Duplicate definitions being generated in retrieveSQLSelectResults

[Trellian]

Hi Sam,

I have a table that has two FK's in it as follows:

public class TblMapVisitor2Entry extends BaseObservable implements Parcelable
    {
    @Column(index = true) private int       id;
    @Column private TblEntry   tblEntry;
    @Column private TblVisitor tblVisitor;

and in Q.Java, I get an error on duplicate symbols as follows:

      for (int i = 0; !cursor.isAfterLast(); i++) {
        com.thorburn.sanscan.scanDB.TblMapVisitor2Entry tblmapvisitor2entry =
            new com.thorburn.sanscan.scanDB.TblMapVisitor2Entry();

        com.thorburn.sanscan.scanDB.TblEntry tblentry = new com.thorburn.sanscan.scanDB.TblEntry();
        tblmapvisitor2entry.setTblEntry(tblentry);

        com.thorburn.sanscan.scanDB.TblVisitor tblvisitor =
            new com.thorburn.sanscan.scanDB.TblVisitor();
        tblmapvisitor2entry.setTblVisitor(tblvisitor);

        com.thorburn.sanscan.scanDB.TblEntry tblentry = new com.thorburn.sanscan.scanDB.TblEntry();
        tblmapvisitor2entry.setTblEntry(tblentry);

        com.thorburn.sanscan.scanDB.TblVisitor tblvisitor =
            new com.thorburn.sanscan.scanDB.TblVisitor();
        tblmapvisitor2entry.setTblVisitor(tblvisitor);

You can see that 'tblentry' and 'tblvisitor' variables are being declared twice. I have had a look at the preprocessor template source for Q.Java but can't see how it could be happening. I'm currently taking a shot at fixing it, but not having much success.

Please can you take a look?

Thanks, Adrian

[samkirton]

Hi Adrian,

Could you paste the source of "BaseObservable" also?

Cheers, Sam

[Trellian]

Hi Sam,

You can see the source here: https://android.googlesource.com/platform/frameworks/data-binding/+/c9a5d6f140f732ca0ff279a4b1ee315072e1c422/library/src/main/java/android/databinding/BaseObservable.java

Thanks for the quick response!

/Adrian

On 02 Sep 2016 11:34, Samuel Kirton wrote:

Hi Adrian,

Could you paste the source of "BaseObservable" also?

Cheers, Sam

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/memtrip/SQLKing/issues/14#issuecomment-244329223, or mute the thread https://github.com/notifications/unsubscribe-auth/ABCOrHj89qUoypJUc3-pBWToCJ-2RyqWks5ql-2LgaJpZM4JzfYh.

[Trellian]

BaseObservable is being used because I'm using the new 2-way DataBinding features, instead of Butterknife

/Adrian

On 02 Sep 2016 11:34, Samuel Kirton wrote:

Hi Adrian,

Could you paste the source of "BaseObservable" also?

Cheers, Sam

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/memtrip/SQLKing/issues/14#issuecomment-244329223, or mute the thread https://github.com/notifications/unsubscribe-auth/ABCOrHj89qUoypJUc3-pBWToCJ-2RyqWks5ql-2LgaJpZM4JzfYh.

[Trellian]

I have tried this in Q.Java (starting at line 167 - in retrieveSQLSelectResults):

I'm building a list of columns already processed for the table, to avoid generating duplicates.

                for (int i = 0; !cursor.isAfterLast(); i++) {
                    ${packagedTableName} ${table.getName()?lower_case} = new ${packagedTableName}();
                    <#assign colsProcessed=[]>
                    <#list table.getColumns() as column>
                        <#if column.isJoinable(tables) >
                            <#if colsProcessed?seq_contains(column.getName()) != true>
                            ${joinReferences(table.getName(),tables)}
                                <#assign colsProcessed = colsProcessed + [column.getName()]>
                            </#if>
                        </#if>
                    </#list>

Preprocessor builds fine, but final Q.Java doesn't generate now, so obviously I've made a stupid mistake.

[samkirton]

If the final Q doesn't generate it will be a FreeMarker issue, if you build the project with mvn clean install -X it should be logged.

Also is 2 way binding available in the support library?

[Trellian]

I've got the final Q to generate now, it still has the double declarations, but FreeMarker seems happy for now.

One thing that concerns me is that the Maven Install installs to: C:\Users\Adrian.m2\repository\com\memtrip\sqlking\preprocessor\1.0.15\preprocessor-1.0.15.jar

And it looks like my gradle build of my project is using what it has in it's cache at

C:\Users\Adrian\m2repository\com\memtrip\sqlking\preprocessor\1.0.15\preprocessor-1.0.15.jar

I couldn't find a way to point either one to a consolidated maven repository.

Have you experienced this before?

Regards, Adrian

On 02 Sep 2016 11:56, Samuel Kirton wrote:

If the final Q doesn't generate it will be a FreeMarker issue, if you build the project with |mvn clean install -X| it should be logged.

Also is 2 way binding available in the support library?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/memtrip/SQLKing/issues/14#issuecomment-244334152, or mute the thread https://github.com/notifications/unsubscribe-auth/ABCOrHSdDh9rjRb4ysSBuQE1zEwT6E6Aks5ql_LLgaJpZM4JzfYh.

[samkirton]

On the client project it might be worth using: gradle build --refresh-dependencies

Also its probably best if you stash your changes and see if you can reproduce the issue, it might have been fixed in 1.1, 1.0.15 is quite old!

[Trellian]

I'm a bit confused here... it turns out that I was indeed using 1.1, but changes to my local 1.0.15 showed up as compile errors in my client project?

Anyway, I have moved to 1.1, the original problem still exists with the duplicate declarations.

Any Ideas?

Thanks, Adrian

[samkirton]

Hi Adrian,

Please see the fix/duplicate_relationships branch, the issue has been fixed. I will push a release to maven central tonight.

Cheers, Sam

[Trellian]

Hi Sam, Did you get a chance to push a release to Maven?

Thanks, Adrian

[samkirton]

All deployed.

Cheers, Sam

[Trellian]

Thanks Sam, works like a charm!