On 4/1/2023 11:21 PM, Carlos via ntg-context wrote:
On Fri, Mar 31, 2023 at 10:22:49PM +0200, Hans Hagen via ntg-context wrote:
On 3/31/2023 10:08 PM, Carlos via ntg-context wrote:
sure, why should it, you want lucida so better quit with an error than kicking in some font; actually cmr math fonts have been obsoleted for way over a decade by latin modern math fonts in 32 bit font engines
Font loading and processing time can be mosty neglected so these 16 seconds come from something else, maybe there are ways to trace file access. Another possibility is that your fonts are not cached in which case every run will involve parsing the otf / ttf and producing whatever resources needed (normally cached).
interesting. Earlier as the output was showing
mkiv lua stats > loaded patterns: en::1, load time: 0.000 mkiv lua stats > loaded fonts: 4 files: lucidabrightregular.otf, lucidasansregular.otf, latinmodern-math.otf, lmroman10-regular.otf
you asked
so why not use lucida math fonts?
which is a valid question but nevertheless unsettling in that lmodern-regular may nat have been called out. I fully understand the inclusion of a latinmodern-math in it as a lucidamath was not previously available but I ponder at the idea and involuntary implementation of having lmroman along the pack. Why? If this is TeX doings, or misdoings (depending how one looks at it), it clearly shows to me that TeX also restricts my freedom to use whatever font I may deem necessary. Don't you think? You can probably disagree with me here, or anyone from the TeX community can, but the roman last was imposed deliberately upon. Someoe may also give a lengthy explanation but that would be just hogwash in thee very end .
The fact that a file is loaded doesn't mean that it is used. In fact there is some fallback loading going on because explaining when nothing gets typeset or why \nullfont is used which results in missing chars messages is a waste of time. It has nothing to do with restrictions: tex will load any font you want it. But for instance when you turn on some tracing the macro package can load a font it needs. Indirectly you get what you ask for. I bet you didn't explicitly ask for e.g. english hyphentation patterns either but still got them, but tex will happily let you use polish ones instead. Of course the best remedy against any automatism is to just start with initex and implement from scratch. In the above case: just add \type {foo} and you might see another font you didn't ask for / set up.
I was actually thinking to ask you about that, and by falling back to cmr math font that perhaps would expedite loading time along the way.
These fonts are small (only huge cjk fonts with tens of thousands of glyphs or fonts with hundreds of accumulated features might have some impact but even then not in the final embedding stage).
Yeah. I guess.
I can also have
mkiv lua stats > loaded patterns: en::1, load time: 0.000 mkiv lua stats > loaded fonts: 3 files: lucidabrightmathsymbol.ttf, lucidanewmathitalic.ttf, lucidabrightregular.otf mkiv lua stats > font engine: otf 3.133, afm 1.513, tfm 1.000, 6 instances, 3 shared in backend, 1 common vectors, 2 common hashes, load time 16.723 seconds
but that loading time gets back at me as the culprit sweet reminder of not using cmr then.
i won't comment because (i verified) that lucidabrightmathsymbol.ttf and lucidanewmathitalic.ttf are illegal copies (if created on someones machine from the typeone its all up to that user). Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | www.pragma-ade.nl | www.pragma-pod.nl -----------------------------------------------------------------