[gentoo-dev] vim spell files

[gentoo-dev] vim spell files

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 2166 bytes --]

Vim 7 includes a native spellchecker. It uses its own spell file
format which can be created using Myspell *.aff/*.dic files (the same
as for Mozilla and OpenOffice). The spell files are rather large, so
it's not feasible to bundle them with vim unconditionally.

Building a spell file from source is tricky and requires huge amounts
of RAM and some fairly obscure dependencies. Upstream ship pre-built
spell files, and I believe it would make most sense for us to use
those.

Note that spell files have to be generated for each internal character
set that may be used. This generally means utf-8 (used by gtk2/qt gvim
internally no matter what) and the appropriate iso-8859- code for the
country in question.

As it stands, English is installed with vim-core. I might remove this
if spell checking ends up as an optional feature. For the rest, I'm
proposing separate packages, one per language, for the different spell
files. Character set decisions will be left up to upstream, and the
user will not get to restrict this further.

I don't intend to maintain these myself (I speak a couple of other
languages well enough to verify that the package works, but I'd rather
not get involved any further), but I'd like to make it easy for
non-Vim-herd people to do so. Naming-wise, vim-spell-${two letter
language code} seems sanest.

Proposed vim-spell.eclass:

    http://dev.gentoo.org/~ciaranm/tmp/vim-spell.eclass

It includes instructions on how to make a spell source package. The
ebuilds should be extremely simple -- only VIM_SPELL_LANGUAGE, LICENSE,
KEYWORDS and IUSE need to be set, along with one inherit line.

Questions:

* app-vim or app-dicts for the spell files?

* Will there be sufficient interest from other developers to make this
worthwhile?

* Is anyone going to get upset about the character set decisions?

* Does anyone think we should make users build spell lists from source,
even though it can require several GBytes of RAM?

-- 
Ciaran McCreesh : Gentoo Developer (Vim, Shell tools, Fluxbox, Cron)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] vim spell files

[Luca Barbato]

Ciaran McCreesh wrote:
> Vim 7 includes a native spellchecker. It uses its own spell file
> format which can be created using Myspell *.aff/*.dic files (the same
> as for Mozilla and OpenOffice). The spell files are rather large, so
> it's not feasible to bundle them with vim unconditionally.

are them interchangeable?

> 
> Building a spell file from source is tricky and requires huge amounts
> of RAM and some fairly obscure dependencies. Upstream ship pre-built
> spell files, and I believe it would make most sense for us to use
> those.

I second that.

> 
> Note that spell files have to be generated for each internal character
> set that may be used. This generally means utf-8 (used by gtk2/qt gvim
> internally no matter what) and the appropriate iso-8859- code for the
> country in question.
> 

That should cover most uses.


> Questions:
> 
> * app-vim or app-dicts for the spell files?

app-dicts if they can be non vim only...

> 
> * Will there be sufficient interest from other developers to make this
> worthwhile?

if is vim only I'm not sure.

> 
> * Is anyone going to get upset about the character set decisions?
> 

As usual I'm afraid...

> * Does anyone think we should make users build spell lists from source,
> even though it can require several GBytes of RAM?

Only if requested by user (think about an useflag)

lu

-- 

Luca Barbato

Gentoo/linux Developer		Gentoo/PPC Operational Leader
http://dev.gentoo.org/~lu_zero

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] vim spell files

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 1122 bytes --]

On Fri, 23 Sep 2005 10:12:27 +0200 Luca Barbato <lu_zero@gentoo.org>
wrote:
| Ciaran McCreesh wrote:
| > Vim 7 includes a native spellchecker. It uses its own spell file
| > format which can be created using Myspell *.aff/*.dic files (the
| > same as for Mozilla and OpenOffice). The spell files are rather
| > large, so it's not feasible to bundle them with vim unconditionally.
| 
| are them interchangeable?

Nope. The format's vim-specific. Vim upstream consider other formats to
be too inefficient for on the fly spellchecking in an editor where you
can make extremely quick changes to a lot of text.

| > * Does anyone think we should make users build spell lists from
| > source, even though it can require several GBytes of RAM?
| 
| Only if requested by user (think about an useflag)

USE flags control things that affect the generated content. Building
spell files from source every time always produces the same output.

-- 
Ciaran McCreesh : Gentoo Developer (Vim, Shell tools, Fluxbox, Cron)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] vim spell files

[Ioannis Aslanidis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Ciaran McCreesh wrote:
> | > * Does anyone think we should make users build spell lists from
> | > source, even though it can require several GBytes of RAM?
> | 
> | Only if requested by user (think about an useflag)
> 
> USE flags control things that affect the generated content. Building
> spell files from source every time always produces the same output.
> 

Then the most logical would be to use a precompiled spell file.

