[m-rev.] for review: io.stream_db.m

[Julien Fischer]

On Fri, 11 Mar 2022, Zoltan Somogyi wrote:

>
> 2022-03-11 22:30 GMT+11:00 "Julien Fischer" <jfischer at opturion.com>:
>>
>> On Fri, 11 Mar 2022, Zoltan Somogyi wrote:
>>
>>> 2022-03-11 18:05 GMT+11:00 "Julien Fischer" <jfischer at opturion.com>:
>>>> Ok, I'm still inclined to place it in the submodule however.
>>>
>>> That is what I committed.
>>>
>>>> Related to this: why is there a declaration for the type stream_id/0
>>>> in the publicly documented interface of io.m?  There's nothing in that
>>>> part of the interface that would allows users to do anything with a
>>>> stream_id value.
>>>
>>> Agreed. The attached diff, most of which addresses one of Peter's
>>> earlier reviews, also deletes this redundant type declaration.
>>
>> Unfortunately, its removal breaks installation in the Java grade.
>>
>>    Making Mercury/javas/jmercury/mdb__browser_info.java
>>    Uncaught Mercury exception:
>>    Software Error: map.lookup: key not found
>>         Key Type: string
>>         Key Value: "stream_id"
>>         Value Type: tree234.tree234(parse_tree.prog_data.type_ctor,
>>           hlds.hlds_data.hlds_type_defn)
>>
>> (At a guess this may have something to do with the handling of
>> equivalence types in the Java backend?)
>
> May be. I will take a look once my Java bootcheck has finished.
> If I don't get any better ideas, I will put the declaration back,
> but this time into the undocumented interface section.

I will do that now with the addtion of an XXX comment about
the above, since this is breaking the ROTD builds.

Julien.