From: Alessio S. <ale...@gm...> - 2009-01-22 22:33:14
Attachments:
abcl-scripting-merge.patch
|
Alright, I have finally found some time to attack the issue of merging the "scripting" branch with trunk. I checked out abcl's trunk head (revision 11575), done a merge in my working copy and verified that it builds. I have run my little test cases (to be added to abcl's own tests sooner or later...) for the jsr-223 support and got no errors. I have also run the ansi-tests for the first time (via the ant target) - it said "build successful" after 25 or so minutes, but I don't know if that means the tests actually passed (shouldn't some fail with current abcl?). Then I generated a patch (which I'm attaching) and briefly contorlled it - it seems to be basically only additions, no deletions, as I expected, so it seems relatively safe. I am aware of at least two problems that imho should be solved before committing the merge: 1. the ant build still checks for java > 1.5 to include jsr-223 support instead of controlling the existence of the javax.script.ScriptEngine class (as someone correctly suggested some time ago) 2. the jsr-223 support is not at all integrated with the lisp-based build (which I don't know since I've only used it once a looong time ago ;-) problem 1 is easy to solve and I'll do it as soon as possible (probably tomorrow). Problem 2 is a bit harder for me, so if anyone has got ideas (or kindly wants to volunteer for such an interesting job ;-), he's welcome. Is everyone ok with this? Bye, Alessio |
From: Erik H. <eh...@gm...> - 2009-01-23 09:31:26
|
Hi Alessio, On Thu, Jan 22, 2009 at 11:01 PM, Alessio Stalla <ale...@gm...> wrote: > Alright, I have finally found some time to attack the issue of merging > the "scripting" branch with trunk. > I checked out abcl's trunk head (revision 11575), done a merge in my > working copy and verified that it builds. I have run my little test > cases (to be added to abcl's own tests sooner or later...) for the > jsr-223 support and got no errors. I have also run the ansi-tests for > the first time (via the ant target) - it said "build successful" after > 25 or so minutes, but I don't know if that means the tests actually > passed (shouldn't some fail with current abcl?). Then I generated a > patch (which I'm attaching) and briefly contorlled it - it seems to be > basically only additions, no deletions, as I expected, so it seems > relatively safe. Thanks for taking the time. The "Build successfull" part is probably misleading, however, a little bit before the end, there should be a list of items which failed. If you have around 45 errors, that's good. > I am aware of at least two problems that imho should be solved before > committing the merge: > > 1. the ant build still checks for java > 1.5 to include jsr-223 > support instead of controlling the existence of the > javax.script.ScriptEngine class (as someone correctly suggested some > time ago) > 2. the jsr-223 support is not at all integrated with the lisp-based > build (which I don't know since I've only used it once a looong time > ago ;-) > problem 1 is easy to solve and I'll do it as soon as possible > (probably tomorrow). Problem 2 is a bit harder for me, so if anyone > has got ideas (or kindly wants to volunteer for such an interesting > job ;-), he's welcome. Problem 2 is "easy" to fix for me; some others may also be able to do it. So, if you merge the change to trunk, possibly you (or Mark) can fix (1) and Mark, Ville or I can fix (2). > Is everyone ok with this? Yes! Thanks again for the great contribution! BTW: I may have already asked you so, but are you using the code yourself? Bye, Erik. |
From: Mark E. <ev...@pa...> - 2009-01-23 09:48:46
|
Erik Huelsmann wrote: [...] > On Thu, Jan 22, 2009 at 11:01 PM, Alessio Stalla > <ale...@gm...> wrote: [...] >> 1. the ant build still checks for java > 1.5 to include jsr-223 >> support instead of controlling the existence of the >> javax.script.ScriptEngine class (as someone correctly suggested some >> time ago) >> 2. the jsr-223 support is not at all integrated with the lisp-based >> build (which I don't know since I've only used it once a looong time >> ago ;-) [...] > > Problem 2 is "easy" to fix for me; some others may also be able to do > it. So, if you merge the change to trunk, possibly you (or Mark) can > fix (1) and Mark, Ville or I can fix (2). > For 1., the general idea would be to use the AVAILABLE task in a target like abcl.init (untested) <available property="jsr-223.p" classname="javax.script.ScriptEngine"/> will set the property 'jsr-223.p' if the class is present in the JVM. -- "A screaming comes across the sky. It has happened before, but there is nothing to compare to it now." |
From: Alessio S. <ale...@gm...> - 2009-01-23 10:24:04
|
On Fri, Jan 23, 2009 at 10:31 AM, Erik Huelsmann <eh...@gm...> wrote: > Thanks for taking the time. The "Build successfull" part is probably > misleading, however, a little bit before the end, there should be a > list of items which failed. If you have around 45 errors, that's good. Well... I didn't see that part, but it's probably an overlook. Unfortunately this weekend I won't have my laptop available, I'll run the tests again on Monday and look more closely at the result ;) >> problem 1 is easy to solve and I'll do it as soon as possible >> (probably tomorrow). ...or much more probably Monday ;) >> Problem 2 is a bit harder for me, so if anyone >> has got ideas (or kindly wants to volunteer for such an interesting >> job ;-), he's welcome. > > Problem 2 is "easy" to fix for me; some others may also be able to do > it. So, if you merge the change to trunk, possibly you (or Mark) can > fix (1) and Mark, Ville or I can fix (2). > >> Is everyone ok with this? > > Yes! Thanks again for the great contribution! Ok :) then Monday I'll rerun the tests, fix point 1 above (thanks Mark), and commit (if tests pass, of course!) > BTW: I may have already asked you so, but are you using the code yourself? I have used it myself to do a little demonstration (a primitive Swing GUI DSL). I'd like to continue developing it and turning it into a real library. Then I'd like to use that library to do the GUI of another idle project of mine. Of course the time required will be of biblical proportions, given the little spare time I have available... but hope never dies :D Ale |
From: Alessio S. <ale...@gm...> - 2009-01-25 23:37:37
|
On Fri, Jan 23, 2009 at 11:23 AM, Alessio Stalla <ale...@gm...> wrote: > On Fri, Jan 23, 2009 at 10:31 AM, Erik Huelsmann <eh...@gm...> wrote: >> Thanks for taking the time. The "Build successfull" part is probably >> misleading, however, a little bit before the end, there should be a >> list of items which failed. If you have around 45 errors, that's good. It says 53 tests failed. I tried with trunk, and I still got 53 failures, so even if I haven't controlled the tests one by one, I think it is pretty probable that the merge won't break anything :) So I have committed it. I also included Mark's contribution for conditionally building JSR-223 support. I tested it with jdk 1.5.0 and 1.6. Let me know if you encounter any problems. Cheers, Alessio |