| <<<Back 1 day (to 2020/11/15) | Fwd 1 day (to 2020/11/17)>>> | 20201116 |
pedr0 | hi all - would you know what are those DRM restrictions ? I have a bizarre PDF I can't edit within an editor even after a mutool clean pass... bever happened | 15:03.24 |
ator | pedr0: sorry, I have no idea what you're talking about | 15:14.34 |
| if you mean an encrypted PDF file with restrictive permissions, try 'mutool clean -D' | 15:15.00 |
sebras | ator: did you look at sebras/sigantures? | 15:30.57 |
| signatures... | 15:31.01 |
ator | +ui_label("Signature field is %slocked.", readonly ? "" : "un"); still mentions "locked" in the rename commit | 15:35.07 |
| " Widgets are not locked, they are read-only, so rename interfaces." does not compile. it references a function that does not exist: pdf_signature_is_readonly | 15:36.54 |
sebras | ator: it does. I was wary about changing that because I'm not sure what non-readonly would be. modifiable? | 15:36.57 |
ator | sebras: maybe only show the line if it is readonly, and show nothing if it's modifiable | 15:37.20 |
sebras | ator: ah, crap. I tried to rebase things properly. I'll fix that. | 15:37.20 |
ator | "Widgets should ignore the annotation read-only flag." should that be squashed in too? | 15:37.56 |
sebras | ator: the intent was to have the first commit just do a rename. | 15:41.33 |
| then do changes in later patches. | 15:41.43 |
| I read in the spec that the read-only flag in /F is ignored for widgets. for those only the read-only flag in /Ff matters. | 15:42.15 |
ator | sebras: ok. | 15:43.16 |
artifexirc-bot | <Fred> @ator are you going to put your CJK font fix on 1.18.x? | 15:51.33 |
ator | @Fred sure, if you need it | 15:52.42 |
artifexirc-bot | <Fred> I do, thanks. | 15:52.56 |
ator | @Fred um, 1.18.x doesn't build for me | 15:53.50 |
artifexirc-bot | <Fred> hm. | 15:54.14 |
ator | No rule to make target 'build/debug/thirdparty/tesseract/src/ccutil/globaloc.o' ... do I need to remove the tesseract thirdparty module or change its version? | 15:54.18 |
| it looks like a remnant of going from master where tesseract is a submodule to the 1.18.x branch where it is not | 15:55.26 |
| and the file lists don't match | 15:55.32 |
| the 1.18.x tesseract submodule must use a different version than master I guess | 15:56.01 |
artifexirc-bot | <Fred> 1.18.x builds for me. | 15:56.41 |
| <ator> I got it building by removing the submodule | 15:56.48 |
| <ator> I pushed the CJK commit to 1.18.x | 15:56.54 |
| <Fred> thx | 15:56.57 |
pedr0 | ator: '-D' is exactly what I am talking about :-) | 16:00.10 |
ator | pedr0: you may need to add -P to set the permissions you want | 16:01.37 |
pedr0 | thanks for that | 16:02.20 |
ator | -1 should grant all permissions | 16:03.13 |
| *but* I don't think you should need that, since with -D we don't write an encryption dictionary which is the place where permissions are stored | 16:03.50 |
sebras | ator: I have updated sebras/signatures with fixup commits and made sure that things build | 16:35.16 |
| ator: more opinions or LGTM? | 16:46.20 |
| ator: opinions on the test cases? | 16:46.26 |
artifexirc-bot | <Robin_Watts> @sebras Moved here from #artifex... | 17:12.28 |
| <Robin_Watts> "Widgets should ignore the annotation read-only flag". Why? | 17:12.44 |
| <Robin_Watts> (I'm sure you have a good reason. Could you state it in the commit message?) | 17:13.18 |
sebras | @Robin_Watts because the spec says so. :) | 17:13.19 |
artifexirc-bot | <Robin_Watts> So an annotation can be read only, but the widget is not? | 17:13.42 |
| <Robin_Watts> hmm, that vaguely makes sense if you want the widget to have (say) its text filled in, but people not to be able to edit the size/shape of the widget, maybe. | 17:14.32 |
sebras | no, both can be read only, but there are two ReadOnly flags, in pdfref17 page 609 it mentions the ReadOnly flag for annotations, with a Note stating that this flag is ignored for widgets, instead it defers to ReadOnly in /Ff | 17:14.36 |
| and yes, I ought to mention this in the comment message, will fix. | 17:14.47 |
artifexirc-bot | <Robin_Watts> Thanks. I know I'd be confused by it, and that extra information would help a lot. | 17:15.11 |
| <Robin_Watts> trace_next_error. Will that return old = 0 on the first call? | 17:16.50 |
| <Robin_Watts> Being *REALLY* picky, in pdf_signature_format_designated_name, I'd have done if (name == NULL) return; | 17:19.05 |
| <Robin_Watts> smaller diff, less right hand creep. | 17:19.15 |
| <Robin_Watts> Being *REALLY* picky, in pdf_signature_format_designated_name, I'd have done if (name == NULL) return NULL; | 17:19.24 |
| <Robin_Watts> Otherwise, all lgtm. | 17:20.18 |
