>>>>> "Jim" == James Carlson <james.d.carlson at Sun.COM> writes:

Jim> nightly.1

Jim>   277: s/which/that/ (It's a restrictive relative clause ... and
Jim>   one of my pet peeves.  :-/)

Will fix.

Jim> nightly.sh

Jim>   279,2326: no need to escape end-of-line when line ends with pipe
Jim>   character.  (Thanks for removing the {} swill, though.)

Will fix.

Jim>   285: this is a little odd.  We go out of our way to support $DEST
Jim>   with embedded spaces at line 283, but then fall over here.  It'd
Jim>   be good to be more consistent.

Oops.  Will fix (and the cd $WS, too).

Jim>   344: can this mkdir fail?

Yes, for things like ENOSPC.  $dest was already created by
cpio.

In general, nightly.sh doesn't deal well with environmental errors like
ENOSPC.  But I'll look at putting in some error checking for
copy_source_hg.

Jim>   346,350: should these ";" actually be "&&" instead?  What happens
Jim>   if the "cd" fails?

I'll have to look at this some more.

Jim>   513: do any of the other make default rules do inadvertent things
Jim>   with svn or hg?  (Any chance of inadvertent SCM invocation during
Jim>   the build?)

Not that I know of.  It's highly unlikely, at least for Mercurial,
because the rule would have to depend on the location of the repo's
root.  

The strings "svn" and "hg" are absent from /usr/share/lib/make/*.

Jim>   515,516: I'd put the >> at the end of the first line and omit
Jim>   "\".

Okay.

Jim>   2230: as long as you're here, you might use "function" (?).
Jim>   (Actually, I'm not too wild about the ksh "function" feature, but
Jim>   you used it at line 313 for new code.)

Okay.

Jim>   2327,2329: might be clearer (less duplication) to pipe 'fi' at
Jim>   2330 through tee.

Okay.

Jim>   2331: does "" do something here?

I've had problems in the past where 'echo' didn't produce the spacer
line that I wanted.  I forget the details.  But just plain 'echo' does
seem to work now, so I'll remove the empty string.

Jim>   2491: I had thought that the SCCSHELPER thing was about files
Jim>   that were _modified_ during the "source delivery" build, not just
Jim>   removed.  But it's been a long time since I've done such a build,
Jim>   so I might not be remembering this right.

Perhaps it was used that way sometime in the past.  But the source build
is currently set up to copy the source tree and mung the copy.  I
believe it's been that way for the past couple years at least (i.e., as
long as I've been paying attention to the source build).

Thanks for the comments.

mike

Reply via email to