You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@camel.apache.org by Claus Ibsen <cl...@gmail.com> on 2014/11/05 20:14:46 UTC
Re: Support of official commons-csv
Hi
Thanks. I think its a good time to migrate this in Camel 2.15. But the
PR has merge conflicts now. Do you mind rebasing so the PR can be
merged on master branch?
On Wed, Oct 15, 2014 at 12:10 PM, Antoine DESSAIGNE
<an...@gmail.com> wrote:
> Hi everyone,
>
> I've submitted this PR https://github.com/apache/camel/pull/292 for
> updating csv to the official version
>
> What do you think ?
>
> Have a nice day.
>
> Antoine.
>
> 2014-09-12 11:59 GMT+02:00 Antoine DESSAIGNE <an...@gmail.com>:
>
>> Hi,
>>
>> I don't see how it can be compatible since the public API of CsvDataFormat
>> relies on 2 classes that were deleted : CSVStrategy and CSVConfig.
>>
>> I'm also not convinced by a camel-csv2 component since it's common-csv
>> version 1. I would rather have another component ("camel-csv-commons" ?)
>> that publishes the <commons-csv> data format (or <csv-commons>).
>>
>> Which is the right way ?
>>
>> Also, I didn't managed to build the master. I keep having "Failed to
>> execute goal
>> org.apache.camel:camel-api-component-maven-plugin:2.15-SNAPSHOT:fromApis
>> (generate-test-component-classes) on project camel-box: Error generating
>> source for com.box.boxjavalibv2.resourcemanagers.IBoxCollaborationsManager:
>> String index out of range: 1"
>>
>> What did I do wrong ?
>>
>> Thanks.
>>
>> Antoine
>>
>> 2014-09-11 21:56 GMT+02:00 Christian Müller <ch...@gmail.com>:
>>
>>> +1
>>>
>>> I think a camel-csv2 component is not necessary.
>>>
>>> Best,
>>>
>>> Christian
>>> -----------------
>>>
>>> Software Integration Specialist
>>>
>>> Apache Member
>>> V.P. Apache Camel | Apache Camel PMC Member | Apache Camel committer
>>> Apache Incubator PMC Member
>>>
>>> https://www.linkedin.com/pub/christian-mueller/11/551/642
>>>
>>> On Thu, Sep 11, 2014 at 2:30 AM, Willem Jiang <wi...@gmail.com>
>>> wrote:
>>>
>>> > When I upgraded the http client version of camel-http4 to 4.3.x, there
>>> > were some API changes in http client just like common-csv. As we can do
>>> > some adaption inside camel-http4, I just changed the configuration a
>>> bit.
>>> > I think we can do the same thing in camel-csv :)
>>> >
>>> > --
>>> > Willem Jiang
>>> >
>>> > Red Hat, Inc.
>>> > Web: http://www.redhat.com
>>> > Blog: http://willemjiang.blogspot.com (English)
>>> > http://jnn.iteye.com (Chinese)
>>> > Twitter: willemjiang
>>> > Weibo: 姜宁willem
>>> >
>>> >
>>> >
>>> > On September 10, 2014 at 10:57:03 PM, Claus Ibsen (
>>> claus.ibsen@gmail.com)
>>> > wrote:
>>> > > Hi
>>> > >
>>> > > Welcome back.
>>> > >
>>> > > I suspect the breakings is less affected by Camel users, eg I would
>>> > > assume the data it creates on the Camel Exchange / Message / body /
>>> > > headers can be the same as before? So the changes affected is maybe
>>> > > how to configure it using the csv data format or its csv config
>>> > > object?
>>> > >
>>> > > I am fine to make this change for Camel 2.15 and add some notes to the
>>> > > release notes that end users need to do some changes.
>>> > >
>>> > >
>>> > >
>>> > > On Wed, Sep 10, 2014 at 4:46 PM, Antoine DESSAIGNE
>>> > > wrote:
>>> > > > Hi everyone,
>>> > > >
>>> > > > I was looking at this issue :
>>> > > > https://issues.apache.org/jira/browse/CAMEL-7716.
>>> > > >
>>> > > > In fact, the commons-csv library changed a lot between the version
>>> that
>>> > > > camel uses and the official one. In addition, there is lot of nice
>>> > stuff
>>> > > > such as:
>>> > > > * immutable configurations, it's no longer required to copy the
>>> > > > configuration
>>> > > > * build-in iterator mechanism
>>> > > > * etc...
>>> > > >
>>> > > > Supporting the commons-csv API would break the compatibility with
>>> > existing
>>> > > > code and usage.
