[icedtera-web] future of shell launchers

[Jiri Vanek]

Yup!

Pushed!

Hopefully it will make its benefits.


J.
On 1/14/19 6:24 PM, Laurent Bourgès wrote:
> Yes please go ahead.
> I am still overbusy / tired, so I miss this train !
> Will find some time very soon to look at itw launchers.
> 
> Sorry,
> Laurent
> 
> Le lun. 14 janv. 2019 à 16:45, Jiri Vanek <jvanek at redhat.com <mailto:jvanek at redhat.com>> a écrit :
> 
>     I will push this soon to head. Hopefully it will be positive change.
>     If it proves itself, it willgo ti 1.7.
> 
>     Any feedback still welcomed, another patche sin this area will flow
> 
>     J.
>     On 12/18/18 2:47 PM, Jiri Vanek wrote:
>     > Hi!
>     >
>     > There is quite minimalist change to make itw shell lunchers portable out of distribution world.
>     > I think it is feasable also for 1.7.
>     >
>     > It should be consistent iwth current windows bats, and I will align future rust ones too.
>     >
>     >
>     > J.
>     >
>     > On 12/17/18 12:03 PM, Laurent Bourgès wrote:
>     >> Hi Jiri,
>     >>
>     >> It looks very promising, but I have to figure out the transition plan.
>     >>
>     >> Would you accept to talk directly (skype?) to help going forward ?
>     >>
>     >> Laurent
>     >>
>     >> Le lun. 17 déc. 2018 à 10:52, Jiri Vanek <jvanek at redhat.com <mailto:jvanek at redhat.com>
>     <mailto:jvanek at redhat.com <mailto:jvanek at redhat.com>>> a écrit :
>     >>
>     >>
>     >>
>     >>     Hi Laurent!
>     >>
>     >>     Maybe a bit more constructive answer to you query.
>     >>     You could noticed, than in rust lunchers(patch still on review), we are not using
>     make-generated
>     >>     classpath, but instead are sent in individual jars.
>     >>     I think they are much more versatile for various changes of paths.
>     >>     usage of those will make (boot)classpaths composing in bats easier, and allows better
>     implementation
>     >>     of ITW_HOME into linux shells.
>     >>     I have very good experience with resolvig scripts dir with following snippet[1].
>     >>     So the workflow can be:
>     >>     1)adding [1] to both head and 1.7
>     >>     2)change shell launchers in 1.8 to accept parameters whcih are flowing to rust lunchers
>     >>     3)in 1.8 compose classapths honor variables in same manner as rust lucnhers do
>     >
>     > 3.5 "cross compile" launchers.
>     >>     4)adapt non breakig parts of 2 an 3  in 1.7. namely hack (sed in (boot)cps>) the usage
>     og  [1]'s
>     >>     SCRIPT_DIR and  bat's ITW_HOME
>     >>       - postpond posisble dangerous parts to 1.7.3
>     >>     In meantime, finish .args file handling everywhere. Especially in 1.8.
>     >>
>     >>     Thoughts?
>     >>
>     >>
>     >>     [1]
>     >>     ## resolve folder of this script, following all symlinks,
>     >>     ##
>     http://stackoverflow.com/questions/59895/can-a-bash-script-tell-what-directory-its-stored-in
>     >>     SCRIPT_SOURCE="${BASH_SOURCE[0]}"
>     >>     while [ -h "$SCRIPT_SOURCE" ]; do # resolve $SOURCE until the file is no longer a symlink
>     >>       SCRIPT_DIR="$( cd -P "$( dirname "$SCRIPT_SOURCE" )" && pwd )"
>     >>       SCRIPT_SOURCE="$(readlink "$SCRIPT_SOURCE")"
>     >>       # if $SOURCE was a relative symlink, we need to resolve it relative to the path where
>     the symlink
>     >>     file was located
>     >>       [[ $SCRIPT_SOURCE != /* ]] && SCRIPT_SOURCE="$SCRIPT_DIR/$SCRIPT_SOURCE"
>     >>     done
>     >>     readonly SCRIPT_DIR="$( cd -P "$( dirname "$SCRIPT_SOURCE" )" && pwd )"
>     >>     --
>     >>     Mgr. Jiri Vanek
>     >>     judovana at email.cz <mailto:judovana at email.cz> <mailto:judovana at email.cz
>     <mailto:judovana at email.cz>>
>     >>
>     >
>     >
> 
> 
>     -- 
>     Jiri Vanek
>     Senior QE engineer, OpenJDK QE lead, Mgr.
>     Red Hat Czech
>     jvanek at redhat.com <mailto:jvanek at redhat.com>    M: +420775390109
> 


-- 
Jiri Vanek
Senior QE engineer, OpenJDK QE lead, Mgr.
Red Hat Czech
jvanek at redhat.com    M: +420775390109