Re: bulky-but-brainless updates 2001.05.26

Jonathan Hogg <[email protected]>
Newsgroups gmane.comp.sysutils.ark.devel
Message-ID <[email protected]>
On 26/5/01 23:05, Will Partain wrote:

>A side-effect of this change is that your .xmla
>("pre-compiled .xml") files are no good any more.  Just do
>
>  find . -name '*.xmla' -print | xargs /bin/rm
>
>and carry on as normal.  (The files will be remade to the
>correct thing.)

Ha! I hit this. [I've been shadowing you all evening Will ;-)]

But I found a nasty problem. I sucked down the new ark code. Noticed the 
problem with the new pickle format, so I ditched all my xmla's and ran:

     ark package reveal --redo arkbase

To update to the new code. However, this rebuilds a bunch of the xmla's 
with the old code, then switches part way through the process to the new 
code, which tries to read them in and barfs.

I bashed in a quick patch to the code to catch this, but I've not 
committed it yet as I noticed you were pretty busy. [Basically, all I've 
done is make it unpack the triple in the exception block before returning 
it recomposed. If the xmla file is the previous version then an unpack 
exception triggers the "ignore this xmla" code.]

I'm hitting a new bug with the rhlinux changes that I can't track down 
though. An attempt to reveal a package is complaining about "no such 
thing: .../ia32-rhlinux.xml", but I'm sure (find&grep sure) that there 
aren't any references to rhlinux in my team directory. It seems to barf 
during dependency checking/evaluation. Ideas?

On an unrelated note: I also made some local changes to only show 
exception tracebacks if the verbose flag has been set. Otherwise it just 
prints the exception and dies. If this is OK with you then I'll submit 
it. Feel free to bounce it though. I was finding that I make mistakes all 
the time that cause a thousand lines of traceback to fly past obscuring 
the thing that actually caused the error ;-)

On a brighter note: I'm getting fewer tracebacks now as I slowly move 
towards having a working system. I'll share my team directory once I'm 
closer to being done. One thing I'm having difficulty with is that I've 
been basing my work on the glasli1 team, but I've discovered that the 
glasli1 team is no longer bootable with the extensive changes to sidai 
over the last few months.

I'm also having some difficulty getting my head around sidai. I like the 
policies generally, but sometimes I can't figure out what the policy is 
to decide if I like it ;-)

For instance: how is ark-diffing-deploy meant to work? Once I've looked 
at the diffs and satisfied myself that it's what I want to do. How do I 
push them through? [Takes me back to OUR days...]

All looking good.

:-j


-- 
jonathan hogg, one good idea ltd, 131 queen margaret dr., glasgow g20 8pd
http://www.onegoodidea.com/ tel:+44-(0)7976-614338 fax:+44-(0)7970-537451
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.