| <<<Back 1 day (to 2016/11/16) | 20161117 |
tor8 | Robin_Watts: so, we ready to do RC2? | 11:28.15 |
Robin_Watts | tor8: yeah, as far as I am concerned at least. | 12:28.36 |
tor8 | Robin_Watts: okay. I've tagged and uploaded a source tarball | 12:36.19 |
| do you want me to bulid the windows binaries so you can sign them? | 12:36.35 |
Robin_Watts | tor8: That'd work. | 12:37.35 |
tor8 | oh ffs. pdf-name-table.h not regenerating properly in MSVC is getting annoying. | 12:38.57 |
Robin_Watts | tor8: If you 'rebuild' it's OK. | 12:39.47 |
tor8 | does 'rebuild' delete the pdf-name-table.h file? the generate.bat script sees it and doesn't run the namedump tool, regardless of time stamp. | 12:40.24 |
Robin_Watts | tor8: I think so. | 12:42.19 |
| del /q ..\..\generated ..\..\include\mupdf\pdf\name-table.h ..\..\source\pdf\pdf-name-table.h | 12:42.38 |
| That's the 'clean' portion of the rebuild command. | 12:42.59 |
tor8 | aha! 'Clean build' removed the file. thanks! | 12:43.12 |
| I've just been deleting the 'Release' directory and doing git reset --hard | 12:43.31 |
malc_ | tor8: git clean -dfx | 12:45.32 |
tor8 | malc_: that removes a bit more than I'd like though... :) | 12:46.22 |
Robin_Watts | 'Rebuild solution' is easier :) | 12:46.27 |
tor8 | I prefer keeping my TODO file around... | 12:46.32 |
malc_ | tor8: :) | 12:48.04 |
tor8 | Robin_Watts: on casper, /home/tor/mupdf-1.10-rc2-windows.zip | 12:50.24 |
Robin_Watts | tor8: Done. Just going to run it through advzip now. | 12:59.06 |
| tor8: OK, signed exes all zipped up as /home/robin/mupdf-1.10-rc2-windows.zip. Advzip saved 2%. | 13:33.31 |
tor8 | Robin_Watts: thanks. files are live on mupdf.com/downloads/ | 13:37.03 |
norbertj | +Robin_Watts: Hi, I re-installed git and now my git pull fails with "Permission denied (publickey)." Can you help? | 13:49.07 |
tor8 | norbertj: can you ssh into casper? | 14:04.26 |
| norbertj: what's the git url listed by "git remote -v"? | 14:04.51 |
norbertj | tor8: with filezilla I can ftp to casper, but ssh 184.73.189.105 fails with Permission denied | 14:13.35 |
| tor8: git remote -v : golden norbert@184.73.189.105:/home/git/ghostpdl.git (fetch) and (push) | 14:15.04 |
Robin_Watts | norbertj: previously, were you using plink rather than ssh? | 14:27.10 |
| I have GIT_SSH set to plink.exe here. | 14:27.38 |
| Try: plink norbertj@184.73.189.105: | 14:28.18 |
| Try: plink norbert@184.73.189.105: Sorry | 14:28.40 |
norbertj | Robin_Watts: yes, plink works (though prompt not nice), putty also works (better link). | 14:30.19 |
Robin_Watts | norbertj: Right, so "export GIT_SSH=plink.exe" | 14:30.45 |
| then try "git fetch golden" | 14:31.08 |
| I suspect you probably had set GIT_SSH=plink.exe in your .profile file, and by reinstalling git you lost that. | 14:31.37 |
norbertj | yes, that was indeed the case. Why can't I remember that :( | 14:32.16 |
Robin_Watts | norbertj: If you're anything like me, you forget stuff when it just works :) | 14:36.43 |
norbertj | Robin_Watts: that's entirely true. So I should write this down somewhere. This is probably the same problem with my home-pc. Will check it when I get home. | 14:38.13 |
Robin_Watts | norbertj: We have a twiki on twiki.ghostscript.com where I try to record stuff like this. | 14:38.50 |
| You would be more than welcome to an account on it. | 14:39.41 |
