; > isQuoting()
> > > > > > >
> > > > > > > they all do the same thing: test an ivar for null.
> > > > > > >
> > > > > > > So the simple thing to do is to make sure we have a good ivar
> > name
> > >
gt; > use
> > > > > > the JAXB inspired pattern: ivar foo has a test method called
> > > isFooSet()
> > > > > >
> > > > > > commentStart : Character
> > > > > > escape : Character
> > > > > > quoteChar : Char
x?
> > > > > Let's do that and match the method names.
> > > > >
> > > > > This also gives us room for String versions later,
> > > > getCommentStartString()
> > > > > for example, without breaking BC.
> > > > >
>
gt; > getCommentStartString()
> > > > for example, without breaking BC.
> > > >
> > > > For the with* methods, having the Character postfix does not make
> > sense,
> > > > since the type of the argument is in the signature. So
> > > >
> > > Now, that's better :-)
> > >
> > > My only question now is should "char delimiter" be "char
> delimiterChar"?
> > >
> > > See revision 1612352.
> > >
> > > Gary
> > >
> > >
> >
> >
> > >
> > > br,
> > > Benedikt
> > >
> > >
> > > 2014-07-20 14:04 GMT+02:00 Gary Gregory :
> > >
> > > > I like using all "get" methods and no "is" methods. It is simpl
It is simpler and
> > > makes the getters easier to access as a group with code completion IMO.
> > The
> > > with methods do not behave like Java bean method so I do not thing we
> > need
> > > to worry about that. Unless we want to register immutability.
ith methods do not behave like Java bean method so I do not thing we
> need
> > to worry about that. Unless we want to register immutability...
> >
> > Gary
> >
> > Original message From: Benedikt Ritter <
> > brit...@apache.org> Date:07/20/2014 04:02
f getters and setters in
> CSVFormat (was: Re: [VOTE] Release Commons CSV 1.0 based on RC1)
> using "get" for methods that return booleans is very uncommon imho...
>
> how about leaving all the gramme stuff out and use:
>
> void withSkipEmptyHeaders(boolean)
> boolea
Le 20/07/2014 22:35, Benedikt Ritter a écrit :
> Okay, but we can agree upon "allow missing column names" instead of "ignore
> empty headers"?
I'm ok for "allow missing column names" or "missing column names allowed".
Emmanuel Bourg
-
ike isActive, isEnabled, is closed
>
> Gary
>
> Original message From: Benedikt Ritter
> Date:07/20/2014 04:06 (GMT-05:00)
> To: Commons Developers List
> Subject: Re: [CSV] Rename isIgnoreEmptyHeaders? (was: Re: [VOTE]
> Release Commons CSV 1.0 based on R
Date:07/20/2014 04:06 (GMT-05:00)
To: Commons Developers List
Subject: Re: [CSV] Rename isIgnoreEmptyHeaders? (was: Re: [VOTE]
Release Commons CSV 1.0 based on RC1)
Any thoughts about this? I still think isAllowingEmptyColumnNames() is a
good name. :)
Benedikt
mmutability...
Gary
Original message From: Benedikt Ritter
Date:07/20/2014 04:02 (GMT-05:00)
To: Commons Developers List
Subject: Re: [CSV] Naming pattern of getters and setters in
CSVFormat (was: Re: [VOTE] Release Commons CSV 1.0 based on RC1)
using "get" for m
Any thoughts about this? I still think isAllowingEmptyColumnNames() is a
good name. :)
Benedikt
using "get" for methods that return booleans is very uncommon imho...
how about leaving all the gramme stuff out and use:
void withSkipEmptyHeaders(boolean)
boolean isSkipEmptyHeaders
that would
- restore symmetry between getter and setter
- almost follow JavaBean conventions (except for the "wi
Although i am not familiar with CSV's codebase, imho "get" is more straight
forward, so +1 to Gary's suggestion.
