This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: scrambled source for breakfest?

On May 5, 2004, at 10:07 AM, Mike Stump wrote:

In gcc/gcc/doc/invoke.texi, you did:

revision 1.423
date: 2004/03/04 19:28:36; author: stuart; state: Exp; lines: +20 -59
2004-03-04 Stuart Hastings <>
* gcc/doc/invoke.texi: Document -mlongcall for Darwin/PPC.

? Surely this is wrong?

Yes, it's wrong. My screwup.

Could you please audit your changes and fix any that went astray?



Usually this can never happen. The usual method of coming up with such a change would be copying whole files around and not reviewing a cvs diff just before check in. The usual way to avoid copying or moving whole files around would be to use diff/patch to move around just the parts one is interested in.

I restored content from my patch in 1.422 to the top of the tree. I didn't restore any other content.

I don't understand exactly how I screwed this up; the "simple answer" would be that I neglected to update my copy of invoke.texi before committing, but of course CVS will prevent that. In any case, my patch unintentionally reverted three other patches:

1.422 by Mike Stump, concerning frameworks (fixed today)
1.421 by David Edelsohn (committing a change on behalf of Mostafa Hagog), concerning gcse-after-reload (patch below)
1.420 by Richard Sandiford, concerning -mfix-4400 and errata (fixed on 24mar04)

Mike reapplied his changes today.

Richard, your "Apply missed hunk" on 24-mar was correcting my screwup.

My apologies to Mike, David, and Richard.

I have a patch to restore the David/Mostafa gcse-after-reload documentation below.

For the record, here is the *bad* patch I applied, creating 1.423 (cvs diff -r1.419 -r1.423 gcc/gcc/doc/invoke.texi). I will *not* be re-applying this:

Attachment: invoke.texi.badpatch
Description: Binary data

Here is a patch to restore the gcse-after-reload documentation added by David on behalf of Mostafa in revision 1.421:

Attachment: invoke.texi.for-dje
Description: Binary data

Here is my original patch for -mlongcall; this is what I was originally trying to apply on 04mar2004:

Attachment: stuarts.invoke.texi.patch
Description: Binary data

I'll commit the gcse-after-reload restoration patch later today unless someone objects. I will make sure to update my sources first. I expect some people will be checking up on me...

2004-05-05 Stuart Hastings <>
* gcc/doc/invoke.texi: Restore -fgcse-after-reload doc from 1.421,
mistakenly clobbered by 1.423.

stuart hastings
Apple Computer

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]