| resolveTemp howto [message #29406] |
Wed, 04 April 2007 15:30  |
Eclipse User |
|
|
|
Originally posted by: r.c.ladan.tue.nl
This is a multi-part message in MIME format.
--------------090109060809050507000707
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Hi,
I've attached a file which shows the following problem:
----
rule testit {
from
t : behaviourView!MessageEnd
to
t2 : behaviourView!MessageEnd (
m <- thisModule.resolveTemp(t.m, 'm_sm2b'),
name <- thisModule.resolveTemp(t.m, 'm_sm2b').name
)
}
rule add_sendMessage {
from
m_sm1 : behaviourView!Message(m_sm1.name.toLower() = 'sendmessage')
to
m_sm2a : behaviourView!Message (
),
m_sm2b : behaviourView!Message (
name <- m_sm1.name + 'b'
)
}
----
In the rule testit, the first assignment (to t2.m) succeeds, the element
is correctly placed inside its container (see the output in
sd-out-sample.ecore)
But the second assignment in testit (to t2.name) fails. It gives an
empty result, but in my opinion it should return m_sm1.name + 'b'
because of analogy. How can this be solved?
The archive contains:
* sd-in-sample.ecore (the input model)
* sd-out-sample.ecore (the output model)
* behaviour.km3 (the metamodel for both the input and output)
* lb4sd.atl (the transformation with more 'failing' resolveTemp calls)
I'm using the bundled ATL+Eclipse download from
http://www.eclipse.org/m2m/atl/download/ and the default atl2004
Regards,
Rene
--------------090109060809050507000707
Content-Type: application/octet-stream;
name="show-problem.zip"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
filename="show-problem.zip"
UEsDBBQAAAAIAClWhDbnZ9Ue9QAAAG4CAAARAAAAYmVoYXZpb3VyVmlldy5r bTOVUj1vwjAQ
nRMp/+HmShnCCHPHiqpUXRCDk7zQaxM7so8ghPLfa0OBxmorsfl8T+/LznNy 3PUtN4yaBljH
Rmdplvaq+lRbUIl3NbDZ2TfGno5ZmlStco6W5QcqeUFzukuUiOVyJyCtOtCc Vn7W20XYWTSw
0BWoQ+FXT3DOUz/qmkzfG8eCZUPGYhbBTSA6C0XIABwDOLka+ma9z866yGcb ospoUaxh/3TX
XRRjvYA7RrQ3lgkF4ng+8TVgaHICLhb/BaE8J4ETfyaj28PZXfQ693Thzawf NpMmfndmLuWP
P7/Js+WOhQe8Hnq4k3CtRImfbnrjF1BLAwQUAAAACABDiYQ2l4QRrOABAABc BAAACQAAAGxi
NHNkLmF0bI1TS2vbQBA+2+D/MO1FErUNNj0lLaWFFgpNDK2bSwlmVxpbm+zD 7I6smpL/3tmV
6mAXm5yEVt9rZz4ZVzUaQcu3oboeDUuPghAWP5dwBRJrsVOu8XcKW1h7Z+Dr 7ek5k0ZDHzVE
