Hi, For those with ftp access .. the new beta version there is somewhat experimental as I rewrote some of the (lua) font code - alternative basemode definitions (context only, experimental) - cleaned up node mode initialization - less memory usage due to adapted data structures - a bit smaller files in the cache - some speedup due to more efficient accessibility - different organization and carry-over of tfm information There will probably be some more changes but I would not be surprised if certain features don't work. This is no big deal and mostly a matter of wrong data access then as nothing changed in the otf processing itself. No testing has been done for the generic variant. This will happen when I'm satisfied with the changes. As the information available in the tfm instance is organized differently one can expect that code that uses it has to be adapted. Anyhow, even after we publish a new beta it is probably good to wait with adapting generic usage till we are stable again. Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
On Fri, Mar 18, 2011 at 02:22:49PM +0100, Hans Hagen wrote:
Hi,
For those with ftp access .. the new beta version there is somewhat experimental as I rewrote some of the (lua) font code
- alternative basemode definitions (context only, experimental) - cleaned up node mode initialization
No time for testing, but a suggestion I'd for a while: currently features supported in base mode need to be registered explicitly based on assumption that certain features tags will be used with lookup types supported in base mode. But in practice a feature tag can contain any lookup, e.g. I know fonts that use contextual alternates under 'liga' tag. My suggestion is to allow any feature tag in base mode and then check the actual type of the registered lookups (which might already be the case). Regards, Khaled -- Khaled Hosny Egyptian Arab
On 18-3-2011 2:42, Khaled Hosny wrote:
On Fri, Mar 18, 2011 at 02:22:49PM +0100, Hans Hagen wrote:
Hi,
For those with ftp access .. the new beta version there is somewhat experimental as I rewrote some of the (lua) font code
- alternative basemode definitions (context only, experimental) - cleaned up node mode initialization
No time for testing, but a suggestion I'd for a while: currently features supported in base mode need to be registered explicitly based on assumption that certain features tags will be used with lookup types supported in base mode. But in practice a feature tag can contain any lookup, e.g. I know fonts that use contextual alternates under 'liga' tag. My suggestion is to allow any feature tag in base mode and then check the actual type of the registered lookups (which might already be the case).
this is somewhat tricky as basemode is a frozen approach: we set up the ligatures once and no real analysis later on takes place indeed checking takes place so in principle any feature that has ligature, alternate of substitution is dealt with; of course it's no problem to enable 'all' features in basemode, as long as we let complaints about funny side effects go into the void; i'll have to think about how to provide that variant efficiently ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
On Fri, Mar 18, 2011 at 02:58:23PM +0100, Hans Hagen wrote:
On 18-3-2011 2:42, Khaled Hosny wrote:
On Fri, Mar 18, 2011 at 02:22:49PM +0100, Hans Hagen wrote:
Hi,
For those with ftp access .. the new beta version there is somewhat experimental as I rewrote some of the (lua) font code
- alternative basemode definitions (context only, experimental) - cleaned up node mode initialization
No time for testing, but a suggestion I'd for a while: currently features supported in base mode need to be registered explicitly based on assumption that certain features tags will be used with lookup types supported in base mode. But in practice a feature tag can contain any lookup, e.g. I know fonts that use contextual alternates under 'liga' tag. My suggestion is to allow any feature tag in base mode and then check the actual type of the registered lookups (which might already be the case).
this is somewhat tricky as basemode is a frozen approach: we set up the ligatures once and no real analysis later on takes place
indeed checking takes place so in principle any feature that has ligature, alternate of substitution is dealt with;
of course it's no problem to enable 'all' features in basemode, as long as we let complaints about funny side effects go into the void; i'll have to think about how to provide that variant efficiently
My real use case is that I've stylistic sets that I need to use in base mode (for math) but I need to manually register ss01..ss20 features in base mode first though they all are gsub_single lookups, so if context would simply check if the feature is gsub_single (or gpos_pair etc.) and reject it if not, it would be more efficient than hand picking certain features as the check will be done anyway. Regards, Khaled -- Khaled Hosny Egyptian Arab
On 18-3-2011 3:16, Khaled Hosny wrote:
My real use case is that I've stylistic sets that I need to use in base mode (for math) but I need to manually register ss01..ss20 features in base mode first though they all are gsub_single lookups, so if context would simply check if the feature is gsub_single (or gpos_pair etc.) and reject it if not, it would be more efficient than hand picking certain features as the check will be done anyway.
indeed Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
participants (2)
-
Hans Hagen -
Khaled Hosny