- --
Ioannis Aslanidis

Hellenic Gentoo (http://hellenicgentoo.sf.net)
Gentoo Forums (http://forums.gentoo.org)
-----BEGIN PGP SIGNATURE-----

iD8DBQFDNC9gGiwEn1Zp/wERApI3AJ9a51rkRcAxOlv9arX7kB9MLP1KLwCg5jXy
IvTgf28dNdQoY+LTC/Xqf9g=
=hUEC
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] vim spell files

[Ioannis Aslanidis]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



Ciaran McCreesh wrote:
> I don't intend to maintain these myself (I speak a couple of other
> languages well enough to verify that the package works, but I'd rather
> not get involved any further), but I'd like to make it easy for
> non-Vim-herd people to do so. Naming-wise, vim-spell-${two letter
> language code} seems sanest.

> * Will there be sufficient interest from other developers to make this
> worthwhile?

> 
> * Is anyone going to get upset about the character set decisions?

I'm really interested in a Greek dictionary. That would prove very
useful. About the character set, as far as it doesn't break documents,
there would be no problem.

> * Does anyone think we should make users build spell lists from source,
> even though it can require several GBytes of RAM?

Instad of the use flag thing, I'd first ask if there is any sort of
costumization on the dictionary. If there are no customization options,
I see it pointless to have to build it from source.

- --
Ioannis Aslanidis

Hellenic Gentoo (http://hellenicgentoo.sf.net)
Gentoo Forums (http://forums.gentoo.org)
-----BEGIN PGP SIGNATURE-----

iD8DBQFDM8r/GiwEn1Zp/wERAtKUAKD0a/DGGojjxKZYIaeztEkpktMRRwCg0RXd
1UThSTGmvEBayEzLCmvP/Bs=
=NuwV
-----END PGP SIGNATURE-----
-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] vim spell files

[Lares Moreau]

> * Does anyone think we should make users build spell lists from source,
> even though it can require several GBytes of RAM?

A Use flag and a warning message with a 5 second delay might be
adviseable.  Some users add USE flags with out worrying about what 
resources it may take.


-Lares

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] vim spell files

[Michael Kohl]

[-- Attachment #1.1: Type: text/plain, Size: 1233 bytes --]

On Fri, 23 Sep 2005 02:20:21 +0100
Ciaran McCreesh <ciaranm@gentoo.org> wrote:

> Questions:
> 
> * app-vim or app-dicts for the spell files?

As they only work with vim, I think they should go into app-vim.
 
> * Will there be sufficient interest from other developers to make this
> worthwhile?

I have finished vim-spell-de (both UTF-8 and Latin1 work like a charm
btw) and could also do others. Since you haven't commited the eclass
yet, I haven't added the package to the tree so far.

> * Is anyone going to get upset about the character set decisions?

Someone will always get upset about something. If the decision is sane,
I wouldn't care about those people too much though.
 
> * Does anyone think we should make users build spell lists from
> source, even though it can require several GBytes of RAM?

No, not really.

Cheers,
Michael


-- 
www.cargal.org 
GnuPG-key-ID: 0x90CA09E3
Jabber-ID: citizen428 [at] cargal [dot] org
Registered Linux User #278726


-- 
web@citizen428.net                                 citizen428@gentoo.org
http://citizen428.net/                http://dev.gentoo.org/~citizen428/
GnuPG key: 0x90CA09E3/4D21 916E DBCE 72B8 CDC5  BD87 DE2D 91A2 90CA 09E3

[-- Attachment #1.2: . --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

[gentoo-dev] Re: vim spell files

[Dan Meltzer]

On 9/22/05, Ciaran McCreesh <ciaranm@gentoo.org> wrote:
<snip>
> * Does anyone think we should make users build spell lists from source,
> even though it can require several GBytes of RAM?
Of course! simply talk to hardware companies, and get them to kick
back memory from RAM upgrades that everyone buys, everyone benefits!
>
> --
> Ciaran McCreesh : Gentoo Developer (Vim, Shell tools, Fluxbox, Cron)
> Mail            : ciaranm at gentoo.org
> Web             : http://dev.gentoo.org/~ciaranm
>
>
>

-- 
gentoo-dev@gentoo.org mailing list

Re: [gentoo-dev] vim spell files

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 839 bytes --]

On Fri, 23 Sep 2005 02:20:21 +0100 Ciaran McCreesh <ciaranm@gentoo.org>
wrote:
| * app-vim or app-dicts for the spell files?

I was persuaded to go with app-vim, since it's more specific.

| * Will there be sufficient interest from other developers to make this
| worthwhile?

Looks like it.

The eclass is now in the tree, as is app-vim/vim-spell-en. I decided to
split out the English dictionary because it's over a meg of unnecessary
bloat otherwise...

Feel free to add other languages. There are instructions in the eclass.
Don't forget to package.mask your package (put it below the existing
vim7 entries please) and mail vim@ when you're done.

-- 
Ciaran McCreesh : Gentoo Developer (Vim, Shell tools, Fluxbox, Cron)
Mail            : ciaranm at gentoo.org
Web             : http://dev.gentoo.org/~ciaranm


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]