| 1 |
* William L. Thomson Jr. <wltjr@g.o> schrieb: |
| 2 |
|
| 3 |
Hi, |
| 4 |
|
| 5 |
> > Great :) |
| 6 |
> > It was really, really ugly getting tomcat emerge'd w/ all this |
| 7 |
> > commercial crap :( |
| 8 |
> |
| 9 |
> Blame upstream for using them. Granted the are somewhat optional |
| 10 |
> more on that below. |
| 11 |
|
| 12 |
Well, they told me, I won't need them ... who the hell's right ? |
| 13 |
|
| 14 |
> > I had some talks w/ tomcat folks |
| 15 |
> |
| 16 |
> Where? Who? Just curious. |
| 17 |
|
| 18 |
users@×××××××××××××.org |
| 19 |
|
| 20 |
> There are ant download targets for most of the deps, we just |
| 21 |
> bypass. So it's not like the deps come from no where in a normal |
| 22 |
> build from source. |
| 23 |
|
| 24 |
In other words: tomcat ships lots of bundled third-party packages, |
| 25 |
while gentoo takes them from the upstream ? |
| 26 |
|
| 27 |
But then I wonder: if the tomcat folks can simply bundle them in, |
| 28 |
why does gentoo has to suffer from these download restrictions ? |
| 29 |
|
| 30 |
> > - they were some bit confused |
| 31 |
> > about the huge dependencies @gentoo and suggested using some |
| 32 |
> > of their (monolithic) packages instead. |
| 33 |
> |
| 34 |
> I have tried to slim it down, but dropping certain deps to seem to |
| 35 |
> change things that are activated or not within Tomcat. For example |
| 36 |
> there seems to be some issues with the java5 USE flag. Either that, |
| 37 |
> or some of the dropped deps when that flag is used is effecting the |
| 38 |
> resulting catalina.jar. Who knows what other jars would be effected |
| 39 |
> by dropping other deps. |
| 40 |
|
| 41 |
hmm, we should do some deeper investigations @ dev@tomcat ... |
| 42 |
|
| 43 |
<snip> |
| 44 |
|
| 45 |
> Now consider this for a moment. Upstream is likely using binaries |
| 46 |
> to build their binary version of Tomcat ;) So it's not like they |
| 47 |
> are compiling all the stuff they are using to build Tomcat from |
| 48 |
> source. Before they compile and ship Tomcat. We do it all from |
| 49 |
> source. :) Which can explain extended deps right there. |
| 50 |
|
| 51 |
Oh hell ... |
| 52 |
|
| 53 |
<snip> |
| 54 |
|
| 55 |
> The only thing I can say about the whole situation is they are working |
| 56 |
> on releasing Tomcat 6.0.x in the next month or two. Tomcat 6.0.x has way |
| 57 |
> less deps, and it's much clearer that the ones it has are needed. No |
| 58 |
> core optional, or optional stuff. I might see about unmasking and |
| 59 |
> versioning either alpha/beta the ebuild despite sources not being |
| 60 |
> tagged :( So people can test it out and hopefully help get it released |
| 61 |
> ASAP. |
| 62 |
> |
| 63 |
> As for 5.5.x, it's not clear if there will be further releases from |
| 64 |
> upstream. It's got lots of legacy code and likely deps. |
| 65 |
|
| 66 |
Okay, since I'm not using it in production yet, I'll try the masked 6.0.6. |
| 67 |
|
| 68 |
BTW: |
| 69 |
|
| 70 |
What does the jni useflag actually do ? Switch between java and native |
| 71 |
implementation of certain things (SSL, and what else?) ? |
| 72 |
|
| 73 |
And it seems that ecj is used instead of the default java compiler. |
| 74 |
Can't jikes be used instead of ecj (IMHO, should be faster) ? |
| 75 |
|
| 76 |
|
| 77 |
cu |
| 78 |
-- |
| 79 |
--------------------------------------------------------------------- |
| 80 |
Enrico Weigelt == metux IT service - http://www.metux.de/ |
| 81 |
--------------------------------------------------------------------- |
| 82 |
Please visit the OpenSource QM Taskforce: |
| 83 |
http://wiki.metux.de/public/OpenSource_QM_Taskforce |
| 84 |
Patches / Fixes for a lot dozens of packages in dozens of versions: |
| 85 |
http://patches.metux.de/ |
| 86 |
--------------------------------------------------------------------- |
| 87 |
-- |
| 88 |
gentoo-dev@g.o mailing list |
Archives