norbertj | Robin_Watts: just registered there. | 14:48.28 |
sebras | tor8: yo. | 15:18.11 |
| tor8: have you built mujs with gcc-6 recently? | 15:18.19 |
tor8 | sebras: hi. I have not. | 15:18.24 |
sebras | tor8: it complains about indentation in jsY_unescape(). | 15:18.33 |
tor8 | new set of warnings? | 15:18.36 |
sebras | tor8: yeah, now gcc complains about indentation for certain statements. | 15:19.03 |
tor8 | I guess it thinks the multiple statements on one line are in error | 15:19.37 |
| if you wrap the "goto error;" in braces like "if (...) {goto error;} x |= jsY_tohex..." does it shut up? | 15:20.16 |
sebras | yes. | 15:21.00 |
| tor8: thanks for fixing the 100+ warnings about the cmaps. | 15:21.46 |
tor8 | sebras: fix on mujs.git:tor/master | 15:22.00 |
sebras | tor8: ui_input() computes ex, but it is not used. how about nuking it..? | 15:22.55 |
| tor8: I have a set of fixes for compiler warnings on sebras/master | 15:35.26 |
| tor8: using those patches gcc 6.2.0 and clang 3.8.1 can both compile mupdf without warnings (with jpegxr and openssl present and without luratech) | 15:36.38 |
| tor8: oh, and it clustered without warnings now. | 15:46.56 |
tor8 | sebras: LGTM. | 15:47.05 |
| but maybe wait to push until we have 1.10 release properly | 15:47.19 |
sebras | tor8: do you want me to push to master? | 15:47.19 |
| tor8: unless we want these to be part of the release. | 15:47.30 |
| that decision is up to you. | 15:47.34 |
tor8 | no more commits until we can make a proper release. hopefully on monday if nothing crops up over the weekend. | 15:47.50 |
| no more commits on 'master' that is | 15:47.59 |
sebras | tor8: is cool. | 15:48.08 |
| tor8: what's the bugzilla state? | 15:48.15 |
| tor8: I think 697337 can be closed as I updated curl. | 15:48.45 |
tor8 | yes, go ahead and close that one. | 15:49.45 |
sebras | tor8: done. | 15:53.29 |
malc_ | huston we have a problem - lovely "ECMA-376, Second Edition, Part 1 - Fundamentals And Markup Language Reference.pdf" strikes again: mupdf-x11 "/home/malc/x/pdf/ECMA-376, Second Edition, Part 1 - Fundamentals And Markup Language Reference.pdf" | 19:07.28 |
| error: invalid page object | 19:07.28 |
| mupdf: error: cannot open document | 19:07.28 |
| | 19:07.28 |
Robin_Watts | malc_: Can you put that file somewhere please? | 19:10.37 |
malc_ | Robin_Watts: i'd rather not - 43M :( | 19:12.13 |
Robin_Watts | Then there is nothing we can do. | 19:12.25 |
malc_ | Yes you can, like googling it for instance yielding http://www.ecma-international.org/publications/files/ECMA-ST/ECMA-376,%20Fourth%20Edition,%20Part%201%20-%20Fundamentals%20And%20Markup%20Language%20Reference.zip | 19:14.22 |
Robin_Watts | That's the Fourth edition. | 19:16.29 |
| I need *exactly* the same file as you have. | 19:16.44 |
| I googled for it, failed to find it, which is why I asked. | 19:16.58 |
| I'm not going on a snark hunt. | 19:17.08 |
malc_ | ~/x/rcs/git/mupdf/build/native/mupdf-x11 ECMA-376,\ Fourth\ Edition,\ Part\ 1\ -\ Fundamentals\ And\ Markup\ Language\ Reference.zip Ecma\ Office\ Open\ XML\ Part\ 1\ -\ Fundamentals\ And\ Markup\ Language\ Reference.pdf error: No pages in document | 19:17.48 |
| mupdf: error: cannot open document | 19:17.49 |
| | 19:17.49 |
| rather, | 19:18.23 |
| ~/x/rcs/git/mupdf/build/native/mupdf-x11 Ecma\ Office\ Open\ XML\ Part\ 1\ -\ Fundamentals\ And\ Markup\ Language\ Reference.pdf | 19:18.24 |
