Adding ttf-sil-abyssinica-udeb to packages used in the GTK installer? - Debian

This is a discussion on Adding ttf-sil-abyssinica-udeb to packages used in the GTK installer? - Debian ; Frans Pop writes: > On Friday 11 January 2008, Eddy Petrișor wrote: >> No, in the font udeb build. > > No. That completely sucks! It would mean that you'd NEVER really know if all > characters needed > > ...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 33 of 33

Thread: Adding ttf-sil-abyssinica-udeb to packages used in the GTK installer?

  1. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    Frans Pop writes:

    > On Friday 11 January 2008, Eddy Petrișor wrote:
    >> No, in the font udeb build.

    >
    > No. That completely sucks! It would mean that you'd NEVER really know if all
    > characters needed
    >
    > Especially with Christian's comment that we display a fair number of strings
    > that are not even in D-I SVN, this whole idea gets a huge NACK from me as
    > it is.


    Same from my side. It'll be very difficult to know if there's no
    missing symbols and like.

    --
    O T A V I O S A L V A D O R
    ---------------------------------------------
    E-mail: otavio@debian.org UIN: 5906116
    GNU/Linux User: 239058 GPG ID: 49A5F855
    Home Page: http://otavio.ossystems.com.br
    ---------------------------------------------
    "Microsoft sells you Windows ... Linux gives
    you the whole house."

  2. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    Quoting Eddy Petri?or (eddy.petrisor@gmail.com):

    > Most fonts include at least:
    > - the latin ranges (basic + latin A, in most cases)
    > - symbols like (and more) : !@#$%^&*()_+{}:"?><,./;'[]=-`~ - many
    > of these not used
    >
    > Sometimes they contain other ranges, too, so I suspect it would be
    > something (remember "The udeb size goes from 220768 bytes to 183090
    > bytes"? ;-) .



    Well, if you're talking about stripping the Latin range from all fonts
    but Freefont (and the CJK ones), this is something that's being
    considered.

    My understanding was that you were proposing to strip out unused
    glyphs from the supported languages. For instance, if a given glyph is
    not used by Amharic translations, it would be removed.

    *that*, I would object to.....stripping the Latin range, I of course
    don't object.



    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHhwJR1OXtrMAUPS0RAtRHAJ9/O6gMSbtA70JPa3je3kyQYL2F9ACgk4Cs
    ZKANW3cUtbB7aJrG/iR8phY=
    =fZ/m
    -----END PGP SIGNATURE-----


  3. Re: [RFC] Stripping Latin range in fonts used by g-i installer?


    On Fri, Jan 11, 2008 at 01:29:51AM +0100, Frans Pop wrote:
    > Especially with Christian's comment that we display a fair number of strings
    > that are not even in D-I SVN, this whole idea gets a huge NACK from me as
    > it is.


    and in the input forms you can type unpredictable chars


    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHhx7cax2slmJA7HoRAg3HAKDueHB1c7XqQoKuXQ0zBx KilxcT5ACggUT/
    2WFnPJlvZCyr6zucZAUgalE=
    =tlRO
    -----END PGP SIGNATURE-----


  4. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    Quoting Christian Perrier (bubulle@debian.org):

    > Im waiting for some ACK by others to propose the strip patch to the
    > font maintainer.
    >
    > I propose ttf-sil-abyssinica as a start, with some attempt to make the
    > "strip" patch less dependent from the package than it is in
    > ttf-freefont (many things are hardcoded there). I may need some help
    > by people more skilled in well-written shell scripting than /me



    I just reported a bug against ttf-sil-abyssinica source package, with
    an attached patch, to propose stripping out the Latin range in the
    udeb.

    In case this proves acceptable, I think we could extend this to other
    specific fonts.



    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHjQ2e1OXtrMAUPS0RAkHdAKCe/2KAKLZyFxRoxFfZ5wcOQmttvQCfedY0
    jORH3a7G7kgtj8hpxaokh1g=
    =FsMZ
    -----END PGP SIGNATURE-----


  5. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    Hi,

    On Tue, Jan 15, 2008 at 08:46:38PM +0100, Christian Perrier wrote:
    > Quoting Christian Perrier (bubulle@debian.org):
    >
    > I just reported a bug against ttf-sil-abyssinica source package, with
    > an attached patch, to propose stripping out the Latin range in the
    > udeb.