>>> > > >
>>> > > > What should I do ? Break compatibility ? Create a new data format ?
>>> > > >
>>> > > > Thanks a lot for your insights on this matter,
>>> > > >
>>> > > > Have a nice day,
>>> > > >
>>> > > > Antoine.
>>> > > >
>>> > > > PS: I was very busy, but now I managed to freed up time for camel :)
>>> > >
>>> > >
>>> > >
>>> > > --
>>> > > Claus Ibsen
>>> > > -----------------
>>> > > Red Hat, Inc.
>>> > > Email: cibsen@redhat.com
>>> > > Twitter: davsclaus
>>> > > Blog: http://davsclaus.com
>>> > > Author of Camel in Action: http://www.manning.com/ibsen
>>> > > hawtio: http://hawt.io/
>>> > > fabric8: http://fabric8.io/
>>> > >
>>> >
>>> >
>>>
>>
>>
--
Claus Ibsen
-----------------
Red Hat, Inc.
Email: cibsen@redhat.com
Twitter: davsclaus
Blog: http://davsclaus.com
Author of Camel in Action: http://www.manning.com/ibsen
hawtio: http://hawt.io/
fabric8: http://fabric8.io/
Re: Support of official commons-csv
Posted by Antoine DESSAIGNE <an...@gmail.com>.
Hi everyone,
This weekend commons-csv 1.1 was released, thus I've updated the
pull-request with the latest version.
Have a nice day,
Antoine.
2014-11-07 18:39 GMT+01:00 Antoine DESSAIGNE <an...@gmail.com>:
> Hi everyone,
>
> Since it's a pretty big pull-request, I'm going to explain the changes
> I've made and the reasons for them.
>
> The most obvious change is the model of commons-csv.
> Previously, commons-csv was using 2 objects, a CSVStrategy and a
> CSVConfig, that we configured using a getter/setter approach. They no
> longer use these objects and this approach but a single CSVFormat object.
> This CSVFormat object is created by applying changes on a reference format,
> each change produces a new CSVFormat object.
> Thus, now you can set the reference format (before applying any changes)
> either using the object directly or its common name.
>
> Also, there is a lot of nice configuration options of the format, I
> decided to make them all configurable using the Java API or the Spring API.
> For those options, null often means that the feature is disabled. Since for
> the data format null means "unconfigured, use default", I've added a
> "disabled" options for each CSVFormat option.
>
> In commons-csv, the most interesting option is the header configuration.
> You can specify the headers using a String array. If you want automatic
> headers, you need to use an empty array. If you want to disable the headers
> you need to pass null. Obviously you can configure all that using either
> the Java API or the Spring API (using children <header> tags). A
> consequence of exposing this configuration is that you no longer need the
> autogenColumns parameter, thus I removed it.
>
> Also, in order to ease the readability of the code, I've created 2
> classes: CsvMarshaller and CsvUnmarshaller. The CsvDataFormat delegates to
> them the marshalling and unmarshalling processes. These classes also handle
> the different meaning of the same option like "skipHeaderRecord". When
> unmarshalling you're indicating that there is an header row that you don't
> want to read. When marshalling you're indicating that the header row should
> not be written.
>
> I really hope that you'll find in this email the reasons behind the
> changes you saw in the code.
>
> Anyhow, If you want to review the code with me, I will be more that happy
> to organize it using Skype.
>
> Also, as you may have noticed it's a pretty big change and thus maybe it
> should be in another data format module (csv1 ?).
>
> Have a nice day,
>
> Antoine.
>
> 2014-11-05 20:14 GMT+01:00 Claus Ibsen <cl...@gmail.com>:
>
>> Hi
>>
>> Thanks. I think its a good time to migrate this in Camel 2.15. But the
>> PR has merge conflicts now. Do you mind rebasing so the PR can be
>> merged on master branch?
>>
>>
>>
>> On Wed, Oct 15, 2014 at 12:10 PM, Antoine DESSAIGNE
>> <an...@gmail.com> wrote:
>> > Hi everyone,
>> >
>> > I've submitted this PR https://github.com/apache/camel/pull/292 for
>> > updating csv to the official version
>> >
>> > What do you think ?
>> >
>> > Have a nice day.
>> >
>> > Antoine.
>> >
>> > 2014-09-12 11:59 GMT+02:00 Antoine DESSAIGNE <
>> antoine.dessaigne@gmail.com>:
>> >
>> >> Hi,
>> >>
>> >> I don't see how it can be compatible since the public API of
>> CsvDataFormat
>> >> relies on 2 classes that were deleted : CSVStrategy and CSVConfig.
