| <<<Back 1 day (to 2020/06/15) | Fwd 1 day (to 2020/06/17)>>> | 20200616 |
malc_ | wonders if (correct rendering of https://en.wikipedia.org/wiki/Double_acute_accent à la https://boblycat.org/~malc/scratch/double-accute.png) is the norm or an exception for "most" people | 09:26.51 |
ator | IAEON: convert to CBZ. unpack the files, then repack them as ZIP, and rename the .zip to .cbz | 09:55.04 |
chrisl | Does no one think that PDF producers should be held responsible for producing crap, rather than bitching at the PDF consumers about it?? :-( | 10:11.41 |
malc_ | chrisl: https://en.wikipedia.org/wiki/Robustness_principle | 10:13.00 |
kens | IMO an error message 'your file does not conform to the specificatioon' is quite robust enough | 10:14.00 |
| Trying to guess what Adobe Acrobat will or will not accept is too much like hard work | 10:14.30 |
| There;'s a specification, its not as good as the PLRM but people should follow it. | 10:14.51 |
| Even if software accepts broken PDF files it should say 'your file is broken' so you know | 10:15.04 |
chrisl | malc_: Being well behaved in the face of broken input is not the same as happily accepting and trying to "fix" blatently broken input. I don't think producers should be permitted to just ignore the spec, and expect consumers to make sense of it | 10:15.11 |
malc_ | chrisl: i am not too fond of Robust principle, iow you are preaching to the choir | 10:16.09 |
| +ness | 10:16.17 |
chrisl | Perhaps the next time gcc errors out on some of my typo ridden C code, I should just open gcc bug, rather than fix my code! | 10:18.11 |
malc_ | you should! after all search engines do that all the fucking time! why can't a bloody compiler just DWIM?? | 10:19.45 |
chrisl | Psychic software: do what the users wants, not what they tell it to do! | 10:23.27 |
malc_ | yep. that's the plan. we all be better off with as little human involvement as possble | 10:28.27 |
| ator: mupdf fails to properly render trema/umlaut over A for all the fonts in distribution (even for Noto Sans a font with which diacritics are correct, at least, in firefox) (demonstration - https://tpaste.us/re4z) | 13:26.06 |
ator | what fonts? which distribution? | 13:39.40 |
malc_ | ator: fonts found in mupdfs resources directory | 13:40.33 |
| Noto, Charis, DroidSans | 13:40.53 |
ator | malc_: I blame Harfbuzz... | 13:48.19 |
| maybe I'm not setting a correct flag, or it's just not capable | 13:48.42 |
malc_ | firefox uses hb too | 13:48.56 |
| (i think)) | 13:48.59 |
| damn extra paren... i should go commit sepuku | 13:49.17 |
ator | malc_: your script doesn't work for me. it's loading a font-face but not using it. | 13:59.07 |
| s/pre/body/ | 13:59.51 |
| with those fixes, some fonts work, others not so much (depends on the presence of a GPOS opentype table) | 14:00.29 |
malc_ | ator: doh, sorry | 14:00.51 |
ator | the built-in Noto Serif font works | 14:01.06 |
malc_ | noto sans works too | 14:01.52 |
ator | the Charis ones don't because we ship a stripped down version | 14:02.05 |
malc_ | so sorry for insinuating that firefox is better | 14:02.12 |
ator | the original (bloated) charis sil font works | 14:02.14 |
malc_ | does original charter work? | 14:02.35 |
| ator: any idea if this https://github.com/moosotc/snippets/blob/master/bin/3ze.py https://github.com/moosotc/snippets/blob/master/bin/3ze.sh can be simplified (not requiring fontforge)? | 16:09.05 |
| <<<Back 1 day (to 2020/06/15) | Forward 1 day (to 2020/06/17)>>> | |