    I've created a test image [1] after applying the above patch; quckly tried it with qemu
    a took a screenshot [2].
    Would be nice to have some feedback from original language speakers

    regards,
    Davide

    [1] http://www.alioth.debian.org/~zinosat-guest/mini.iso
    [2] http://www.alioth.debian.org/~zinosat-guest/am_gi.png

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHjStUax2slmJA7HoRApaBAJ9T8XVlRqs3FCnPcswuCp sEQDRDDACeJIGt
    Kk5eERQC0aXsIpkGdehM3Bo=
    =xuQ0
    -----END PGP SIGNATURE-----


  6. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    On Wednesday 16 January 2008, Praveen A wrote:
    > 2008/1/10, Christian Perrier :
    > > My understanding was that you were proposing to strip out unused
    > > glyphs from the supported languages. For instance, if a given glyph is
    > > not used by Amharic translations, it would be removed.
    > >
    > > *that*, I would object to.....

    >
    > What about the conjuncts (combination of basic characters which does not
    > have a unicode code point) that comes in the private area of the font?


    Can you give specific examples please?

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHjZ/lgm/Kwh6ICoQRAkpxAJ0RqC8Z7LFXl3/Crbsl+RQY/0LYZwCgy33S
    VDrqQxGVYB75qKBOlO+nZQo=
    =oHqe
    -----END PGP SIGNATURE-----


  7. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    On Tuesday 15 January 2008, Davide Viti wrote:
    > I've created a test image [1] after applying the above patch; quckly
    > tried it with qemu a took a screenshot [2].
    > Would be nice to have some feedback from original language speakers


    To me the latin text in that screenshot looks virtually identical to what
    you get without the font stripped.

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHjaTCgm/Kwh6ICoQRAjRCAKCYqwGPU1k7ccUfso89wGG6ERpcewCgpoAs
    TZdkLC0vqd1FZxmqx5xOuLI=
    =Oh95
    -----END PGP SIGNATURE-----


  8. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    On Tuesday 15 January 2008, Christian Perrier wrote:
    > In case this proves acceptable, I think we could extend this to other
    > specific fonts.


    As I've said before, this should be tested separately for each font using
    custom images before submitting the patches. IMO you really need to see how
    it looks before you can judge whether it is acceptable or not.

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHjacggm/Kwh6ICoQRAsqrAJkBD/g6pE1eKaCQB6+L+pk1oPB1HwCggrsb
    QdwykzP6ipfveX2ycU+Xg4E=
    =Aj/C
    -----END PGP SIGNATURE-----


  9. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    (Please do not CC me when replying!)

    On Wednesday 16 January 2008, you wrote:
    > 2008/1/15, Frans Pop :
    > > On Wednesday 16 January 2008, Praveen A wrote:
    > > > What about the conjuncts (combination of basic characters which does
    > > > not have a unicode code point) that comes in the private area of the
    > > > font?

    > >
    > > Can you give specific examples please?
    > >

    > The combination of 3 characters (each has a unicode code point) on the
    > left produce the conjunct (does not have a unicode codepont) on the
    > right. There are a lot of such conjuncts (you can see those if you look
    > at the screenshot of the fontforge window).


    These examples are for combinations of characters _within_ the
    language-specific range, so they would not be affected.
    Please give specific examples of conjuncts that use characters from both the
    language-specific range and the latin range.

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHjahdgm/Kwh6ICoQRAqESAJ48eBP/9OOaYOejpg5mDGlKm3k9xgCaAj0p
    EgzfHIk4bqCuQOF5cwHOUtY=
    =Rteu
    -----END PGP SIGNATURE-----


  10. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    On Wednesday 16 January 2008, you wrote:
    > On Tuesday 15 January 2008, Davide Viti wrote:
    > > I've created a test image [1] after applying the above patch; quckly
    > > tried it with qemu a took a screenshot [2].
    > > Would be nice to have some feedback from original language speakers

    >
    > To me the latin text in that screenshot looks virtually identical to what
    > you get without the font stripped.


    Oops. That was with an image that did not yet have the updated gtk-set-font.

    With a good image there are differences (the latin characters look
    "thinner"). It should be up to the translators to decide.
    The most significant difference is probably in punctuation: periods, commas,
    dashes, brackets, etc. This is an aspect of the exclusion I had not really
    realized before.

    Something else: at least one codepoint is missing in the Abyssinica font
    ATM. It shows up in the mirror country selection list: codepoint 1316.

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHjazHgm/Kwh6ICoQRAhTXAJ9Gth7wlVRnf/s2hB2SG9S++6+BRQCgozSx
    GCenv+bEHtcc2NfxYgipsfk=
    =l2rG
    -----END PGP SIGNATURE-----


  11. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    Quoting Frans Pop (elendil@planet.nl):

    > With a good image there are differences (the latin characters look
    > "thinner"). It should be up to the translators to decide.
    > The most significant difference is probably in punctuation: periods, commas,
    > dashes, brackets, etc. This is an aspect of the exclusion I had not really
    > realized before.


