| <<<Back 1 day (to 2016/07/30) | 20160731 |
HenryStiles | ugh laptop died, disk failure, I've had it with Lenovo, back to MacBook for the laptop | 00:19.11 |
Robin_Watts | HenryStiles: SSD time. | 00:20.59 |
HenryStiles | I don't like Lenovo for a lot of reasons and I miss the macbook | 00:23.57 |
| they are expensive of course | 00:24.15 |
Robin_Watts | Macbooks are style over substance, unexpandable, too expensive, and far to prone to breaking for my taste. | 00:28.15 |
| Your Lenovo is still in guarantee, right? | 00:28.42 |
| (ISTR it's just a bit older than mine, and mine had a 3 year guarantee) | 00:29.02 |
HenryStiles | hate it, that simple | 00:38.29 |
| bulky awkward ugly trackpad off center ... on and on | 00:39.13 |
Robin_Watts | hmm. I don't feel mine is bulky. The trackpad is central for me too. Different models, plus different tastes etc, I guess. | 00:48.08 |
| anyhooo, bedtime. | 00:48.32 |
fredross-perry | I love my Macbook Pro w/SSD. Light and powerful (but expensive). I think the extra $$ pay for themselves in productivity within the first year. trackpad: I use a mouse whenever possible. | 01:27.35 |
sivoais | Hi all. I am trying to build MuPDF on FreeBSD and came across the same bug as in <http://bugs.ghostscript.com/show_bug.cgi?id=696828>. | 03:16.34 |
| The patch that removes HAVE_INCBIN works, but I was wondering if that would make it into a future release. | 03:17.43 |
sebras | Robin_Watts: but you are fine with us expanding ligatures? it's a pedantic difference, but it is there. ;) | 04:36.43 |
| Robin_Watts: if we leave tabs in the raw text then the only option is to special case c == 9 in fz_copy_selection(), or would you want us to remove the c < 32 comparison altogether? | 04:39.54 |
HenryStiles | fredross-perry: 13 or 15? | 04:43.06 |
fredross-perry | 13. 500 Gb SSD. 8 Gb RAM. | 04:43.48 |
HenryStiles | fredross-perry: went to the store this evening and looked. Little worried that all the models are mid 2015, should have an update soon I guess. | 04:45.06 |
fredross-perry | Mine's late 2013 (refurb). good price, works great. | 04:45.36 |
HenryStiles | http://buyersguide.macrumors.com/#Mac , the 12 is the only one that is "buy now" | 04:47.45 |
| may not need the latest stuff, but skylake would be nice | 04:48.42 |
| off to bed, I'll mull over it. Good night | 04:49.02 |
fredross-perry | 'night | 04:49.21 |
| But, you still want compute power, which the MBP has over the MB (I think): http://www.laptopmag.com/articles/macbook-vs-air-vs-pro | 04:58.31 |
sebras | Robin_Watts: we also do translation of strings containing 0x00 in pdf_show_char() to '?'. shouldn't that be .notdef? | 05:12.06 |
| Robin_Watts: or what am I missing here? | 05:12.11 |
| I have a hard time separating what interfaces we want to return the raw untranslated character ids through and what interfaces we want to do rudimentary translation as a convenience for the applications. | 05:13.13 |
| Robin_Watts: when you have time there are a slew of commits for review on sebras/master. you did look at them prior to tor8's comments which I believe that I have addressed now. | 06:55.21 |
Robin_Watts | sebras: In all honesty, no. | 11:16.31 |
| I'm not OK with us expanding ligatures - but that's something tor8 did before my opnions formed. | 11:17.00 |
| I think I'd like either: | 11:17.14 |
| 1) for us to give raw data, and then to have helper functions that 'simplify' the data appropriately that we can call later. | 11:17.51 |
| or 2) to have some options to the functions that say whether we expand ligatures, simplify whitespace etc. | 11:18.11 |
| fred, henry: The trackpad on my old macbook is SOOO much nicer than the lenovos one, I'll give you that. | 11:19.29 |
| but it's not so different that it justifies the awful keyboard the mac has. | 11:19.52 |
sebras | Robin_Watts: right. I don't really have any opinions of my own on these matters I just want to add code that fits with what is already there. And if we want to make a redesign later we can. | 13:53.51 |
| Robin_Watts: we synthesize spaces between words also but hueuristics from the distance between characters. so you envision that these parts should also be separated out with ligature expansion and possible tab simplification? | 14:40.07 |
| s/but/by/ | 14:40.16 |
Robin_Watts | sebras: Hadn't considered space synthesis until now, but yes, now you come to mention it. | 17:01.33 |
| I'd be tempted to have a flags word for options in the structured text device. | 17:01.55 |
| One bit to mean "synthesise spaces", one bit to mean "expand ligatures", one bit to mean "simplify whitespace" etc. | 17:02.32 |
sebras | Robin_Watts: right, then I think I understand where you are coming from. :_ | 17:22.36 |
| :) | 17:22.39 |
| Robin_Watts: what about converting 0x00 to '?' inside the pdf-parser? | 17:24.09 |
| Robin_Watts: 0x00 is a valid PDF string character, but I'm guessing that there are assumptions along the way that assume that strings are NULL-terminated. | 17:24.43 |
| maybe 0x00 should be represented by .notdef if present, I'm not sure how those things work. | 17:25.09 |
| all of this ought to be before the struct text device is even involved. | 17:25.37 |
ray_laptop | sebras: (for the logs) I say your comment "0x00 is a valid PDF string character, but I'm guessing that there are assumptions along the way that assume that strings are NULL-terminated.", but doubt that (at least in gs) since multi-byte fonts often have a '0' byte | 20:53.28 |
Robin_Watts | sebras: The PDF parser shouldn't change at all. | 23:20.35 |
| It handles all the encoding etc that we want. | 23:20.44 |
| What I am talking about is just the structured text extraction. | 23:20.55 |
| That handles purely the values once then get to the text fz_device calls. | 23:21.17 |
| Whatever coding is used internally (PDFDocEnc, or whatever) doesn't matter before we get to the fz_device calls. | 23:21.51 |
| but after that, we start to 'extract' the text into unicode etc. | 23:22.02 |
| At that point, it matters how we 'sanitize' the results - infer whitespace, collapse whitespace, expand ligatures etc. | 23:22.27 |
| Forward 1 day (to 2016/08/01)>>> | |