Thursday 20 March 2008

News from other worlds

I was recently forced to face once more the limitations and quirks of rpm. This is more painful, when looking at a more capable system like .deb.

When building a .deb package, the debian/rules file is pretty much the heart of all things happening, with debian/control and debian/changelog as the lungs and the liver(? maybe) .

Having different files for the maintainer scripts makes possible things like automatic (but repetitive) generation/modifications of those scripts. This can be useful in some situations when the postinst should do stuff depending on the current contents of the released upstream source (sorry, I know I am being vague, but I can't detail too much). The source package is then generated and the resulting package can be rebuilt 1000 times and the resulting postinst will not be different for a given release. But this is always done in the same way, no matter the upstream version.

Fatality one - the spec is fixed, saint, but still can be altered... still, not in due time

So the basic idea is to have postinst.in as a source for postinst, which is created during the configure stage and later removed on clean. This is possible and works on deb.

Try the same on rpm and you will fail. This is because, although the spec file supports file inclusions, the altered files are merged into the spec right at the beginning of the build process, and you end up with no or an outdated postinst. A later rebuild of the same package would do the trick, but that is wrong for more than one reason.

Fatality two - why fatality one was met

All of this is made in order to overcome yet another rpm limitation: rpm can't assist too much on upgrades; you can't simply upgrade packages properly from one version to another, because rpm doesn't have the notion of letting the scripts know which version is installed over which; it only tells that it is an upgrade or not.

This has lead some people believe rpm isn't able to upgrade at all. Glad to tell them they're wrong (still rpm sucks and is brain dead).

Fatality three - why rpm based system fail to have upgrade support today

Want another RPM limitation? RPM has this brain dead idea that on upgrade the "proper" order and way to call the maintainer script(lets) is:
  • new-preinst 2
  • new-postinst 2
  • old-prerm 1
  • old-postrm 1
The numbers 1 and 2 are literally what is passed to the scripts and mean "the number of the installed versions of the same package that will exist after the current operation finishes". No versions of any kind, no other parameters, no rollback, no nothing, just 1 or 2.

And by letting the scripts of the old version run AFTER the scripts of the new one means that you have to explicitly prepare the packages for a later upgrade support (and the best idea is to do NOTHING in the old prerm and postrm scripts on upgrades, since nobody is clairvoyant to see what would be the right course of action for upgrade in future releases).

More bad ideas from the rpm land (glad not to be there)

In somewhat related news, the fedora people are discussing now around the rejection of being able to have package names contain non-ascii characters, characters like kanji, kanas, cyrillic characters or any other characters. Even if I am pro-l10n, this wreaks of bad idea scent from a mile afar.

They aren't probably aware that even unicode is broken (if you see the same glyphs in the comparison table, that is proof unicode is broken) but that is their decision and I am glad I am not in that lost land.

Saturday 8 March 2008

mistranslations - how to deal with it

Aigars, you could use what I already made for the compedia creation. The code is available on i18n.debian.net in my home dir (I thought I exposed it on alioth, too, but I just checked and is not).

The results for latvian are also present.

Wednesday 5 March 2008

email harvesting prevention - simple and efficient

On this page I've seen this nice way of preventing email addresses harvesting (I have noscripts installed):



When javascript is enabled/permited, it looks like this:


The code is really simple and could be easily automatically applied to pages (the code obfuscates even more, but that is irrelevant - see the original source for the code):
<script language='JavaScript' type='text/javascript'>
<!--
var prefix = 'ma' + 'il' + 'to';
var path = 'hr' + 'ef' + '=';
var addy35780 = 'g.martinez' + '@';
addy35780 = addy35780 + 'pcbsd' + '.' + 'es';
document.write( '<a ' + path + '\'' + prefix + ':' + addy35780 + '\'>' );
document.write( addy35780 );
document.write( '<\/a>' );
//-->\n </script><script language='JavaScript' type='text/javascript'>
<!--
document.write( '<span style=\'display: none;\'>' );
//-->
</script>This e-mail address is being protected from spam bots, you need JavaScript enabled to view it
<script language='JavaScript' type='text/javascript'>
<!--
document.write( '</' );
document.write( 'span>' );
//-->
</script>


BTW, the PC-BSD site looks really nice, unlike Debian's.