>> >>
>> >> I'm also not convinced by a camel-csv2 component since it's common-csv
>> >> version 1. I would rather have another component ("camel-csv-commons"
>> ?)
>> >> that publishes the <commons-csv> data format (or <csv-commons>).
>> >>
>> >> Which is the right way ?
>> >>
>> >> Also, I didn't managed to build the master. I keep having "Failed to
>> >> execute goal
>> >>
>> org.apache.camel:camel-api-component-maven-plugin:2.15-SNAPSHOT:fromApis
>> >> (generate-test-component-classes) on project camel-box: Error
>> generating
>> >> source for
>> com.box.boxjavalibv2.resourcemanagers.IBoxCollaborationsManager:
>> >> String index out of range: 1"
>> >>
>> >> What did I do wrong ?
>> >>
>> >> Thanks.
>> >>
>> >> Antoine
>> >>
>> >> 2014-09-11 21:56 GMT+02:00 Christian Müller <
>> christian.mueller@gmail.com>:
>> >>
>> >>> +1
>> >>>
>> >>> I think a camel-csv2 component is not necessary.
>> >>>
>> >>> Best,
>> >>>
>> >>> Christian
>> >>> -----------------
>> >>>
>> >>> Software Integration Specialist
>> >>>
>> >>> Apache Member
>> >>> V.P. Apache Camel | Apache Camel PMC Member | Apache Camel committer
>> >>> Apache Incubator PMC Member
>> >>>
>> >>> https://www.linkedin.com/pub/christian-mueller/11/551/642
>> >>>
>> >>> On Thu, Sep 11, 2014 at 2:30 AM, Willem Jiang <willem.jiang@gmail.com
>> >
>> >>> wrote:
>> >>>
>> >>> > When I upgraded the http client version of camel-http4 to 4.3.x,
>> there
>> >>> > were some API changes in http client just like common-csv. As we
>> can do
>> >>> > some adaption inside camel-http4, I just changed the configuration a
>> >>> bit.
>> >>> > I think we can do the same thing in camel-csv :)
>> >>> >
>> >>> > --
>> >>> > Willem Jiang
>> >>> >
>> >>> > Red Hat, Inc.
>> >>> > Web: http://www.redhat.com
>> >>> > Blog: http://willemjiang.blogspot.com (English)
>> >>> > http://jnn.iteye.com (Chinese)
>> >>> > Twitter: willemjiang
>> >>> > Weibo: 姜宁willem
>> >>> >
>> >>> >
>> >>> >
>> >>> > On September 10, 2014 at 10:57:03 PM, Claus Ibsen (
>> >>> claus.ibsen@gmail.com)
>> >>> > wrote:
>> >>> > > Hi
>> >>> > >
>> >>> > > Welcome back.
>> >>> > >
>> >>> > > I suspect the breakings is less affected by Camel users, eg I
>> would
>> >>> > > assume the data it creates on the Camel Exchange / Message / body
>> /
>> >>> > > headers can be the same as before? So the changes affected is
>> maybe
>> >>> > > how to configure it using the csv data format or its csv config
>> >>> > > object?
>> >>> > >
>> >>> > > I am fine to make this change for Camel 2.15 and add some notes
>> to the
>> >>> > > release notes that end users need to do some changes.
>> >>> > >
>> >>> > >
>> >>> > >
>> >>> > > On Wed, Sep 10, 2014 at 4:46 PM, Antoine DESSAIGNE
>> >>> > > wrote:
>> >>> > > > Hi everyone,
>> >>> > > >
>> >>> > > > I was looking at this issue :
>> >>> > > > https://issues.apache.org/jira/browse/CAMEL-7716.
>> >>> > > >
>> >>> > > > In fact, the commons-csv library changed a lot between the
>> version
>> >>> that
>> >>> > > > camel uses and the official one. In addition, there is lot of
>> nice
>> >>> > stuff
>> >>> > > > such as:
>> >>> > > > * immutable configurations, it's no longer required to copy the
>> >>> > > > configuration
>> >>> > > > * build-in iterator mechanism
>> >>> > > > * etc...
>> >>> > > >
>> >>> > > > Supporting the commons-csv API would break the compatibility
>> with
>> >>> > existing
>> >>> > > > code and usage.
>> >>> > > >
>> >>> > > > What should I do ? Break compatibility ? Create a new data
>> format ?
>> >>> > > >
>> >>> > > > Thanks a lot for your insights on this matter,
>> >>> > > >
>> >>> > > > Have a nice day,
>> >>> > > >
>> >>> > > > Antoine.
