Re: What happened to tank ?

[ Thread Index | Date Index | More Archives ]


> > I think it show that new tools can be dangerous for the host and are
> > not ready for production. Did you see all trouble we had and we
> > still did'n build anything. I belive in new tools, but there seems
> > not inuf mature to used them on Tank.
> I can't accept to endorse problem I didn't create ! When yo run rm -r
> on a dir with mounted dir into it, it erases all. rm -r **IS**
> dangerous, it's in all GNU/Linux manuals, and everybody using shell
> experimented this at least one time... We created this sandbox
> manually and we don't have any clue about why it stopped to works. My
> scripts where not used to manage it, so how can you say it's their
> fault ???

I dont put fault on them, I say with the new way to do things we can
get-into BIG trouble (after all we did I forgot we mount /home/slitaz
into the sandbox. And I'm not sure rm -r did the bug because only a few
repos was removed, but not www/, strange.

That said, now I recreated a sandbox and only
monted /home/slitaz/cooking into it, not the all slitaz work directory.

For other devs, on Tank we use a sandbox because Tank runs stable, in
the future or if you use cooking, you just have to install
tazwok-experimental and cook package like usuall but there will be
build in a chroot. In short:

# mkdir /home/slitaz/cooking && cd /home/slitaz/cooking
# hg clone && cd -
# tazwok chroot
# tazwok cook pkg

Am I right GoKhlaYeh ?

> > Now I will try to make a chroot as before (tazdev gen-chroot) and
> > use experimental to build package but without bootstraping the
> > toolchain and without any sandbox. Or you might have an other
> > suggestion ?
> Yes. We success to create this sandbox and run tazwok cook-toolchain
> into it two days ago. A bug in the last release prevented to run the
> cook-toolchain until end, so I fixed it and as I tell you I builded a
> toolchain successfully with this fix; and now you want to let fall
> all we did to make it works and use it in a way nobody tried... Are
> you gonna to tell it's my bad if it didn't works then ? (ok, it's
> ironic, I'm sure that no)
> So my suggestion is getting back this sandbox, cook the toolchain as
> planned, cook packages we need to make iso and release them, then cook
> the whole repository :)

It's what I did this this mornig. Recreate a clean sandbox manually an
then tried to build the toolchain. Actually it's the fourth try of the
day. I add trouble with linux-api-header and disconnection due to
packet lang... At the moment glibc is building

In day Imust say something was missing me, an UPDATED tazwok manual!!!
This is the first thing on should do we adding commands to a tool, add
the coresponding information into the manual, so one can RTFM.

I'm persevering using your tools... so please be kind! And I realy
think ther improve slitaz but are not yet ready for production.

AND, we can't just build iso packages, release 4.0 ISO's and then finish
build. People must be able to test all packages before we release, it's
why wok is froozen 2-3 mounth before stable (I saure update in wok

So everyone: STOP UPDATING packages in wok and focus you energy on fixes
and writing doc.

> GoKhlaYeh <gokhlayeh@xxxxxxxxxx>

- Christophe

SliTaz GNU/Linux Mailing list -

Mail converted by MHonArc 2.6.19+