On Saturday, 19 July 2014, Gary Gregory wrote:
> On Sat, Jul 19, 2014 at 12:14 PM, Emmanuel Bourg > wrote:
>
> > Le 19/07/2014 13:48, Gary Gregory a écrit :
> >
> > > Can we go back
On Sat, Jul 19, 2014 at 12:14 PM, Emmanuel Bourg wrote:
> Le 19/07/2014 13:48, Gary Gregory a écrit :
>
> > Can we go back to use "get"?
>
> We are running in circles Gary, Benedikt and I, if others could weigh in
> that would help.
>
Circles, back and forth, to and fro, call it what you will. I
Le 19/07/2014 13:48, Gary Gregory a écrit :
> Can we go back to use "get"?
We are running in circles Gary, Benedikt and I, if others could weigh in
that would help.
Emmanuel Bourg
-
To unsubscribe, e-mail: dev-unsubscr...@com
Using *ing and *ed versions are worse IMO. *ed like
isHeaderRecordSkipped makes it sound like we are asking if something has
happened in the past. *ing (
isSkippingHeaderRecord) makes is sounds like you are controlling something
that is happening right now.
I think it might make the getter methods
... moving this to a new thread.
so how do we reach consensus here? :o)
if(format.isSkippingHeaderRecord()) {
...
}
sounds better to me than
if(format.isHeaderRecordSkipped) {
...
}
Gary made up a good argument about using active voice in method names. How
ever there will be the asymmetry b
... moving this to a new thread.
The name isIgnoreEmptyHeaders seems to cause confusion, since in this case
an entry in the header is meant and not the whole header (like in
skipHeaderRecord)
The JavaDoc of isIgnoreEmptyHeaders is:
/**
* Specifies whether empty headers are ignored when
On Fri, Jul 18, 2014 at 4:36 PM, Emmanuel Bourg wrote:
> Le 18/07/2014 22:17, Benedikt Ritter a écrit :
> > Just to make it clear... We're happy with:
> > - boolean isIgnoringSurroundingSpaces()
> > - withIgnoreSurroundingSpaces(boolean)
> >
> > So nothing has to change?
>
> Not sure, I find the
Le 18/07/2014 22:17, Benedikt Ritter a écrit :
> Just to make it clear... We're happy with:
> - boolean isIgnoringSurroundingSpaces()
> - withIgnoreSurroundingSpaces(boolean)
>
> So nothing has to change?
Not sure, I find the asymmetry a bit disturbing, but if I'm the only one
I can pass on that.
On Fri, Jul 18, 2014 at 4:01 PM, Benedikt Ritter
wrote:
> I think "isAllowingMissingColumnNames" is a good name.
>
That does not real well at all unfortunately.
Gary
>
> Send from my mobile device
>
> > Am 18.07.2014 um 17:43 schrieb Emmanuel Bourg :
> >
> > Le 18/07/2014 17:37, Gary Gregory
Le 18/07/2014 22:05, Benedikt Ritter a écrit :
> While looking at CSVFormat again, I noticed that the methods currently are
> withQuotePolicy and getQuotePolicy. Now I'm feeling we should go with the
> more verbose but more explicit name for that enum.
Ok. We could use QuoteMode instead of Quot
Just to make it clear... We're happy with:
- boolean isIgnoringSurroundingSpaces()
- withIgnoreSurroundingSpaces(boolean)
So nothing has to change?
Benedikt
Send from my mobile device
> Am 18.07.2014 um 16:12 schrieb Gary Gregory :
>
> There is a nice pattern now with ivar and method names whe
Send from my mobile device
> Am 15.07.2014 um 20:07 schrieb Benedikt Ritter :
>
> Hello Emmanuel,
>
> thanks for your thorough review. I've worked through your list. Here are my
> comments:
>
>
> 2014-07-12 19:02 GMT+02:00 Emmanuel Bourg :
>> I took a fresh look at the API and here is my re
I think "isAllowingMissingColumnNames" is a good name.
Send from my mobile device
> Am 18.07.2014 um 17:43 schrieb Emmanuel Bourg :
>
> Le 18/07/2014 17:37, Gary Gregory a écrit :
>
>> So what do you call the individual names in a header record?
