[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gpsd-dev] Working towards 3.10 release
From: |
Eric S. Raymond |
Subject: |
Re: [gpsd-dev] Working towards 3.10 release |
Date: |
Fri, 8 Nov 2013 16:41:33 -0500 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
Hal Murray <address@hidden>:
> scons -c is still leaving a lot of stuff that I expect to get deleted.
>
> after scons -c:
> address@hidden gpsd2]$ ll -R .sc*
> -rw-rw-r-- 1 murray murray 0 Nov 7 20:36 .scons-option-cache
> -rw-rw-r-- 1 murray murray 153360 Nov 7 20:36 .sconsign.dblite
>
> .sconf_temp:
> total 296
> -rw-rw-r-- 1 murray murray 79 Nov 6 09:22 conftest_0.c
> -rw-rw-r-- 1 murray murray 924 Nov 6 09:22 conftest_0.o
> -rwxrwxr-x 1 murray murray 7309 Nov 6 09:22 conftest_0
> -rw-rw-r-- 1 murray murray 79 Nov 6 09:22 conftest_1.c
> -rw-rw-r-- 1 murray murray 924 Nov 6 09:22 conftest_1.o
> ....
Those are scons state files. If I try to delete them from with scons,
various postconditions fail and havoc ensues. (One version of the
havoc was the carnivorous config bug.) Sorry, I did try.
> It also misses www/*.html
I've sort of fixed that. There's now a 'distclean' target that does an
ordinary clean, then removes lingering test binaries, then a
web-directory clean. The only things it leaves in place are the
scons state files.
> I've got two different copies of www/cycle.png
> -rw-rw-r-- 1 murray murray 12690 Oct 31 16:12 gpsd/www/cycle.png
> -rw-rw-r-- 1 murray murray 11626 Nov 2 22:58 gpsd2/www/cycle.png
>
> I've never run scons in gpsd. It is just a clean reference to see what scons
> -c is leaving behind.
>
> The newer one looks better. The text for RTIME is mangled in the older one.
> The older one came from git.
> I see a www/cycle.dia and I had to install dia to run scons www so I assume
> the newer one came from scons www as run on my system. (The time stamps
> match the html files.)
>
> I don't know what you are trying to do. I'd expect either cycle.png wouldn't
> be in git or it wouldn't get created by scons www. The latter would be
> handy for users like me so dia isn't required but it makes life harder for
> you to release a new cycle.dia.
It's the result of a mess that's too tedious to explain. I'll clean it up,
probably by geetting rid of the diagram.
--
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>