Debian Games Team - have source, but do not import :-/

JoeyH wrote a while back some thoughts on why he hates the DGT SVN repo (mainly our policy to import just the incomplete source of packages - usually just the debian/ directory).

He updated that blog entry and latter said:

Update: I realized after posting this that while the space gains on alioth might be illusory, there is certianly a space gain for people checking the whole repo out. What's really needed is a way to keep all the upstream sources in svn, but check them out only when wanted, and check out all the debian directories for cross-game work.

I was thinking, how about having a different area of the repo where the full source is present and which pulls the debian directory as an external from the current location?

I know it has the huge disadvantage that anything outside of the debian/ directory, but part of the diff, is, for sure, lost. (Still, there is also another team policy to not track anything outside debian directly, except through patches[1], except for really exceptional cases, but even then, there is a need for a strong justification).


Would that work? (I know git-svn doesn't support externals properly, yet, so there's not much gain).


[1] I know Joey, you hate those, too

Sunday 2 March 2008

pbuilder - /etc/pbuilderrc no longer a conffile (333294)

I have just finished a completely functional patch for pbuilder's bug #333294. I started this during debconf7, to be more precise, just after Junichi's talk about pbuilder. This patch makes pbuilder smarter smarter about the information it places in the MIRRORSITE variable of the /etc/pbuilderrc.

And now that file is no longer a conffile[1], and it asks some questions via debconf. And since debconf is translatable, there is a Romanian translation, too.

I have published my changes in my pbuilder git repo on alioth.

Anyone interested in the changes can pull from that repo:

git://git.debian.org/git/users/eddyp-guest/pbuilder.git
or
http://git.debian.org/git/users/eddyp-guest/pbuilder.git


[1] this was the main reason the patch wasn't good enough, since modifying a conffile is RC.

Saturday 1 March 2008

TDebs - arch is not important, is it?

Updates:
  1. the arch issue raised by Frans was not the object of my post. Also, not all points are invalid/reiterated, just the ones iterated here. Thanks Frans.
  2. add a note that there is an image that proves that .mo endianness is irelevant - usefull for readers that might miss that



While looking at Neil Williams' talk(hi-res/lo-res) from FOSDEM about cross building and tdebs I was STUNNED about the many (not all, thanks Frans) points that were reiterated during that talk when there were already answered.

Well, I guess you have to repeat yourself a lot to get your ideas through to the other side. So, let's do that.

All (except the first) of the following have been already answered one way or another on the TDebs wiki page.

Q: .mo files are arch independent or not? Neil asks again on his blog.
A: It doesn't matter, does it?


0 eddy@bounty ~ $ file /usr/share/locale/ro/LC_MESSAGES/wormux.mo
/usr/share/locale/ro/LC_MESSAGES/wormux.mo: GNU message catalog (big endian), revision 0, 228 messages
0 eddy@bounty ~ $ uname -a
Linux bounty 2.6.24.2-bounty #1 SMP Wed Feb 13 02:34:09 EET 2008 x86_64 GNU/Linux
(there is a picture here which proves my point, if you can't see it, click here).



Q: The packages file will explode
A: No, that's addressed.

Q: How do you select languages?
A: Simply use what we have and expand on that.

Q: Do tdebs support anything else than gettext?
A: Yes.

Building in a sarge chroot might not always work

Adeodato correctly points out that building under a Sarge chroot, and expecting the package to work on Sid, might not always[1][2] work mainly because of ABI changes and similar changes. Thus, I am forced to clarify this a little.


Generally, the binaries in the packages we produce do not dynamically link to anything other than libc6, libstdc++5 (yes, 5, not 6) or some internal libraries, so we don't have to worry about ABI changes, at least, not yet.


So, a disclaimer is in order: unless you know what you're doing, you probably are better off building packages for each of the supported Debian distros.

Thanks Adeodato for the heads-up.

[1] unless your package is a special case
[2] or, better said, in most cases


udate: fixed the link to dato's post