Issue13

Title merge the various translators
Priority feature Status resolved
Superseder Nosy List gabi, malte, silvia
Assigned To gabi Keywords
Optional summary

Created on 2009-10-07.18:24:38 by malte, last changed by malte.

Messages
msg52 (view) Author: malte Date: 2009-10-09.19:04:02
OK, I removed the LAMA translators.

About the no-invariants.patch, I copied to the trunk translator to save it for
now. In the future, I guess we should make this an option of the translator.
I'll open a separate ticket for that.
msg50 (view) Author: silvia Date: 2009-10-09.18:50:09
Malte: Sure, the three LAMA translators can be removed now.

Gabi: The file "no-invariants.patch" is a patch that turns off invariant
generation in the translator (i.e., the resulting output will have exactly the
binary variables of the PDDL input). This has been helpful for some people in
tasks where the invariant generation was not working properly, so I've been
giving this patch to people as a work-around for their problems. It might be
worthwile to keep this file... or put an option into the translator that skips
invariant generation.
msg46 (view) Author: malte Date: 2009-10-09.17:24:21
I deleted the aij-translator branch.

Silvia, I would also like to delete the remaining three branches of the translator:

branches/lama/trunk/downward/translate
branches/lama/branches/lama-integration/downward/translate
branches/lama/branches/experimental/downward/translate

Is that OK with you? They should all be superseded by the translator in trunk,
which is essentially the LAMA translator, but with some more detailed
information about which parts of the translator take how much time.

Once that is done, we're finished unifying the translator. :-)
msg45 (view) Author: gabi Date: 2009-10-09.16:40:02
I merged the aij-translator into the trunk.

Now there is an additional file relaxed_task.py that might do 
something similar as relaxed_tasks.py and translate_relaxed.py, 
but I did not look into that.
msg42 (view) Author: gabi Date: 2009-10-09.11:13:16
I merged the lama translator into the trunk, which
implies three major changes to the trunk version:
- support of action-costs (r3283)
- the invariant synthesis is aborted if it takes too
  long or finds too many invariant candidates (r3284)
- conflicting add and delete effects are now handled
  differently (r3285)

Open issues:
- I did *not* yet adapt the following files that exist only
  in the trunk:
  * additive-hmax.py
  * relaxed_tasks.py
  * run-additive-hmax
  * translate-relaxed.py
- I ignored the file no-invariants.patch in the lama
  translator directory. Is this file still in use and
  should it be merged to the trunk?
msg39 (view) Author: malte Date: 2009-10-08.23:02:00
I removed the translators that we don't need to merge and checked that indeed
there is nothing interesting in their code. The current codebase (r3272) now
contains the following translator versions, excluding ext and temporal branches:

# find . | grep '/translate$' | grep -v '/results/translate$' | grep -v temporal
| grep -v /ext/

./trunk/downward/translate
./branches/lama/trunk/downward/translate
./branches/lama/branches/lama-integration/downward/translate
./branches/lama/branches/experimental/downward/translate
./branches/aij-translator/downward/translate
./branches/aij-translator-exp/downward/translate

Of these, the three LAMA translators are completely identical, and I'll get rid
of aij-translator-exp soon. So what remains to do is to merge the translators of
aij-translator and lama into the trunk.
msg38 (view) Author: malte Date: 2009-10-08.22:35:22
The translators in the aij-translator and aij-translator-exp branches have been
merged to the extent that this makes sense (r3269), and the translator in the
aij-translator branch is the one that we should integrate into the trunk.
msg24 (view) Author: malte Date: 2009-10-07.18:24:38
We need to merge the translators from the various branches of the code, namely:

 * everything
 * lama/trunk
 * aij-translator
 * aij-translator-exp
 * any others that are different from everything
History
Date User Action Args
2009-10-09 19:04:02maltesetstatus: in-progress -> resolved
messages: + msg52
2009-10-09 18:50:09silviasetmessages: + msg50
2009-10-09 17:24:21maltesetnosy: + silvia
messages: + msg46
2009-10-09 16:40:02gabisetmessages: + msg45
2009-10-09 11:13:17gabisetmessages: + msg42
2009-10-08 23:02:00maltesetmessages: + msg39
2009-10-08 22:35:22maltesetmessages: + msg38
2009-10-08 00:01:25maltesetpriority: wish -> feature
2009-10-07 18:24:38maltecreate