| <<<Back 1 day (to 2018/03/11) | 20180312 |
gour | morning | 10:07.12 |
| how is one supposed to print pdfs when using mupdf? | 10:07.25 |
kens | 'print' in what sense ? MuPDF is (essentially) a library for rendering PDF files, there are a number of example viewers, as far as I know none of them drives a physical ink-on-paper printer as such. I could be mistaken, I'm not a MuPDF developer. | 10:09.20 |
gour | kens: well, if one uses e.g. qpdfview, evince, atril, okular, xpdf etc. it's possible to send viewed document directly to the printer | 10:12.18 |
kens | You're going to have to wait until someone else comes online then, I can't answer the question. It could be that its possible (depending on the operating system you are using, and the specific example viewer) but I don't know. | 10:13.20 |
| Of coruse, it should be relatively simple to render a bitmap to the desired resolution and have the operatign system send that via the printer driver to a specific printer. | 10:14.04 |
gour | i'm looking for a viewer to replace evince/atril which produce incorrect result on (fedora) linux | 10:18.53 |
tor8 | cosfskt (for the logs): there is no function to display the creation date of a pdf on the android app. | 11:42.57 |
| gour (for the logs): sorry, the mupdf viewer can't drive the system printer directly. you should be able to convert to PWG or a high resolution bitmap and print that, using say "mutool convert -o out.pwg -r600 input.pdf" | 11:44.01 |
| sorry, not "-r600" that's old habits | 11:44.41 |
gour | ok, that would be ok for me, but not for not-so-savvy relative :-) | 11:52.32 |
tor8 | it should be -O resolution=600 instead | 12:04.32 |
vtorri | hello | 13:29.26 |
mubot | Welcome to #mupdf, the channel for MuPDF. If you have a question, please ask it, don't ask to ask it. Do be prepared to wait for a reply as devs will check the logs and reply when they come on line. | 13:29.26 |
vtorri | tor8 sinc mupdf 1.12, fz_new_pixmap_with_bbox_and_data() has a new argument : a separator | 13:29.51 |
| tor8 : what is it ? | 13:30.10 |
tor8 | vtorri: not a separator, it's a list of color separations (i.e. spot colors) | 13:30.28 |
| you can pass NULL safely | 13:30.31 |
vtorri | ok | 13:30.38 |
| yes, sorry, separations | 13:30.49 |
| tor8 it works, thanks | 13:38.48 |
| though i have to create a test with the autotools to check if I have mupdf 1.11 or 1.12 :( | 13:39.18 |
tor8 | vtorri: for the next release we're adding FZ_VERSION_MAJOR _MINOR and _PATCH #defines | 13:52.23 |
| 1.13 should be out later this month I hope | 13:52.36 |
vtorri | thank you | 13:52.45 |
| PATCH is just for notifying bug fixing, right ? | 13:53.23 |
tor8 | if we need to do an intermedate release, we'll use the PATCH level | 13:56.00 |
| instead of 1.11b like we did before | 13:56.10 |
| fredross-perry (for the logs) there's a fix for non-showing annotations in java on origin/master, and a bit of sample code on tor/java | 13:58.49 |
| and there's also a crucial reference counting fix there too | 13:59.38 |
Robin_Watts | Moving here for release discussion. | 14:34.41 |
| I'm not sure I buy that you can't usefully use text_bits !=0 with any or center rasterisers. | 14:35.27 |
tor8 | Robin_Watts: the any or center rasterizers are non-aa | 14:41.20 |
| so we should not antialias text for these, IMO | 14:41.34 |
| unless you set the text=aa8 explicitly | 14:41.41 |
| or do you think we should convert to paths and use the any/center rasterizer directly instead of using freetype's own rasterizer? | 14:42.45 |
Robin_Watts | tor8: My point is simply that there are PDF files out there that render badly if you don't use the any rasteriser. | 14:46.04 |
| and if you DO use that for rendering to screen, it can still look much nicer to use text antialiasing. | 14:46.23 |
tor8 | Robin_Watts: that's why we have separate flags for text and graphics, isn't it? | 14:46.39 |
Robin_Watts | Yes. | 14:46.45 |
tor8 | Robin_Watts: if I set 'any' for text rendering, I would expect it to be non-AA like when I set graphics rendering to 'any' | 14:47.02 |
Robin_Watts | Ah, maybe I'm misreading your patch. | 14:47.31 |
tor8 | if (aa > 8) I set text_bits to 0 | 14:47.47 |
| that's all it does | 14:47.50 |
| so it detects the app/cop rules and zeroes text_bits instead of clamping to 8 | 14:48.12 |
Robin_Watts | tor8: yeah, being dim. LGTM. | 14:49.04 |
tor8 | Robin_Watts: Thanks. | 14:49.14 |
sebras | Robin_Watts: tor8: do we have a file with crypt->r == 6? | 15:07.52 |
Robin_Watts | Pass. | 15:08.16 |
tor8 | for f in *.pdf; do mutool show $f g 2>/dev/null ; done|grep '/V '|grep '/R 6' finds 1931_-_crypt_version_5_revision_6.pdf in tests_private/pdf/sumatra | 15:15.34 |