>
> A column name. That's the term used in CSVRe
Le 18/07/2014 17:37, Gary Gregory a écrit :
> So what do you call the individual names in a header record?
A column name. That's the term used in CSVRecord.
Emmanuel Bourg
-
To unsubscribe, e-mail: dev-unsubscr...@commons.apac
On Fri, Jul 18, 2014 at 11:35 AM, Emmanuel Bourg wrote:
> Le 18/07/2014 17:29, Gary Gregory a écrit :
>
> > There is indeed one header record. But that record contains one header
> per
> > column, so one could conceivably says that there are one or more headers
> > per record, just like there is
Le 18/07/2014 17:29, Gary Gregory a écrit :
> There is indeed one header record. But that record contains one header per
> column, so one could conceivably says that there are one or more headers
> per record, just like there is one or more value per data record.
That's what I found confusing. Fr
On Fri, Jul 18, 2014 at 11:20 AM, Emmanuel Bourg wrote:
> Le 18/07/2014 16:13, Gary Gregory a écrit :
> > Would using "Headers" (plural) be better: withSkipHeadersRecord(boolean)
> ?
>
> Isn't the plural form withSkipHeaderRecords() instead?
>
> But since there is only one header record I'd keep
Le 18/07/2014 16:13, Gary Gregory a écrit :
> Would using "Headers" (plural) be better: withSkipHeadersRecord(boolean) ?
Isn't the plural form withSkipHeaderRecords() instead?
But since there is only one header record I'd keep the singular form.
Emmanuel Bourg
-
Would using "Headers" (plural) be better: withSkipHeadersRecord(boolean) ?
Gary
On Fri, Jul 18, 2014 at 7:54 AM, Emmanuel Bourg wrote:
> Thank you for the fixes Benedikt.
>
> Le 15/07/2014 20:07, Benedikt Ritter a écrit :
>
> > skipHeaderRecord refers to the header records as a whole (so it's
There is a nice pattern now with ivar and method names where we have the
ivar ignoreSomething (note that it is verb first in the active void) and
withIgnoreSomething.
It is important IMO to use the active void instead of passive
(somethingIgnored) to make it clear who performs the action.
It's wi
Thank you for the fixes Benedikt.
Le 15/07/2014 20:07, Benedikt Ritter a écrit :
> skipHeaderRecord refers to the header records as a whole (so it's
> singular). ignore empty headers refers to header column values, so it's
> plural. I guess that makes sense.
Ok, I misunderstood the intent then.
Hello Emmanuel,
thanks for your thorough review. I've worked through your list. Here are my
comments:
2014-07-12 19:02 GMT+02:00 Emmanuel Bourg :
> I took a fresh look at the API and here is my review based on the
> Javadoc and the content of the site only:
>
> - The Javadoc for CSVFormat menti
Le 13/07/2014 14:01, Gary Gregory a écrit :
> I do not think we should bother with Java 5.
>
> For me the discussion should be whether we require Java 6 or 7.
+1 for Java 7
Emmanuel Bourg
-
To unsubscribe, e-mail: dev-unsubscr
I do not think we should bother with Java 5.
For me the discussion should be whether we require Java 6 or 7.
Gary
Original message From: Adrian Crum
Date:07/13/2014 05:27
(GMT-05:00) To: Commons Developers List
Subject: Re: [CANCEL][VOTE] Release Commons CSV 1.0 based on
I would be inclined to leave Java 5 compatibility to the few (if any)
people who still use it. They have access to the CSV source code and
they can modify it to make it Java 5 compatible.
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 7/13/2014 10:30 AM, Benedikt Ritter wrote:
2
2014-07-12 3:17 GMT+02:00 Gary Gregory :
> +1
>
> Apache Maven 3.2.2 (45f7c06d68e745d05611f7fd14efb6594181933e;
> 2014-06-17T09:51:42-04:00)
> Maven home: C:\Java\apache-maven-3.2.2
> Java version: 1.7.0_60, vendor: Oracle Corporation
> Java home: C:\Program Files\Java\jdk1.7.0_60\jre
> Default lo
2014-07-13 11:27 GMT+02:00 Adrian Crum :
> Why would we target Java 5? It is long past end-of-life.
