From: Sternbach, W. [IT] <wil...@ci...> - 2003-04-04 17:57:51
|
Hello, I wanted to offer a suggestion to change the default compiler option from -O0 (No optimization) to -O3 (Maximum optimization). The reason is that alot of the posts to this group which I have read show that people are not using any optimization options, and because it defaults to -O0, their EXE's are not benefiting from the outstanding Gnu back end optimizations. As a result, most people using GCC are getting horribly sloppy code generated, instead of a finely tuned optimized EXE which runs much faster. Its only a suggestion. What do the members of this group think of this suggestion to make full optimization (-O3) the default compiler option. - Bill -----Original Message----- From: Sean Young [mailto:se...@me...] Sent: Friday, April 04, 2003 10:35 AM To: min...@li... Subject: [Mingw-users] gcc/w32api issues. All, Some of these questions are more gcc related, but I'm hoping this is the right place to ask. We've been porting a database replicator to mingw (it already compiled with gcc on *nix and msvc on Windows). So far, we've come across the following (minor) issues: 1) C:\>gcc -o x x.c -E Output ends up in `x.exe', not `x'. Is that supposed to happen? 2) When compiling an program that uses Oracle OCI and -Wshadow, I get: D:\data\jsycu\src\jd>gmake jig_or90.dll gcc -g3 -Wall -Werror -Wshadow -Wpointer-arith -Wmissing-declarations -Wno-char-subscripts -I../../src -c -Id:/oracle/ora920/oci/include -DJP_DBMS_ORA=90 -DJP_DBMS_ORA_DYN jdorc.c -o ora90/jdorc.po cc1.exe: warnings being treated as errors In file included from d:/oracle/ora920/oci/include/oci.h:2150, from ../../src/jp/jpora.h:14, from jdorc.c:7: d:/oracle/ora920/oci/include/orl.h:3235: warning: declaration of `index' shadows a global declaration <built-in>:0: warning: shadowed declaration is here d:/oracle/ora920/oci/include/orl.h:3318: warning: declaration of `index' shadows a global declaration <built-in>:0: warning: shadowed declaration is here -snip- Now one could argue that Oracle shouldn't define variables with a general name like `index', but the same can be said about gcc. What is `index' for? Wouldn't __builtin_index be a better name? The same happens on *nix with gcc-3.2. 3) For SQL Server, we need odbcss.h and odbcbcp.lib for bulk copy. gcc (rightfully) doesn't like the msvc or SQL Server includes, so they had to be re-written for w32api. When I've finished polishing them (where) shall I submit them? I've posted a version of odbcss.h and odbcbcp.def at mingw-patches. Is there any reason they haven't been included yet? Sean ------------------------------------------------------- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ _______________________________________________ MinGW-users mailing list Min...@li... You may change your MinGW Account Options or unsubscribe at: https://lists.sourceforge.net/lists/listinfo/mingw-users |