Hi Steve!
Awesome, thank you! Those look like a great place to start, I just need to
update the dev tools on this Mac.
I had had a quick look at DAFFODIL-1714… The operator code appears to be
broken. It divides the number by 2 to fake unsigned maths using signed longs,
but then needs to correct for that. There’s a commented out fix for the
quotient, but something very similar is needed for the remainder. Basically:
val r = n - q0 * d
// val q = q0 + (if (ULong(r) >= ULong(d)) 1 else 0)
ULong(r.toLong)
Should be:
val r = n - q0 * d
if (ULong(r) >= ULong(d))
ULong(r - d)
else
ULong(r)
Or alternatively, the commented-out line could be included and moved up, then r
calculated using q instead of q0.
Cheers,
—
Russ
> On 23 May 2018, at 18:51, Steve Lawrence <[email protected]> wrote:
>
> Hi Russ,
>
> That's probably a good place to start. The code contributor workflow [1]
> should have all the steps necessary to get the code, build, test, and
> make contributions. If you get stuck anywhere, let us know, it probably
> means the instruction are bad.
>
> Until we get bugs tagged as good tasks for new contributors, perhaps a
> good introductory bug is DAFFODIL-1928 and maybe related bug
> DAFFODIL-1420. The fix for these should just be a matter of updating a
> libraryDependency and moving existing test cases out of
> src/test/scala-debug and into src/test/scala. Should be a fairly small
> and hopefully painless first contribution just to get you familiar with
> the build/test/pull-request process of Daffodil.
>
> Thanks!
> - Steve
>
> [1]
> https://cwiki.apache.org/confluence/display/DAFFODIL/Code+Contributor+Workflow
>
>
> On 05/23/2018 01:21 PM, Russ Williams wrote:
>> Hi Mike!
>>
>> I guess I should grab a copy of the source, get it building locally, and
>> then take a look through the JIRAs to find some small tasks that I can
>> figure out how to do with ~zero knowledge of the codebase. That’ll probably
>> be a good shortlist for newbies?
>>
>> Cheers,
>> —
>> Russ
>>
>>> On 23 May 2018, at 17:51, Mike Beckerle <[email protected]> wrote:
>>>
>>> Hi Russ,
>>>
>>>
>>> Glad you contacted us. Would be great to have another contributor.
>>>
>>>
>>> Your insights are correct - DFDL evolved driven by standardization of
>>> existing data integration tool capabilities, which were focused on files
>>> that you would probably call "data set files", as opposed to image,
>>> document, archive, etc. file formats. We didn't have examples of data
>>> integration systems/tools that did things like file-offsets, so things like
>>> the index structure in zip you described, were beyond the state of the art
>>> for declarative description.
>>>
>>>
>>> But now we have Daffodil open-source, which is the perfect vehicle for
>>> prototyping and creating features enabling these sorts of data
>>> descriptions, and subsequently then, based on that experience, we can feed
>>> ideas back to the DFDL standard for incorporation into a future version of
>>> the standard.
>>>
>>>
>>> As for specifics of how you can start to contribute, longer discussion.
>>>
>>>
>>> We have ambition to label the JIRA tickets for Daffodil to identify good
>>> first projects for new contributors, .... The whole pool of JIRA tickets,
>>> which are our project-wide TODO list, is at
>>>
>>> https://issues.apache.org/jira/projects/DAFFODIL
>>>
>>>
>>> There are 450+ tickets open, so there is lots to work on.
>>>
>>>
>>> I'll leave it at that for this message.
>>>
>>>
>>> ...mike beckerle
>>>
>>>
>>>
>>>
>>>
>>>
>>> ________________________________
>>> From: Russ Williams <[email protected]>
>>> Sent: Wednesday, May 23, 2018 5:54:14 AM
>>> To: [email protected]
>>> Subject: Can I help?
>>>
>>> Hi!
>>>
>>> I stumbled across DFDL / Apache Daffodil yesterday while looking for a way
>>> to specify file formats in a machine-readable form. I was surprised that
>>> there hasn’t been a lot more effort in this space, given the importance for
>>> archival, and I’m very keen to see the project succeed. I’ve not done much
>>> Open Source work before, but I’m a commercial software engineer/architect
>>> with over two decades’ experience, so hopefully I could be of some use.
>>>
>>> I’m particularly interested in the handling of large binary files, which I
>>> see from the wiki and JIRA (e.g. DAFFODIL-1735) is a key area of concern
>>> for you guys as well, but I’m a little concerned that the DFDL 1.0 spec
>>> seems to have been written with some XML-like assumptions of how parsing
>>> should work, rather than how various binary formats are actually parsed
>>> (e.g. ZIP, with a signature at the start, then an index at the end, doesn’t
>>> seem to fit the document model).
>>>
>>> Are you looking for people to get involved with Daffodil? Is there anything
>>> I can help with to get started?
>>>
>>> Warm regards,
>>> —
>>> Russ
>>
>