Well I don't have a strong preference here. The reason people bring up for
targeting older Java versions is, to reach a broader user group. I'm fine
with targeting Java 7.
>
>
> Adrian Crum
> Sa
Why would we target Java 5? It is long past end-of-life.
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 7/13/2014 10:12 AM, Benedikt Ritter wrote:
Hi all,
this vote it canceled to address the various issues identified. I'll need
some time to work through the list, so I expect the
Hi all,
this vote it canceled to address the various issues identified. I'll need
some time to work through the list, so I expect the next RC by mid of the
week (help is always welcome ;-).
Till then I'd like to hear comments regarding the Java version to target.
>From what I've seen it looks lik
So that means another RC which is fine, since we have a long list of work items
from emmanuel email.
Original message From: Jörg Schaible
Date:07/12/2014 14:05 (GMT-05:00)
To: dev@commons.apache.org Subject: Re: [VOTE] Release
Commons CSV 1.0 based on RC1
Gary Gregory
I'm fine with Java 7. YMMV.
Gary
Original message From: Emmanuel Bourg
Date:07/12/2014 13:22 (GMT-05:00)
To: dev@commons.apache.org Subject: Re: [VOTE] Release
Commons CSV 1.0 based on RC1
Le 12/07/2014 17:57, Jörg Schaible a écrit :
> According release notes J
[VOTE] Release Commons CSV 1.0 based on RC1
I took a fresh look at the API and here is my review based on the
Javadoc and the content of the site only:
- The Javadoc for CSVFormat mentions a parseFile() method, but this
method doesn't exist.
- CSVFormat.DEFAULT states that empty lines are "
Gary Gregory wrote:
> Let's change the docs then.
The release notes are included in the artifacts.
- Jörg
-
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.or
Emmanuel Bourg wrote:
> Le 12/07/2014 17:57, Jörg Schaible a écrit :
>
>> According release notes Java 5 is supported, but it uses already methods
>> available in Java 6:
>
> It's even Java 7, CSVParser implements AutoCloseable.
I did no code analysis, but my Java 6 JDKs could compile and run t
Le 12/07/2014 17:57, Jörg Schaible a écrit :
> According release notes Java 5 is supported, but it uses already methods
> available in Java 6:
It's even Java 7, CSVParser implements AutoCloseable.
Emmanuel Bourg
-
To unsubscr
Let's change the docs then.
Gary
Original message From: Jörg Schaible
Date:07/12/2014 11:57 (GMT-05:00)
To: dev@commons.apache.org Subject: Re: [VOTE] Release
Commons CSV 1.0 based on RC1
Hi Benedikt,
-1
According release notes Java 5 is supported, but it uses al
I took a fresh look at the API and here is my review based on the
Javadoc and the content of the site only:
- The Javadoc for CSVFormat mentions a parseFile() method, but this
method doesn't exist.
- CSVFormat.DEFAULT states that empty lines are "allowed". Does it mean
the empty lines are ignored
Hi Benedikt,
-1
According release notes Java 5 is supported, but it uses already methods
available in Java 6:
== %< ===
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-
plugin:3.1:compile (default-compile) on project commons
+1
Apache Maven 3.2.2 (45f7c06d68e745d05611f7fd14efb6594181933e;
2014-06-17T09:51:42-04:00)
Maven home: C:\Java\apache-maven-3.2.2
Java version: 1.7.0_60, vendor: Oracle Corporation
Java home: C:\Program Files\Java\jdk1.7.0_60\jre
Default locale: en_US, platform encoding: Cp1252
OS name: "windows
We have worked a long time on CSV and more an more people seem to be using
the SNAPSHOT since there is no stable release. Although we still have some
outstanding issues regarding some corner cases and uncommon formats, It
feels like the API as reached a solid state. So I'd like to finally release
55 matches
Mail list logo