>> >>> > > >
>> >>> > > > PS: I was very busy, but now I managed to freed up time for
>> camel :)
>> >>> > >
>> >>> > >
>> >>> > >
>> >>> > > --
>> >>> > > Claus Ibsen
>> >>> > > -----------------
>> >>> > > Red Hat, Inc.
>> >>> > > Email: cibsen@redhat.com
>> >>> > > Twitter: davsclaus
>> >>> > > Blog: http://davsclaus.com
>> >>> > > Author of Camel in Action: http://www.manning.com/ibsen
>> >>> > > hawtio: http://hawt.io/
>> >>> > > fabric8: http://fabric8.io/
>> >>> > >
>> >>> >
>> >>> >
>> >>>
>> >>
>> >>
>>
>>
>>
>> --
>> Claus Ibsen
>> -----------------
>> Red Hat, Inc.
>> Email: cibsen@redhat.com
>> Twitter: davsclaus
>> Blog: http://davsclaus.com
>> Author of Camel in Action: http://www.manning.com/ibsen
>> hawtio: http://hawt.io/
>> fabric8: http://fabric8.io/
>>
>
>
Re: Support of official commons-csv
Posted by Antoine DESSAIGNE <an...@gmail.com>.
Hi everyone,
Since it's a pretty big pull-request, I'm going to explain the changes I've
made and the reasons for them.
The most obvious change is the model of commons-csv.
Previously, commons-csv was using 2 objects, a CSVStrategy and a CSVConfig,
that we configured using a getter/setter approach. They no longer use these
objects and this approach but a single CSVFormat object. This CSVFormat
object is created by applying changes on a reference format, each change
produces a new CSVFormat object.
Thus, now you can set the reference format (before applying any changes)
either using the object directly or its common name.
Also, there is a lot of nice configuration options of the format, I decided
to make them all configurable using the Java API or the Spring API. For
those options, null often means that the feature is disabled. Since for the
data format null means "unconfigured, use default", I've added a "disabled"
options for each CSVFormat option.
In commons-csv, the most interesting option is the header configuration.
You can specify the headers using a String array. If you want automatic
headers, you need to use an empty array. If you want to disable the headers
you need to pass null. Obviously you can configure all that using either
the Java API or the Spring API (using children <header> tags). A
consequence of exposing this configuration is that you no longer need the
autogenColumns parameter, thus I removed it.
Also, in order to ease the readability of the code, I've created 2 classes:
CsvMarshaller and CsvUnmarshaller. The CsvDataFormat delegates to them the
marshalling and unmarshalling processes. These classes also handle the
different meaning of the same option like "skipHeaderRecord". When
unmarshalling you're indicating that there is an header row that you don't
want to read. When marshalling you're indicating that the header row should
not be written.
I really hope that you'll find in this email the reasons behind the changes
you saw in the code.
Anyhow, If you want to review the code with me, I will be more that happy
to organize it using Skype.
Also, as you may have noticed it's a pretty big change and thus maybe it
should be in another data format module (csv1 ?).
Have a nice day,
Antoine.
2014-11-05 20:14 GMT+01:00 Claus Ibsen <cl...@gmail.com>:
> Hi
>
> Thanks. I think its a good time to migrate this in Camel 2.15. But the
> PR has merge conflicts now. Do you mind rebasing so the PR can be
> merged on master branch?
>
>
>
> On Wed, Oct 15, 2014 at 12:10 PM, Antoine DESSAIGNE
> <an...@gmail.com> wrote:
> > Hi everyone,
> >
> > I've submitted this PR https://github.com/apache/camel/pull/292 for
> > updating csv to the official version
> >
> > What do you think ?
> >
> > Have a nice day.
> >
> > Antoine.
> >
> > 2014-09-12 11:59 GMT+02:00 Antoine DESSAIGNE <
> antoine.dessaigne@gmail.com>:
> >
> >> Hi,
> >>
> >> I don't see how it can be compatible since the public API of
> CsvDataFormat
> >> relies on 2 classes that were deleted : CSVStrategy and CSVConfig.
> >>
> >> I'm also not convinced by a camel-csv2 component since it's common-csv
> >> version 1. I would rather have another component ("camel-csv-commons" ?)
> >> that publishes the <commons-csv> data format (or <csv-commons>).
> >>
> >> Which is the right way ?
> >>
> >> Also, I didn't managed to build the master. I keep having "Failed to
> >> execute goal
> >> org.apache.camel:camel-api-component-maven-plugin:2.15-SNAPSHOT:fromApis
> >> (generate-test-component-classes) on project camel-box: Error generating
> >> source for
> com.box.boxjavalibv2.resourcemanagers.IBoxCollaborationsManager:
> >> String index out of range: 1"
> >>
> >> What did I do wrong ?