VX0SWtgS/crjGv6MhoNI4sfAlLN0dkJ+tZAPWNJ3XDOIXERqeRkIeUQNrDAI 7yaQOfnAp7f8
msUPxWj4dEhEGEjRURD6T/kGQxAb/GyrQwaaX0D1AUx0p1qFmzTBqcfg9A6X aLY5Tc0YMrMK
Zi6zYhzzMGMyAYu/CQLxiA1agsphAOsIWucfx9DW+w890nmcR4PzDhHBJt24 2ARY3nm1UVZo
CBy7RKhFAMNAVhEWZkDCb5BA2dNd5cVJRq0swkbtOF++FjogbF1QxAcFvC5d o6uUe614Hm6L
XpByFlg8LqJaJp8vPPIfXRD+dudUBXGidgOhYQrttxiu4Nei1B/t/r63/7fX l957mggcu3Yt
kAMjHjGRj5oB3VwHg3OXZG3aAzs1miCXDUHSDXW6qkSIy5x1Zm8gkxkURzV8 URESPdGOWsq7
WAW0VV+x4/8m2p4rY/4cakrum2vR5wW8hyyqmQ6TFYdWpxjinFhX61jWHikv Iw9XP53M81/4
F1BLAwQUAAAACAAliYQ2O4Ajm60AAAD+AAAAEwAAAHNkLW91dC1zYW1wbGUu ZWNvcmVVj0EL
gkAQhe+C/2GZu24GgYmrZw8mFERXtUlX2t1wTf35jWhBc5k38743MHE6qycb sbfSaAGBvwOG
ujZ3qRsB2aXwwvBw9AJIE9eJZyWjW56xpf8ye8q4DqOiU9pGZApoh+EVcT5N k29U45u+4RSE
FRFQYVuO0rz7q8QJkiUeF1WH9XDGB9OlQgGm6mg4kQa+EjlaWzb4P220RX3f NtV6kBD1Zflm
LU/w7QvSH1BLAwQUAAAACAAkiYQ2dpDJlfQAAACFAQAAEgAAAHNkLWluLXNh bXBsZS5lY29y
ZU1QwW7DIAw9t1L/weXSywhrpUltRNJzD10Pk6Zd08RNmAJUQJru7+ckZBoS GPv5PT+Qx6du
4YHOK2sytk1eGaApbaVMnbHTx4Xv928HvmXHfLWUT63Sr/MJhvjH2RFntQRa JGV8SmDGmhDu
qRB93ydW14l1tSAim1oydsWmeCjbuU+FPcsn+nDKM3pf1Aim0Jgxj6aKlblL 6hkM6MM73WZk
Ia3DXQQd3viwdclEDnLNORRl6Iq2/YHNBGzgBSrlAz21U76Bm3UwaFIOnA+i Cyk0jupSRBtj
NlYu128sQxxHY2Z/ZOKfh3H+pDARooAU8TMp/wVQSwECFAAUAAAACAApVoQ2 52fVHvUAAABu
AgAAEQAAAAAAAAABACAAAAAAAAAAYmVoYXZpb3VyVmlldy5rbTNQSwECFAAU AAAACABDiYQ2
l4QRrOABAABcBAAACQAAAAAAAAABACAAAAAkAQAAbGI0c2QuYXRsUEsBAhQA FAAAAAgAJYmE
NjuAI5utAAAA/gAAABMAAAAAAAAAAQAgAAAAKwMAAHNkLW91dC1zYW1wbGUu ZWNvcmVQSwEC
FAAUAAAACAAkiYQ2dpDJlfQAAACFAQAAEgAAAAAAAAABACAAAAAJBAAAc2Qt aW4tc2FtcGxl
LmVjb3JlUEsFBgAAAAAEAAQA9wAAAC0FAAAAAA==
--------------090109060809050507000707--
|
|
|
| Re: [ATL] resolveTemp howto [message #29480 is a reply to message #29406] |
Wed, 04 April 2007 15:38  |
 Frédéric Jouault
Messages: 572
Registered: July 2009 |
Senior Member |
|
|
Hello,
Firstly, could you please prefix the subject of your posts with [ATL] to
avoid collisions with messages about the other components of M2M?
Secondly, navigating the target model is forbidden.
Because resolveTemp returns a target model element, you should not try
to get its name.
In order to avoid recomputing values (and having to change two pieces of
code in case of updates), you may use attribute helpers.
