUBS, DUBS EVERYWHERE!
Don't change these.
Name: Email:
Entire Thread Thread List
All trademarks and copyrights on this site are owned by their respective parties. All comments and posts are the responsibility of their own posters.
- Tinychan + 2ch Mode -