This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PING^2][PATCH] Reformat malloc to gnu style.
- From: OndÅej BÃlka <neleai at seznam dot cz>
- To: "Joseph S. Myers" <joseph at codesourcery dot com>
- Cc: Allan McRae <allan at archlinux dot org>, Siddhesh Poyarekar <siddhesh at redhat dot com>, libc-alpha at sourceware dot org
- Date: Thu, 2 Jan 2014 19:55:09 +0100
- Subject: Re: [PING^2][PATCH] Reformat malloc to gnu style.
- Authentication-results: sourceware.org; auth=none
- References: <20131211102701 dot GA25483 at domone dot podge> <20131217095418 dot GA23717 at domone dot podge> <20131230152001 dot GA5214 at domone> <52C2BF2A dot 9010009 at archlinux dot org> <20140101132020 dot GA12317 at domone> <52C4A758 dot 9070802 at archlinux dot org> <20140102034542 dot GG5374 at spoyarek dot pnq dot redhat dot com> <52C50A54 dot 5060906 at archlinux dot org> <20140102090824 dot GA18222 at domone> <Pine dot LNX dot 4 dot 64 dot 1401021616070 dot 28168 at digraph dot polyomino dot org dot uk>
On Thu, Jan 02, 2014 at 04:28:18PM +0000, Joseph S. Myers wrote:
> Several of the changes committed in commit
> 6c8dbf00f536d78b1937b5af6f57be47fd376344 are wrong - sufficiently many
> that I think it should be reverted, and then smaller patches submitted
> after the freeze each of which addresses exactly one formatting issue
> (e.g. bad indentation), everywhere it occurs (or only some places it
> occurs, if needed to keep the patch below 1000 lines). For example:
>
That is missing a point, a formatting now was wastly different than gnu
one so goal was not get perfect on first try but to converge to
solution.
In formatting patches there are three factors in play:
1) Amount of code formatted correctly.
2) Amount of code left intact.
3) Amount of code formatted incorrectly.
Now factors 2 and 3 sum to amount of code that needs futher formatting.
If a formatter gets 95% of code right then you need to fix only 5% of
code which means that patches that fix these are twenty times smaller
than in first case.
For our purposes it is mostly irrelevant if flaw was caused by formatter
or left over unless formatting causes loss of informattion. Where a style
allows only one possibility you cannot encode any information so that is
not a factor.