| error: invalid page object | 19:18.24 |
| mupdf: error: cannot open document | 19:18.24 |
| | 19:18.27 |
sebras | malc_: so you are opening the .zip-file? | 19:18.27 |
malc_ | so yeah fourth edition exhibits the same problem | 19:18.37 |
| sebras: not anymore | 19:19.27 |
| also - ni hao | 19:19.34 |
sebras | Robin: I see the same error as malc reported. | 19:20.17 |
Robin_Watts | mudraw can draw pages. | 19:20.32 |
sebras | Robin_Watts: so in pdf_lookup_page_number() we see node == null. | 19:21.24 |
| well, a null object. | 19:21.50 |
| Robin_Watts: ehm... pdf_dict_get() returns NULL if the dict is a null object. | 19:23.11 |
Robin_Watts | yup. | 19:23.24 |
sebras | Robin_Watts: then pdf_name_eq() compares NULL to PDF_OBJ_NAME__LIMIT. | 19:23.25 |
| Robin_Watts: is the correct behaviour? | 19:23.41 |
Robin_Watts | Why wouldn't it be? | 19:23.51 |
sebras | should we be returning PDF_OBJ_NULL from pdf_dict_get()? | 19:23.57 |
| I'm not sure. | 19:24.00 |
Robin_Watts | NULL = No object found. | 19:24.16 |
| PDF_OBJ_NULL = null object found. | 19:24.30 |
sebras | Robin_Watts: so then pdf_dict_get() returns NULL and we're trying to compare to PDF_NAME_Type which is < PDF_OBJ_NAME__LIMIT. | 19:27.10 |
| which causes pdf_name_eq() to return 0 | 19:27.21 |
| and therefore we throw. | 19:27.24 |
Robin_Watts | Right. | 19:27.35 |
sebras | from inside pdf_lookup_page_number(). | 19:27.37 |
| this is while loading the outline. | 19:27.50 |
Robin_Watts | We look for "Type" inside node. | 19:27.51 |
sebras | right. | 19:27.56 |
| so there is some link destination that points to null. | 19:28.07 |
Robin_Watts | Right. | 19:28.41 |
sebras | Robin_Watts: yes, the link dest is [ null 0 0 1 ] | 19:28.48 |
Robin_Watts | So we need to catch the throw somehow, and deal with the fact that it's a duff entry. | 19:29.49 |
| Possibly we change the throw to a return 0; | 19:31.12 |
| but that would be bad, as it's an API function, and throwing is really the right thing to do. | 19:32.41 |
sebras | I can confirm that this is a regression since 1.9a | 19:33.27 |
Robin_Watts | Urgh. | 19:34.57 |
sebras | is bisecting. | 19:35.10 |
| 8a07b7fb14f11204a0d840792ab9f4bd54b066e5 | 19:35.57 |
| that commit doesn't add the fz.throw though. | 19:36.26 |
| ok, I'm too tired to debug this right now. | 19:37.18 |
| robin, if you are to look at it a bit (not sure if you're preoccupied) I think that'd be splendid. | 19:37.53 |
Robin_Watts | I will look. | 19:38.01 |
sebras | Robin_Watts: tor8 mentioned wanting to do the 1.10 soonish so. | 19:38.06 |
Robin_Watts | sebras: We've done the rc2 already. | 19:38.15 |
sebras | malc_: thanks for bringing this to our attention. | 19:38.17 |
Robin_Watts | so this should be fixed. | 19:38.21 |
| malc_: Indeed, thank you. | 19:38.25 |
sebras | Robin_Watts: agreed. | 19:38.26 |
malc_ | np | 19:40.59 |
Robin_Watts | I have a fix. | 19:41.59 |
malc_ | fab | 19:42.08 |
Robin_Watts | but I don't like it. | 19:45.45 |
| let me do a better one. | 19:45.51 |
malc_ | 8a07b7fb14f11204a0d840792ab9f4bd54b066e5 is lovely though: 248 insertions(+), 533 deletions(-) | 19:47.42 |
Robin_Watts | tor8, sebras, malc_: http://git.ghostscript.com/?p=user/robin/mupdf.git;a=commitdiff;h=04118b6432d6662aaf05c10a82e2de117805e466 | 20:24.51 |
| Forward 1 day (to 2016/11/18)>>> | |