You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@corinthia.apache.org by jan i <ja...@apache.org> on 2015/01/18 20:35:14 UTC

Apache Commons is not really an option in corinthia

Hi

Since apache commons was mentioned as a possible solution to several
issues, I checked if my memory failed me.

Apache Commons seems to only deliver java code, and can therefore not be
used in corinthia, we need a C library.

rgds
jan I.

Re: Apache Commons is not really an option in corinthia

Posted by Peter Kelly <ke...@gmail.com>.
> On 19 Jan 2015, at 3:27 am, jan i <ja...@apache.org> wrote:
> 
> On 18 January 2015 at 21:18, Dennis E. Hamilton <de...@acm.org>
> wrote:
> 
>> Yes, I keep pointing out that it is the methods and design, not the code
>> that I am interested in, but I think I am not being emphatic enough when I
>> do that [;<).
>> 
> I know you said that, but at the same time somebody said the word license.
> When I look at methods and design I do not need to care about the license,
> because I
> have to reimplement it anyhow.
> 
> 
>> 
>> It also provides a way that a transposition to a different programming
>> language is correct [;<), by having tests that succeed for both.
>> 
> +1
> 
>> 
>> Likewise, I am perfectly willing to mine POI and the ODF Toolkit for
>> whatever they have to offer in this respect (although this may apply more
>> to the Zip case than the long-form getopt).
>> 
> I am close with zip now, so no need for that. But getopt_long() must be
> available in one of the C projects.

I think we should consider whether the getopt-style syntax or a git-style syntax [1] would be most suitable. I’ve recently been using the Click library for Python in another project I’m working on, and it provides a nice way of providing a hierarchical set of commands, in addition to options for said commands.

The discussion is perhaps a bit premature though until we decide what command-line tools we actually want to have. We’ve got one for command-line conversion (and the mixed-bad dfutil which is kind of a collection of various things which I added in a very ad-hoc way) but perhaps we might want to have other tools, or “subcommands” of dfconvert. Once we know exactly what we’re building, it’ll be easier to decide on a suitable command-line syntax :)

[1] By this, I mean “git commit” instead of “git —commit”. Dashes are used for options, but not sub-commands.

--
Dr. Peter M. Kelly
kellypmk@gmail.com
http://www.kellypmk.net/

PGP key: http://www.kellypmk.net/pgp-key <http://www.kellypmk.net/pgp-key>
(fingerprint 5435 6718 59F0 DD1F BFA0 5E46 2523 BAA1 44AE 2966)


Re: Apache Commons is not really an option in corinthia

Posted by jan i <ja...@apache.org>.
On 18 January 2015 at 22:22, Dave Fisher <da...@comcast.net> wrote:

>
> On Jan 18, 2015, at 12:27 PM, jan i wrote:
>
> > On 18 January 2015 at 21:18, Dennis E. Hamilton <dennis.hamilton@acm.org
> >
> > wrote:
> >
> >> Yes, I keep pointing out that it is the methods and design, not the code
> >> that I am interested in, but I think I am not being emphatic enough
> when I
> >> do that [;<).
> >>
> > I know you said that, but at the same time somebody said the word
> license.
> > When I look at methods and design I do not need to care about the
> license,
> > because I
> > have to reimplement it anyhow.
>
> Take a look at Apache APR -
> http://apr.apache.org/docs/apr/1.5/group__apr__getopt.html
>
> Lots of other portable goodies.
>
> When I mentioned Commons I was thinking APR!
>

and I forgot about APR and only thoughts of big commons. Actually quite
funny because I use APR in quite a number of projects.

A short look at the getopt looks promising.

rgds
jan I.