> >>
> >> Thanks.
> >>
> >> Antoine
> >>
> >> 2014-09-11 21:56 GMT+02:00 Christian Müller <
> christian.mueller@gmail.com>:
> >>
> >>> +1
> >>>
> >>> I think a camel-csv2 component is not necessary.
> >>>
> >>> Best,
> >>>
> >>> Christian
> >>> -----------------
> >>>
> >>> Software Integration Specialist
> >>>
> >>> Apache Member
> >>> V.P. Apache Camel | Apache Camel PMC Member | Apache Camel committer
> >>> Apache Incubator PMC Member
> >>>
> >>> https://www.linkedin.com/pub/christian-mueller/11/551/642
> >>>
> >>> On Thu, Sep 11, 2014 at 2:30 AM, Willem Jiang <wi...@gmail.com>
> >>> wrote:
> >>>
> >>> > When I upgraded the http client version of camel-http4 to 4.3.x,
> there
> >>> > were some API changes in http client just like common-csv. As we can
> do
> >>> > some adaption inside camel-http4, I just changed the configuration a
> >>> bit.
> >>> > I think we can do the same thing in camel-csv :)
> >>> >
> >>> > --
> >>> > Willem Jiang
> >>> >
> >>> > Red Hat, Inc.
> >>> > Web: http://www.redhat.com
> >>> > Blog: http://willemjiang.blogspot.com (English)
> >>> > http://jnn.iteye.com (Chinese)
> >>> > Twitter: willemjiang
> >>> > Weibo: 姜宁willem
> >>> >
> >>> >
> >>> >
> >>> > On September 10, 2014 at 10:57:03 PM, Claus Ibsen (
> >>> claus.ibsen@gmail.com)
> >>> > wrote:
> >>> > > Hi
> >>> > >
> >>> > > Welcome back.
> >>> > >
> >>> > > I suspect the breakings is less affected by Camel users, eg I would
> >>> > > assume the data it creates on the Camel Exchange / Message / body /
> >>> > > headers can be the same as before? So the changes affected is maybe
> >>> > > how to configure it using the csv data format or its csv config
> >>> > > object?
> >>> > >
> >>> > > I am fine to make this change for Camel 2.15 and add some notes to
> the
> >>> > > release notes that end users need to do some changes.
> >>> > >
> >>> > >
> >>> > >
> >>> > > On Wed, Sep 10, 2014 at 4:46 PM, Antoine DESSAIGNE
> >>> > > wrote:
> >>> > > > Hi everyone,
> >>> > > >
> >>> > > > I was looking at this issue :
> >>> > > > https://issues.apache.org/jira/browse/CAMEL-7716.
> >>> > > >
> >>> > > > In fact, the commons-csv library changed a lot between the
> version
> >>> that
> >>> > > > camel uses and the official one. In addition, there is lot of
> nice
> >>> > stuff
> >>> > > > such as:
> >>> > > > * immutable configurations, it's no longer required to copy the
> >>> > > > configuration
> >>> > > > * build-in iterator mechanism
> >>> > > > * etc...
> >>> > > >
> >>> > > > Supporting the commons-csv API would break the compatibility with
> >>> > existing
> >>> > > > code and usage.
> >>> > > >
> >>> > > > What should I do ? Break compatibility ? Create a new data
> format ?
> >>> > > >
> >>> > > > Thanks a lot for your insights on this matter,
> >>> > > >
> >>> > > > Have a nice day,
> >>> > > >
> >>> > > > Antoine.
> >>> > > >
> >>> > > > PS: I was very busy, but now I managed to freed up time for
> camel :)
> >>> > >
> >>> > >
> >>> > >
> >>> > > --
> >>> > > Claus Ibsen
> >>> > > -----------------
> >>> > > Red Hat, Inc.
> >>> > > Email: cibsen@redhat.com
> >>> > > Twitter: davsclaus
> >>> > > Blog: http://davsclaus.com
> >>> > > Author of Camel in Action: http://www.manning.com/ibsen
> >>> > > hawtio: http://hawt.io/
> >>> > > fabric8: http://fabric8.io/
> >>> > >
> >>> >
> >>> >
> >>>
> >>
> >>
>
>
>
> --
> Claus Ibsen
> -----------------
> Red Hat, Inc.
> Email: cibsen@redhat.com
> Twitter: davsclaus
> Blog: http://davsclaus.com
> Author of Camel in Action: http://www.manning.com/ibsen
> hawtio: http://hawt.io/
> fabric8: http://fabric8.io/
>