gnu@toad.com (03/15/89)
GCC should not put extraneous differences into files that would otherwise be identical. If you upgrade from gcc-1.xx to gcc-1.yy and run a regression test, only the files whose assembler code actually changes should show up as different. This is a major botch in "coff", where they put a timestamp in the object files, and you have to take special precautions to try to compare two versions of object code, e.g. to see if a compiler fix has changed anything. John Gilmore
rfg@MCC.COM (Ron Guilmette) (03/15/89)
Dear John, This has been mentioned before to the gcc bug mailing lists. Somebody pointed out that for COFF files, you can get them to compare the same if you skip over the first few bytes of both files before starting the compare. I've forgotten now what the suggested way of doing that was. Anybody care to refresh my memory? // Ron Guilmette - MCC - Experimental (parallel) Systems Kit Project // 3500 West Balcones Center Drive, Austin, TX 78759 - (512)338-3740 // ARPA: rfg@mcc.com // UUCP: {rutgers,uunet,gatech,ames,pyramid}!cs.utexas.edu!pp!rfg
drh@notecnirp.Princeton.EDU (Dave Hanson) (03/15/89)
In article <8903142103.AA18095@riunite.aca.mcc.com> rfg@MCC.COM (Ron Guilmette) writes:
Dear John,
This has been mentioned before to the gcc bug mailing lists.
Somebody pointed out that for COFF files, you can get them
to compare the same if you skip over the first few bytes of
both files before starting the compare. I've forgotten now
what the suggested way of doing that was. Anybody care to
refresh my memory?
strip them, then skip the first 8 bytes.
nevertheless, time/version stamping .o files
is a nuisance as regression testing on a mips
machine will reveal. i hope gcc doesn't do it.
weening@Gang-of-Four.Stanford.EDU (Joe Weening) (03/15/89)
The assembler creates .o files, not GCC. I can't think of any change one can make to GCC that will affect whether .o files are time-stamped. -- Joe Weening Computer Science Dept. weening@Gang-of-Four.Stanford.EDU Stanford University
drh@notecnirp.Princeton.EDU (Dave Hanson) (03/15/89)
In article <WEENING.89Mar14212528@Gang-of-Four.Stanford.EDU> weening@Gang-of-Four.Stanford.EDU (Joe Weening) writes: The assembler creates .o files, not GCC. I can't think of any change one can make to GCC that will affect whether .o files are time-stamped. .asciz "gcc version 1.34.0/Wed Mar 15 07:49:48 EST 1989" stamps the .o file regardless of object format.