Bug #4457
closed
we apparently change .comment of almost every userland object
100%
Description
In investigating another issue, I'me stumbled into a case where it appears that we use mcs to adjust the .comment of every userland object (that is, .o, not what we ship).
cmd/mkfifo/mkfifo.o is changed
lib/libctf/i386/pics/ctf_hash.o is changed
cmd/dtrace/i386/dtrace.o is ok
cmd/iscsiadm/cmdparse.o is changed
cmd/zpool/zpool_iter.o is changed
Fixing the comment section of objects is pointless. We stomp the .comment of the final build product anyway, fixing up all these objects is just wasting time (and causing me a great deal of separate confusion).
Related issues
Updated by Josef Sipek over 9 years ago
- Assignee set to Josef Sipek
- % Done changed from 0 to 20
Updated by Josef Sipek over 8 years ago
As a data point, I used the following DTrace script during a nightly build to see how many invocations to mcs occured before and after:
proc:::exec
/args[0]=="/usr/bin/mcs"/
{
@[args[0]]=count();
}
tick-10sec
{
printa(@)
}
Before, there were 23895 invocations. After, there were only 6032.
Updated by Electric Monk over 8 years ago
- Status changed from New to Closed
- % Done changed from 20 to 100
git commit f3477bb2505a35e160b09c9dbf5476112be4fa7d
commit f3477bb2505a35e160b09c9dbf5476112be4fa7d
Author: Josef 'Jeff' Sipek <josef.sipek@nexenta.com>
Date: 2014-11-06T15:39:04.000Z
4457 we apparently change .comment of almost every userland object
Reviewed by: Keith Wesolowski <keith.wesolowski@joyent.com>
Reviewed by: Alex Reece <alex@delphix.com>
Reviewed by: Richard Lowe <richlowe@richlowe.net>
Approved by: Garrett D'Amore <garrett@damore.org>