diff options
Diffstat (limited to 'zarb-ml/mageia-dev/2012-July/017269.html')
| -rw-r--r-- | zarb-ml/mageia-dev/2012-July/017269.html | 130 |
1 files changed, 130 insertions, 0 deletions
diff --git a/zarb-ml/mageia-dev/2012-July/017269.html b/zarb-ml/mageia-dev/2012-July/017269.html new file mode 100644 index 000000000..3c955dae1 --- /dev/null +++ b/zarb-ml/mageia-dev/2012-July/017269.html @@ -0,0 +1,130 @@ +<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> +<HTML> + <HEAD> + <TITLE> [Mageia-dev] Planning the /usr move + </TITLE> + <LINK REL="Index" HREF="index.html" > + <LINK REL="made" HREF="mailto:mageia-dev%40mageia.org?Subject=Re%3A%20%5BMageia-dev%5D%20Planning%20the%20/usr%20move&In-Reply-To=%3C4FFD792B.1090504%40colin.guthr.ie%3E"> + <META NAME="robots" CONTENT="index,nofollow"> + <META http-equiv="Content-Type" content="text/html; charset=us-ascii"> + <LINK REL="Previous" HREF="017268.html"> + <LINK REL="Next" HREF="017270.html"> + </HEAD> + <BODY BGCOLOR="#ffffff"> + <H1>[Mageia-dev] Planning the /usr move</H1> + <B>Colin Guthrie</B> + <A HREF="mailto:mageia-dev%40mageia.org?Subject=Re%3A%20%5BMageia-dev%5D%20Planning%20the%20/usr%20move&In-Reply-To=%3C4FFD792B.1090504%40colin.guthr.ie%3E" + TITLE="[Mageia-dev] Planning the /usr move">mageia at colin.guthr.ie + </A><BR> + <I>Wed Jul 11 15:01:31 CEST 2012</I> + <P><UL> + <LI>Previous message: <A HREF="017268.html">[Mageia-dev] Planning the /usr move +</A></li> + <LI>Next message: <A HREF="017270.html">[Mageia-dev] Planning the /usr move +</A></li> + <LI> <B>Messages sorted by:</B> + <a href="date.html#17269">[ date ]</a> + <a href="thread.html#17269">[ thread ]</a> + <a href="subject.html#17269">[ subject ]</a> + <a href="author.html#17269">[ author ]</a> + </LI> + </UL> + <HR> +<!--beginarticle--> +<PRE>'Twas brillig, and Olav Vitters at 11/07/12 13:39 did gyre and gimble: +><i> On Wed, Jul 11, 2012 at 12:52:17PM +0100, Colin Guthrie wrote: +</I>>><i> 2. Add a patch to RPM that introduces a new check that must pass before +</I>>><i> a given RPM can be installed. +</I>>><i> 3. Create a new "filesystem" build that uses this check. +</I>><i> [..] +</I>>><i> Several other packages will be subsequently broken (some packages ship +</I>>><i> their binaries in /bin but symlink them to /usr/bin but when /bin is +</I>>><i> itself a symlink to /usr/bin, they package ultimately conflicts with +</I>>><i> itself!). We need to identify such packages and fix them and have them +</I>>><i> ready to go. In order to do the transition correctly, we may need to fix +</I>>><i> them first, build them and then wait until all such packages are fixed, +</I>>><i> THEN update the filesystem rpm and then rebuild all such packages with a +</I>>><i> dep on the filesystem > x package. This might be needed to avoid any +</I>>><i> problems on the build system chroots. +</I>><i> +</I>><i> Shouldn't those packages also rely on the new RPM check? Or you mean +</I>><i> just update those packages and don't allow for symlinks between /bin and +</I>><i> /usr/bin (+same for sbin)? +</I> +They don't need to. Provided they "Require(pre): filesystem >= foo" and +rely on the check being in filesystem rpm packge. That prevents too many +"quirks" being in too many packages (the error message would also be +more confusing for users but a simple "I need a newer filesystem +package" is a bit clearer. + +><i> Would the new RPM do a runtime check to see if things have been +</I>><i> migrated? +</I> +Yes, this is what the patch does, but as mentioned it would only be the +"filesystem" rpm that would actually implement this check. + +><i> Does Dracut usrmove handle the packages which at the moment contain +</I>><i> symlinks (e.g. /bin/foo -> /usr/bin/foo and e.g. /usr/bin/bar -> +</I>><i> /bin/bar)? Or should either dracut or the packages have conflicts to +</I>><i> deal with this? +</I> +The code in dracut deals with this currently. It will overwrite symlinks +with the real binary instead. + +The code is actually quite simple: + +/usr/lib/dracut/modules.d/30convertfs/convertfs.sh + + +There may be some junk left behind when packages provide separate +binaries of the same name in /bin vs /usr/bin (and sbin), but this is +relatively rare (although I will mention this more in my next message). + +Col + + + + +-- + +Colin Guthrie +colin(at)mageia.org +<A HREF="http://colin.guthr.ie/">http://colin.guthr.ie/</A> + +Day Job: + Tribalogic Limited <A HREF="http://www.tribalogic.net/">http://www.tribalogic.net/</A> +Open Source: + Mageia Contributor <A HREF="http://www.mageia.org/">http://www.mageia.org/</A> + PulseAudio Hacker <A HREF="http://www.pulseaudio.org/">http://www.pulseaudio.org/</A> + Trac Hacker <A HREF="http://trac.edgewall.org/">http://trac.edgewall.org/</A> + + +</PRE> + + + + + + + + +<!--endarticle--> + <HR> + <P><UL> + <!--threads--> + <LI>Previous message: <A HREF="017268.html">[Mageia-dev] Planning the /usr move +</A></li> + <LI>Next message: <A HREF="017270.html">[Mageia-dev] Planning the /usr move +</A></li> + <LI> <B>Messages sorted by:</B> + <a href="date.html#17269">[ date ]</a> + <a href="thread.html#17269">[ thread ]</a> + <a href="subject.html#17269">[ subject ]</a> + <a href="author.html#17269">[ author ]</a> + </LI> + </UL> + +<hr> +<a href="https://www.mageia.org/mailman/listinfo/mageia-dev">More information about the Mageia-dev +mailing list</a><br> +</body></html> |