sebras | oh! I'll log in and grab that file! | 15:16.19 |
Robin_Watts | 1 commit on robin/master for review. | 15:16.28 |
| plus the begin_layer/end_layer stuff that you've looked at before, and I've forgotten where we landed. | 15:16.51 |
tor8 | Robin_Watts: "Bug 699106" LGTM | 15:17.06 |
Robin_Watts | Ta. | 15:17.15 |
sebras | tor8: nice, then I have tested combinations crypt->v/r suceessfully: 1/2, 2/3, 4/4, 5/5, 5/6 | 15:23.18 |
| tor8: so I feel more confident now that sebras/master really works. | 15:23.34 |
| tor8: you LGTMed the first three patches, and asked whether I had tested the top two ones. now I have. | 15:24.19 |
| Robin_Watts: if you feel like taking a look, that's also ok. :) | 15:35.14 |
tor8 | sebras: sebras/master LGTM | 16:00.20 |
sebras | ty | 16:08.08 |
aksr | hi, if the pdf is cyrillic, command "mutool draw -F txt file > file2", transliterates the text in latin. is there a way to get this right? | 16:53.02 |
tor8 | aksr: mupdf emits the text as encoded in the PDF file as utf-8 | 16:53.56 |
| quite likely the PDF is badly constructed | 16:54.04 |
aksr | hm, this happened (if i remember correctly) almost every time... | 16:54.33 |
| are you certain? | 16:54.37 |
tor8 | very often a PDF claims to use WinANSI encoding, but then use a font where the characters are *not* what they should be | 16:54.40 |
aksr | do you have any suggestion about this? | 16:54.59 |
tor8 | I have to leave for dinner, but if you post a link to a file I can take a look when I get back | 16:55.00 |
| aksr: look at what "mutool info" says about the fonts | 16:55.24 |
sebras | tor8: I found you some more icons. | 16:55.43 |
aksr | tor8: can you be more specific, on what should i pay my intention? | 16:57.29 |
sebras | tor8: they are on sebras/ui | 16:58.20 |
| aksr: if you run mutool info -F my-document.pdf you may see lines like: | 16:59.45 |
| 345 (890 0 R): TrueType 'EENBJG+TimesNewRoman,Italic' WinAnsiEncoding (5956 0 R) | 16:59.57 |
aksr | i already did it | 16:59.57 |
| MacRomanEncding (here) | 17:00.07 |
| Encoding | 17:00.14 |
sebras | aksr: that's what tor8 wanted to know. | 17:00.16 |
| aksr: is the file publicly available? I'm not good at debugging text encodings, but it is better ot post the link than not to. :) | 17:01.38 |
aksr | just a sec | 17:03.31 |
| sebras: yes, but on the scribd, is that okay? | 17:06.12 |
sebras | aksr: would we need to register for an account? (that's annoying) :) | 17:06.55 |
aksr | for years now, everything on this site is annoying ... | 17:07.24 |
| ;) | 17:07.30 |
| i'm guessing there's no way to re-encode it via ghostscript or something? | 17:08.41 |
sebras | aksr: send the scribd link at least. | 17:09.36 |
kens | Idf the fotn doesn't have a ToUnicode CMap, and isn't usin a recognised encoding, then no, there's no way to correct it. | 17:10.09 |
sebras | aksr: I don't know this encoding stuff well enough to help you though, but when tor8 returns he will see you link. | 17:10.15 |
kens | THe information is gone at that point | 17:10.16 |
| I csna look at it | 17:10.25 |
| can* | 17:10.29 |
aksr | i encountered by now many files to exibit this problem, one of them: https://www.scribd.com/doc/260265132/Dobrotoljublje-v-Tom | 17:10.49 |
kens | Though I'd rather dropbox or something I don't have to register to use | 17:10.58 |
| It seems i have to give this site my credit card details for a 'free tiral' in order to download the file | 17:12.51 |
| I'm not prepared to do that | 17:12.55 |
| Can you put it soemwhere else please ? | 17:13.02 |
Robin_Watts | aksr: Can you not just open a bug and add it there? bugs.ghostscript.com | 17:14.00 |
aksr | Robin_Watts: ok, but what should i state as a problem? wrong encoding or... | 17:15.15 |
Robin_Watts | aksr: Just do your best to describe the issue. | 17:15.46 |
aksr | i filed a bug. but there's not much to tell | 17:30.34 |
| https://bugs.ghostscript.com/show_bug.cgi?id=699107 | 17:30.42 |
sebras | aksr: please attach a file that exhibits the issue, if you want to it can be marked private. | 17:39.34 |
aksr | sebras: ok, done. | 17:43.27 |
tor8 | aksr: Yeah, there's not much we can do about these files. They pretend to use latin fonts, and use latin text, but the embedded font has cyrillic outlines | 18:28.13 |
| aksr: Even when I extract the embedded font and look at it, it has glyphs with latin character names. | 18:28.46 |
| so there's no way for us to know that this is actually cyrillic text, without doing OCR. | 18:29.09 |
HenryStiles | Robin_Watts: sorry to bother you can you send an email to tech again from wss | 19:43.35 |
| ? | 19:43.41 |
| Forward 1 day (to 2018/03/13)>>> | |