> it did. It was because we didn't had time to catch-up since argentina
> trip. up until you came back at january.
> you, of course, can explain it as my lazyness or whatever.. but the
> fact is that i have no idea what you did..

thats why I write mails. The mails here are not just random blabbering, they
are communication for the people that are not physically present. 

Now I hope you understand the value of incremental mails and even more of 
reading mails...

> and how to use it.
> having a documentation, is not the same as seeing the whole picture.
> i want to understand, how it works and why it works like that, not how
> to use it.

If you know how it works you're done as a user of the NativeBoost job. 
Everything else
is mostly interesting for the ci/jenkins administrator.

The main motivation is to NOT use jenkins. Jenkins is bad software, it does 
it's job but
setting it up is not that nice. I wrote all these scripts that we can simply 
write a single
bash file which does everything.

My use-case is always: What do I have to do, to reproduce this locally?

Before: 
- check that I copy the scripts-download
- check that I download the right VM from jenkins (I cannot run it locally 
since the urls were relative)
- check that I us the right scripts from pharo-build
- check that I put the images in the right location otherwise build.sh won't 
work
- pray that everything works

Now:
- copy the jenkins job bash into a file
- execute it
- pray that it works

> like following:
> 
> ciCog.sh              
> ciCogVM.sh            
> ciNBCog.sh            
> ciNBCogVM.sh          
> ciPharo13.sh          
> ciPharo14.sh          
> ciPharo14Cog.sh               
> ciPharo14CogVM.sh             
> ciPharo20.sh          
> ciPharo20Cog.sh               
> ciPharo20CogVM.sh             
> ciPharo20NBCog.sh             
> ciPharo20NBCogVM.sh           
> ciPharo20PharoVM.sh           
> ciPharo20Rizel.sh             
> ciPharo20RizelVM.sh           
> ciPharoVM.sh          
> ciRizel.sh            
> ciRizelVM.sh
> 
> 
> this list burns my eyes.

Initially the scripts did not have the VM suffix, stupid me. So I added it, but 
I could not
change all the jenkins jobs since my internet connection was very bad in 
argentina.

=> we can remove all the non -VM scripts there
The rest are just combinations of the image and the VMs

> and having same problem,as Marcus asks about VMs: which one i have to use?
> Is there a way to have less? just wonder.

Yes, if you make the PharoVM compile we only need 2/3 of them...

- PharoVM the default VM including "everything"
- StackVM the debug vm, in case we have issues with the JIT
- RizelVM for profiling


What can we do to simplify this?
--------------------------------
=> rename them
- PharoVM is fine
- StackVM => Pharo-Debug-VM
- RizelVM => Pharo-Profiler-VM

like that nobody has to decode the name of the VM's anymore...

Reply via email to