[OFBiz] Dev - why set operation for minilang?
Locked
[OFBiz] Dev - why set operation for minilang?
 
|
Re: [OFBiz] Dev - why set operation for minilang?
|
|
Si, The idea of putting the set operation into the simple-method has been on the table for a while, part of a general effort to make tools more consistent and such. The related operations in the simple-method are a mess right now and I don't like them a whole lot (and a lot of people have been quite open about sharing similar sentiments...). The field-to-field, env-to- field and field-to-env operations are all there for legacy reasons. In fact the whole field-name/map-name pairs that exist all over the place only exist because originally we didn't have the dot and string expansion syntax stuff in place so it was the only way to get at a Map entry. After the map accessor and string expander were introduced the preferred operation to use was the env-to-env operation, but the set operation and the way it works is much more flexible and can take the place of all of those, plus the string-to-field operation, plus it does type conversions and supports a default-value and such. Sooner or later I'd like to get rid of all field-to-field, field-to- env, env-to-field, env-to-env, and even string-to-field operations and replace them all with usage of set, and then just get rid of those as they are confusing (as demonstrated by this question...). -David On Sep 26, 2005, at 10:00 AM, Si Chen wrote: > David, > > I saw your solution of creating a "set" operation for typecasting > in minilang. So does "set" cause the old "field-to-field" to be > deprecated? When should one use "set" versus "field-to-field"? Is > it bad to put the typecasting code into "field-to-field," as our > patch suggested? > > Thanks, > > Si > _______________________________________________ > Dev mailing list > [hidden email] > http://lists.ofbiz.org/mailman/listinfo/dev > _______________________________________________ Dev mailing list [hidden email] http://lists.ofbiz.org/mailman/listinfo/dev |
smime.p7s (3K)
Re: [OFBiz] Dev - why set operation for minilang?
|
|
Re: [OFBiz] Dev - why set operation for minilang?
|
|
I'm not sure about this.... If we were to do something like this we might as well do it in the Entity Engine itself, but I'm not sure if I like that idea. In other words, we could always try to convert the data to the right type in the field.set/put operation, and then if it failed throw an exception rather than throwing an exception when it is the wrong type. Would this be a good thing or perhaps cause more problems than it solves? -David On Sep 26, 2005, at 12:53 PM, Si Chen wrote: > Ok. So now I see that field-to-field is deprecated. I also did a > > sichen@linux:~/eclipse/workspace/ofbiz/applications> grep -r field- > to-field * | wc -l > 2164 > > There are still a lot of field-to-field operations out there. > Would it be bad thing, for now, to use our type conversion code in > FieldToField.java? > > Si > > Si Chen wrote: > > >> David, >> >> I saw your solution of creating a "set" operation for typecasting >> in minilang. So does "set" cause the old "field-to-field" to be >> deprecated? When should one use "set" versus "field-to-field"? >> Is it bad to put the typecasting code into "field-to-field," as >> our patch suggested? >> >> Thanks, >> >> Si >> >> _______________________________________________ >> Dev mailing list >> [hidden email] >> http://lists.ofbiz.org/mailman/listinfo/dev >> >> > _______________________________________________ > Dev mailing list > [hidden email] > http://lists.ofbiz.org/mailman/listinfo/dev > _______________________________________________ Dev mailing list [hidden email] http://lists.ofbiz.org/mailman/listinfo/dev |
Re: [OFBiz] Dev - why set operation for minilang?
|
|
| Free forum by Nabble | Edit this page |