    This is something I was considering: leave all punctuation characters
    and maybe also digits as well. So, in short only strip the letters
    (simple and accented).

    >
    > Something else: at least one codepoint is missing in the Abyssinica font
    > ATM. It shows up in the mirror country selection list: codepoint 1316.


    Hmmm, that should be reported to the font package. From what I see in
    fontforge, a few other codepoints are missing: 1315, 1311, 12D7, etc.

    --



    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHjjh91OXtrMAUPS0RAgH9AKCLcc/jvKeDgb1OcInBdJBypUhvQACffOZv
    erikry0lB2nKwWNhb1sQe5Y=
    =5nxP
    -----END PGP SIGNATURE-----


  12. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    On Wed, Jan 16, 2008 at 06:01:49PM +0100, Christian Perrier wrote:
    > Quoting Frans Pop (elendil@planet.nl):
    >
    > > With a good image there are differences (the latin characters look
    > > "thinner"). It should be up to the translators to decide.
    > > The most significant difference is probably in punctuation: periods, commas,
    > > dashes, brackets, etc. This is an aspect of the exclusion I had not really
    > > realized before.

    >
    > This is something I was considering: leave all punctuation characters
    > and maybe also digits as well. So, in short only strip the letters
    > (simple and accented).
    >
    > >
    > > Something else: at least one codepoint is missing in the Abyssinica font
    > > ATM. It shows up in the mirror country selection list: codepoint 1316.

    >
    > Hmmm, that should be reported to the font package. From what I see in
    > fontforge, a few other codepoints are missing: 1315, 1311, 12D7, etc.


    It might be helpful to check existing codepoints in the ttf file using the pdf chart [1];
    in case anyone wanted to do this sort of checks note that the "codepoints" column
    in the spellchecker page [2] lists the set of codepoints used in the various po files

    [1] http://www.alioth.debian.org/~zinosa...abyssinica.pdf
    [2] http://d-i.alioth.debian.org/spellch...el1/index.html

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHjmXQax2slmJA7HoRAh38AKCN2xQbvRJUDqLempvRf3 Nu47gVsgCgxN2+
    4OTfnANdgIHuOelqqTBj5Mg=
    =9Hyy
    -----END PGP SIGNATURE-----


  13. Re: [RFC] Stripping Latin range in fonts used by g-i installer?

    On Wednesday 16 January 2008, Christian Perrier wrote:
    > Quoting Frans Pop (elendil@planet.nl):
    > > With a good image there are differences (the latin characters look
    > > "thinner"). It should be up to the translators to decide.
    > > The most significant difference is probably in punctuation: periods,
    > > commas, dashes, brackets, etc. This is an aspect of the exclusion I had
    > > not really realized before.

    >
    > This is something I was considering: leave all punctuation characters
    > and maybe also digits as well. So, in short only strip the letters
    > (simple and accented).


    In this case it would result in e.g. the string ".com, .net, .edu" to have
    relatively fat periods and commas when compared to the the latin letters.
    I guess it depends on whether the script uses any of the punctuation
    characters itself (Amharic seems not to from the little bit I've seen), or
    whether they're only used "in the context" of latin characters.

    Anyway, seeing this has made me personally *less* in favor of stripping,
    except of course in cases where such differences are a lot less noticeable.

    > > Something else: at least one codepoint is missing in the Abyssinica
    > > font ATM. It shows up in the mirror country selection list: codepoint
    > > 1316.

    >
    > Hmmm, that should be reported to the font package. From what I see in
    > fontforge, a few other codepoints are missing: 1315, 1311, 12D7, etc.


    I think that should be up to the translator as maybe he has the alternative
    of using a different (combination of) characters.

    On Wednesday 16 January 2008, Davide Viti wrote:
    > It might be helpful to check existing codepoints in the ttf file using
    > the pdf chart [1]; in case anyone wanted to do this sort of checks note
    > that the "codepoints" column in the spellchecker page [2] lists the set
    > of codepoints used in the various po files
    >
    > [1] http://www.alioth.debian.org/~zinosa...abyssinica.pdf
    > [2] http://d-i.alioth.debian.org/spellch...el1/index.html


    Sounds like a job for the translator :-)

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHjvZpgm/Kwh6ICoQRAgCmAJ4wvzXeo3EBwwdjgNHoFzJNtg3rCwCfYEX8
    tEAp/B4u9O9IIexrBqP+b/I=
    =b2MR
    -----END PGP SIGNATURE-----


+ Reply to Thread
Page 2 of 2 FirstFirst 1 2