| <paulgardiner> Were we setting the correct flag for locked fields when signing? | 17:26.16 |
| <Robin_Watts> @sebras @cgdae @ator https://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=31950b3a77c9b2ab373f537daf10dc4a911ec7cc | 17:44.15 |
malc_ | Robin_Watts: the patch is inconsistent, in one case you cast w in other n | 17:46.51 |
artifexirc-bot | <Robin_Watts> malc_: I cast the second operand to mul each time. | 17:47.28 |
| <Robin_Watts> I can never remember if (cast)a * b is ((cast)a) * b or (cast)(a * b) | 17:48.01 |
| <Robin_Watts> The former would be what we want, the latter would not. | 17:48.24 |
| <Robin_Watts> whereas a * (cast)b has no such problem. | 17:48.37 |
malc_ | Robin_Watts: okay, but that doesn't make it any less inconsistent, you alwys have an option to flip the args | 17:50.07 |
artifexirc-bot | <Robin_Watts> In fact, the last one is bad there. | 17:51.37 |
| <Robin_Watts> Let me rejig | 17:51.42 |
| <Robin_Watts> malc: https://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=8aad2efc637ef00f87fb158ee9d51fdf2ba698a8 | 17:52.39 |
malc_ | Robin_Watts: no consistency nags from me, excellent job :) | 17:53.32 |
| all that said if you really care you should devise and use some kind of mul function (ala MSs pranoid CRT) and use that, otherwise signed unsigned an C's lack of type safety will eventually bite | 17:55.49 |
Hamiro | Hey guys everything ok? | 18:30.53 |
| I came accross this post regarding macos clipboard | 18:31.06 |
| https://ghostscript.com/mupdfirclogs/2020/10/21.html | 18:31.08 |
| just wondering if there is any updates on it> | 18:31.21 |
artifexirc-bot | <Robin_Watts> Hamiro: ator commented on that topic just a few days ago. | 18:35.39 |
| <Robin_Watts> https://ghostscript.com/mupdfirclogs/2020/11/10.html | 18:36.45 |
| <Robin_Watts> 14:!6 | 18:36.47 |
| <Robin_Watts> 14:16 | 18:36.56 |
Hamiro | Oh ok, would that even work xD building on linux and using it on macos? I have a linux machine lemme give it atry ty | 18:38.12 |
malc_ | Hah i remember that, i managed to convince skejg to build llpp, in the end only one file refused to be built - cocoa.m... all hail macOS | 18:39.36 |
Hamiro | I mean i'm an IOS dev I feel the pain on a daily basis, but the job is the job.... | 18:40.22 |
malc_ | that is honestly the first time i'm speaking with an iOS device, or maybe it's just my oldfartness(the one that refuses to accept excessive shortening) speaking | 18:42.05 |
| Hamiro: you can try and succeed where skejg failed btw.. being an iOS developer suggests that you might have a more uptodate macOS xcode-tools installation | 18:43.49 |
Hamiro | I'm not really sure but the issue seems to be just freeglut not finding any headers or something, unfortunately my C knowledge is basically 0 :( | 18:45.41 |
| Ill give it a try tho | 18:45.54 |
malc_ | Hamiro: if you are talking to me, please prefix it with my nick, the benefits are twofold a) i get notified when looking elsewhere b) am sure that is me who is being addressed | 18:47.51 |
artifexirc-bot | <Robin_Watts> Hamiro I think malc is trying to persuade you to try building llpp. llpp is NOT mupdf, but uses it. | 18:48.28 |
Hamiro | malc_ Got it thanks | 18:49.15 |
malc_ | Hamiro: I am trying to persuade you to build llpp which is merely a mupdf usng application, it does have a coca backend and utilizes pb{copy,paste} on macOS for clipboard integration. Also thanks. | 18:50.18 |
| Robin_Watts: and you foiling my attempts to covertly take over document world are most unfortunate.. (also sorry if my proselytizing has gone too far) | 18:52.26 |
artifexirc-bot | <Robin_Watts> malc_ No, I just wanted to be clear what the situation was. | 18:53.18 |
malc_ | Robin_Watts: gotcha | 18:53.41 |
| Native speakers: "... has quit: Remote host closed the connection" isn't "the connection" a bit weird the "the" part of it? | 19:09.31 |
sam_ | nah | 19:10.00 |
sebras | @Robin_Watts no errored will return 2 the first time becuase I set errored = 2 immediately after we do trace_action("var doc, page, annot, widget, widgetstr, hits, tmp,...._); | 19:56.51 |
| in pdf_signature_format_designated_name() I needed the NULL-check for name to preceede the assignment of parts. | 19:58.26 |
| @Robin_Watts so I just created a scope for it. | 19:58.47 |
| sure I can do if (name == NULL) return NULL; else { const char *parts....; } | 19:59.08 |
| but it doesn't avoid the extra indentation level anyway. | 19:59.26 |
| or else I need to separate the declaration of parts from the assignments of its elements (the elements based on name in particular). | 20:00.04 |
| @Robin_Watts I have updated sebras/signatures with fixup commits. | 20:05.25 |
| <<<Back 1 day (to 2020/11/15) | Forward 1 day (to 2020/11/17)>>> | |