For instance:
helper context behaviourView!Message def: bName : String =
self.name + 'b';
Then, in rule testit, you would write:
name <- t.m.bName
and in rule add_sendMessage:
m_sm2b : behaviourView!Message (
name <- m_sm1.bName
)
Best regards,
Frédéric Jouault
Rene Ladan wrote:
> Hi,
>
> I've attached a file which shows the following problem:
>
> ----
>
> rule testit {
> from
> t : behaviourView!MessageEnd
> to
> t2 : behaviourView!MessageEnd (
> m <- thisModule.resolveTemp(t.m, 'm_sm2b'),
>
> name <- thisModule.resolveTemp(t.m, 'm_sm2b').name
> )
> }
>
>
> rule add_sendMessage {
> from
> m_sm1 : behaviourView!Message(m_sm1.name.toLower() = 'sendmessage')
> to
> m_sm2a : behaviourView!Message (
> ),
> m_sm2b : behaviourView!Message (
> name <- m_sm1.name + 'b'
> )
> }
>
> ----
>
> In the rule testit, the first assignment (to t2.m) succeeds, the element
> is correctly placed inside its container (see the output in
> sd-out-sample.ecore)
>
> But the second assignment in testit (to t2.name) fails. It gives an
> empty result, but in my opinion it should return m_sm1.name + 'b'
> because of analogy. How can this be solved?
>
> The archive contains:
> * sd-in-sample.ecore (the input model)
> * sd-out-sample.ecore (the output model)
> * behaviour.km3 (the metamodel for both the input and output)
> * lb4sd.atl (the transformation with more 'failing' resolveTemp calls)
>
> I'm using the bundled ATL+Eclipse download from
> http://www.eclipse.org/m2m/atl/download/ and the default atl2004
>
> Regards,
> Rene
|
|
|
| Re: [ATL] resolveTemp howto [message #29616 is a reply to message #29480] |
Thu, 05 April 2007 07:47 |
| Eclipse User |
|
|
|
Originally posted by: r.c.ladan.tue.nl
Frédéric Jouault wrote:
> Hello,
>
> Firstly, could you please prefix the subject of your posts with [ATL] to
> avoid collisions with messages about the other components of M2M?
>
>
> Secondly, navigating the target model is forbidden.
> Because resolveTemp returns a target model element, you should not try
> to get its name.
>
Just like in XSLT... But ATL doesn't complain when I write something like
rule A {
from
m1 : model!elem
to
m2 : model!elem (
name <- f(m1.name) -- some function f(String) -> String
),
m3 : model!elem2 (
name <- g(m2.name) -- some function g(String) -> String
),
...
}
In this case m3.name gets the value of g(m2.name), with m2 being an
target model element. Is this a 'feature' in the ATL parser? I
probably should have used g(f(m1.name)) here.
Regards,
Rene
|
|
|
| Re: [ATL] resolveTemp howto [message #30069 is a reply to message #29616] |
Thu, 05 April 2007 14:59 |
Frédéric Jouault
Messages: 572
Registered: July 2009 |
Senior Member |
|
|
Hello,
When you navigate the target model, results can vary: it *might* work
sometimes. But you should not rely on this.
In the future, ATL will be able to tell you when you navigate the target
model. However, the current version lets you try.
Best regards,
Frédéric Jouault
Rene Ladan wrote:
> Frédéric Jouault wrote:
>> Hello,
>>
>> Firstly, could you please prefix the subject of your posts with [ATL]
>> to avoid collisions with messages about the other components of M2M?
>>
>>
>> Secondly, navigating the target model is forbidden.
>> Because resolveTemp returns a target model element, you should not try
>> to get its name.
>>
> Just like in XSLT... But ATL doesn't complain when I write something like
>
> rule A {
> from
> m1 : model!elem
> to
> m2 : model!elem (
> name <- f(m1.name) -- some function f(String) -> String
> ),
> m3 : model!elem2 (
> name <- g(m2.name) -- some function g(String) -> String
> ),
> ...
> }
>
> In this case m3.name gets the value of g(m2.name), with m2 being an
> target model element. Is this a 'feature' in the ATL parser? I
> probably should have used g(f(m1.name)) here.
>
> Regards,
> Rene
|
|
|
Powered by
FUDForum. Page generated in 0.03201 seconds
] 
Search
Help
Register
Login
Home