>
> Regards,
> Dave
>
> >
> >
> >>
> >> It also provides a way that a transposition to a different programming
> >> language is correct [;<), by having tests that succeed for both.
> >>
> > +1
> >
> >>
> >> Likewise, I am perfectly willing to mine POI and the ODF Toolkit for
> >> whatever they have to offer in this respect (although this may apply
> more
> >> to the Zip case than the long-form getopt).
> >>
> > I am close with zip now, so no need for that. But getopt_long() must be
> > available in one of the C projects.
> >
> > If you only search for the methods and design, have a look at the
> > opensource C variants on github (or the official one from AT&T).
> >
> > rgds
> > jan i.
> >
> >
> >>
> >> -----Original Message-----
> >> From: jan i [mailto:jani@apache.org]
> >> Sent: Sunday, January 18, 2015 11:35
> >> To: dev@corinthia.incubator.apache.org
> >> Subject: Apache Commons is not really an option in corinthia
> >>
> >> Hi
> >>
> >> Since apache commons was mentioned as a possible solution to several
> >> issues, I checked if my memory failed me.
> >>
> >> Apache Commons seems to only deliver java code, and can therefore not be
> >> used in corinthia, we need a C library.
> >>
> >> rgds
> >> jan I.
> >>
> >>
>
>

RE: Apache Commons is not really an option in corinthia

Posted by "Dennis E. Hamilton" <de...@acm.org>.
I'm going to clip some of this into JIRA issues where I might find it again [;<).

 -- replying below to --
From: Dave Fisher [mailto:dave2wave@comcast.net] 
Sent: Sunday, January 18, 2015 13:23
To: dev@corinthia.incubator.apache.org
Cc: Dennis Hamilton
Subject: Re: Apache Commons is not really an option in corinthia

[ ... ]

Take a look at Apache APR - http://apr.apache.org/docs/apr/1.5/group__apr__getopt.html

Lots of other portable goodies.

When I mentioned Commons I was thinking APR!

<orcmid>
	I was thinking of Commons CLI.    It is the data structure
     and the rules for it that "steer" the command-line parser (and also 
     provide the brief help page) that interest me.  I like Commons CLI
     for how it handles the line and think that perhaps that structure is much 
     easier to localize too.  (The usage can be localized but I would be
     cagy about localizing the option words.)

     APR is all C.  I had been looking for that but I forgot the name of it.
     I think I was turned off the first time I looked at it because it redoes
     all of the memory allocation and various dynamic-memory structures.
     I am willing to take another look.  It is designed to do what Corinthia
     wants out of "Platform" but has gone deeper into what is involved, I'd
     say.  The platforms that APR has worked with are worth checking on.

     I suspect that getting a Portable C (i.e., standard) version is more
     likely to be found in via FreeBSD or OpenBSD.

     Side Question: Does POI use Apache Commons Compress and if so is that
     API found very suitable?.
</orcmid>

Regards,
Dave

[ ... ]


Re: Apache Commons is not really an option in corinthia

Posted by Dave Fisher <da...@comcast.net>.
On Jan 18, 2015, at 12:27 PM, jan i wrote:

> On 18 January 2015 at 21:18, Dennis E. Hamilton <de...@acm.org>
> wrote:
> 
>> Yes, I keep pointing out that it is the methods and design, not the code
>> that I am interested in, but I think I am not being emphatic enough when I
>> do that [;<).
>> 
> I know you said that, but at the same time somebody said the word license.
> When I look at methods and design I do not need to care about the license,
> because I
> have to reimplement it anyhow.

Take a look at Apache APR - http://apr.apache.org/docs/apr/1.5/group__apr__getopt.html

Lots of other portable goodies.

When I mentioned Commons I was thinking APR!

Regards,
Dave

> 
> 
>> 
>> It also provides a way that a transposition to a different programming
>> language is correct [;<), by having tests that succeed for both.
>> 
> +1
> 
>> 
>> Likewise, I am perfectly willing to mine POI and the ODF Toolkit for
>> whatever they have to offer in this respect (although this may apply more
>> to the Zip case than the long-form getopt).
>> 
> I am close with zip now, so no need for that. But getopt_long() must be
> available in one of the C projects.
> 
> If you only search for the methods and design, have a look at the
> opensource C variants on github (or the official one from AT&T).
> 
> rgds
> jan i.
> 
> 
>> 
>> -----Original Message-----
>> From: jan i [mailto:jani@apache.org]
>> Sent: Sunday, January 18, 2015 11:35
>> To: dev@corinthia.incubator.apache.org
>> Subject: Apache Commons is not really an option in corinthia
>> 
>> Hi
>> 
>> Since apache commons was mentioned as a possible solution to several
>> issues, I checked if my memory failed me.
>> 
>> Apache Commons seems to only deliver java code, and can therefore not be
>> used in corinthia, we need a C library.
>> 
>> rgds
>> jan I.
>> 
>> 


