| 1 |
On Sunday 17 September 2006 02:40, Vlastimil Babka wrote: |
| 2 |
> Alon Bar-Lev wrote: |
| 3 |
> > Hello, |
| 4 |
> > |
| 5 |
> > Maybe you can help answering some questions I have regarding the |
| 6 |
> > generation-2 java eclass. |
| 7 |
> > |
| 8 |
> > I've seen that in openoffice ebuild there is no call to |
| 9 |
> > java-pkg-opt-2_pkg_setup fro pkg_setup, the reason is seems to be |
| 10 |
> > because it is called automatically. |
| 11 |
> > |
| 12 |
> > 1. Is there any documentation regarding portage phase hooks? I |
| 13 |
> > could not see any reference to this in eclass HOWTO guide. |
| 14 |
> |
| 15 |
> It's a new feature of portage 2.1. See |
| 16 |
> /usr/share/doc/portage-<ver>/NEWS.gz. Not mentioned in eclass HOWTO |
| 17 |
> because it's not meant to be used by eclasses, but by users. Java |
| 18 |
> eclass is a exception, but we should stop using it when portage is |
| 19 |
> able to save env properly between phases (which should already be |
| 20 |
> now IIRC?). It's bad thing that now an user can override our phase |
| 21 |
> hooks with /etc/portage/bashrc. |
| 22 |
|
| 23 |
Thanks for explaining that. |
| 24 |
|
| 25 |
> > 2. It seems a bit strange that two eclasses can override the same |
| 26 |
> > function name... How such conflict is resolved? |
| 27 |
> |
| 28 |
> Order of inheriting matters, the latter eclass overrides the |
| 29 |
> former. Ebuild inheriting conflicting eclasses should then override |
| 30 |
> the function itself and call the functions of both eclasses from |
| 31 |
> there. Now if there was a repoman check for that... |
| 32 |
|
| 33 |
Right... But since, as you said, the elcass should not use the pre |
| 34 |
stuff, it should not be a problem. |
| 35 |
|
| 36 |
> > 3. Looking at java-pkg-2.eclass I see function name |
| 37 |
> > pre_pkg-2_setup, shouldn't it be pre_pkg_setup? I see that |
| 38 |
> > pre_pkg_setup is specified in java-pkg-opt-2.eclass... Why is |
| 39 |
> > there a difference? |
| 40 |
> |
| 41 |
> Must be a typo. As a result, java env is probably not set properly |
| 42 |
> inside ebuild's pkg_setup() (for ebuilds that define it). But since |
| 43 |
> there are correct hooks for other phases (especially src_compile) |
| 44 |
> it didn't cause any harm so far. |
| 45 |
|
| 46 |
Should I open a bug for it, or you can fix it? |
| 47 |
|
| 48 |
> > 4. Can you please fix the documentation of java development and |
| 49 |
> > on the eclass it-self so that there will be a comment that the |
| 50 |
> > unlike other eclasses, java-pkg*-2_pkg_setup should not be called |
| 51 |
> > from pkg_setup? |
| 52 |
> |
| 53 |
> Well, since the usage of phase hooks is only a workaround (not |
| 54 |
> meant to free ebuild writer from calling java-pkg*-2_pkg_setup, |
| 55 |
> that's just a consequence) I would say it's better to document that |
| 56 |
> it should be called from ebuild explicitly, so when we stop using |
| 57 |
> the phase hooks (and maybe it's time already?), number of ebuilds |
| 58 |
> won't get broken instantly. |
| 59 |
|
| 60 |
Right, I think so too. |
| 61 |
So people should be told to call the pkg_setup. |
| 62 |
|
| 63 |
For example openoffice ebuild removed the call to |
| 64 |
java-pkg*-2_pkg_setup since it is already called. |
| 65 |
History is on bug#139340. |
| 66 |
|
| 67 |
Best Regards, |
| 68 |
Alon Bar-Lev. |
| 69 |
-- |
| 70 |
gentoo-java@g.o mailing list |
Archives