Re: Apache Commons is not really an option in corinthia

Posted by jan i <ja...@apache.org>.
On 18 January 2015 at 21:18, Dennis E. Hamilton <de...@acm.org>
wrote:

> Yes, I keep pointing out that it is the methods and design, not the code
> that I am interested in, but I think I am not being emphatic enough when I
> do that [;<).
>
I know you said that, but at the same time somebody said the word license.
When I look at methods and design I do not need to care about the license,
because I
have to reimplement it anyhow.


>
> It also provides a way that a transposition to a different programming
> language is correct [;<), by having tests that succeed for both.
>
+1

>
> Likewise, I am perfectly willing to mine POI and the ODF Toolkit for
> whatever they have to offer in this respect (although this may apply more
> to the Zip case than the long-form getopt).
>
I am close with zip now, so no need for that. But getopt_long() must be
available in one of the C projects.

If you only search for the methods and design, have a look at the
opensource C variants on github (or the official one from AT&T).

rgds
jan i.


>
> -----Original Message-----
> From: jan i [mailto:jani@apache.org]
> Sent: Sunday, January 18, 2015 11:35
> To: dev@corinthia.incubator.apache.org
> Subject: Apache Commons is not really an option in corinthia
>
> Hi
>
> Since apache commons was mentioned as a possible solution to several
> issues, I checked if my memory failed me.
>
> Apache Commons seems to only deliver java code, and can therefore not be
> used in corinthia, we need a C library.
>
> rgds
> jan I.
>
>

Re: Apache Commons is not really an option in corinthia

Posted by Peter Kelly <ke...@gmail.com>.
> On 19 Jan 2015, at 3:18 am, Dennis E. Hamilton <de...@acm.org> wrote:
> 
> Yes, I keep pointing out that it is the methods and design, not the code that I am interested in, but I think I am not being emphatic enough when I do that [;<).
> 
> It also provides a way that a transposition to a different programming language is correct [;<), by having tests that succeed for both.
> 
> Likewise, I am perfectly willing to mine POI and the ODF Toolkit for whatever they have to offer in this respect (although this may apply more to the Zip case than the long-form getopt).

Most definitely - in fact I’m envisioning a port of (parts of) POI to C would be extremely helpful to us.

--
Dr. Peter M. Kelly
kellypmk@gmail.com
http://www.kellypmk.net/

PGP key: http://www.kellypmk.net/pgp-key <http://www.kellypmk.net/pgp-key>
(fingerprint 5435 6718 59F0 DD1F BFA0 5E46 2523 BAA1 44AE 2966)


RE: Apache Commons is not really an option in corinthia

Posted by "Dennis E. Hamilton" <de...@acm.org>.
Yes, I keep pointing out that it is the methods and design, not the code that I am interested in, but I think I am not being emphatic enough when I do that [;<).

It also provides a way that a transposition to a different programming language is correct [;<), by having tests that succeed for both.

Likewise, I am perfectly willing to mine POI and the ODF Toolkit for whatever they have to offer in this respect (although this may apply more to the Zip case than the long-form getopt).

-----Original Message-----
From: jan i [mailto:jani@apache.org] 
Sent: Sunday, January 18, 2015 11:35
To: dev@corinthia.incubator.apache.org
Subject: Apache Commons is not really an option in corinthia

Hi

Since apache commons was mentioned as a possible solution to several
issues, I checked if my memory failed me.

Apache Commons seems to only deliver java code, and can therefore not be
used in corinthia, we need a C library.

rgds
jan I.