| <<<Back 1 day (to 2020/02/09) | Fwd 1 day (to 2020/02/11) >>> | 20200210 |
albertus1 | hi, I don't really know ghostscript, but obviously I rely on it in a couple of ways. now lilypond stopped working for me with ghostscript spitting these errors: https://pastebin.com/jAQScTXG -- any idea what I can do to find my problem? | 18:50.21 |
chrisl | albertus1: Well, something seems to have gone wrong interpreting a font - that's about all I can say.... | 18:52.02 |
albertus1 | that's more than I would have seen from that, thanks | 19:04.14 |
chrisl | albertus1: It is a little odd - I would have expected an invalidfont error, rather than the error you are seeing. But CFF is a type of font, and .parsecff is the operator we use to read those fonts | 19:05.33 |
albertus1 | seems to be TeXGyreSchola-Regular | 19:06.26 |
chrisl | I would guess that starts life as an OTF font | 19:06.50 |
kens | Could try with NOSAFER | 19:07.37 |
chrisl | Does lillypond open font files explicitly? <shudders> | 19:08.49 |
kens | I'm only guessing based on the fact that its 9.50 | 19:09.03 |
chrisl | Yeh, but there's no message about loading a font from disk | 19:09.18 |
albertus1 | hm, first I must figure out how to separate the gs invocation from lilypond, so that I can try different gs arguments | 19:10.24 |
| if this helps, this is the entire output with lilypond's verbose option: http://dpaste.com/0AH4S0M | 19:12.02 |
kens | has to go, getting late | 19:13.59 |
chrisl | From our perspective that's not any more information | 19:14.16 |
kens | If you can get the file that Lilypond writes in /tmp/lilypond.... then we could try it | 19:14.41 |
| Not sure if it would help not having the right fonts and stuff but its possible | 19:14.59 |
chrisl | There are also a couple of lilypond PS files that may or may not be required | 19:15.33 |
kens | I know, but they don't seem to be there. and if they are referenced from teh PS file they won't work, due to safer | 19:16.00 |
| Nothing appears to be getting --permit_fil_read added in the command line | 19:16.27 |
albertus1 | I can have lilypond output a ps file, would that be enough? | 19:16.28 |
kens | Well its the PS file that its trying to run in order to produce a PDF file that we need. | 19:16.43 |
albertus1 | hmm, the ps file only contains the error messages then | 19:16.45 |
kens | In your example the file is called /tmp/lilypond-QdrWvr | 19:17.09 |
| I'm assuming the required lilypond files (eg lilyponddefs.ps) are rolled up into that file | 19:17.56 |
| Ah it says: | 19:18.12 |
| Layout nach »/tmp/lilypond-QdrWvr« ausgeben... | 19:18.13 |
| And then there's some stuff about CFF fonts: | 19:18.40 |
| Eingebettete CFF-Schriftartenliste wird initialisiert. | 19:18.40 |
| CFF-Schriftart ??TeXGyreSchola-Bold?? einbetten. | 19:18.40 |
| So its using CFF fonts (as we noted), I'm assuming that these are embedded in that 'layout' | 19:19.16 |
albertus1 | okay, I've got that tmp file, now let's see where I can upload this | 19:19.29 |
kens | I'm afraid I can't stick around just now, I'll try and look at the logs later. | 19:19.49 |
albertus1 | many thanks so far | 19:20.02 |
chrisl | albertus1: If there's nowhere convenient to share the file, and if it's not too large, we can do e-mail | 19:26.47 |
albertus1 | here's the file, it's below 300 KB: https://v2201904398988001.luckysrv.de/lilypond-QdrWvr | 19:48.39 |
chrisl | Well, I see the same error, which is a good start..... | 19:50.45 |
kens | albertus1: That file fails for me in the same way. | 19:53.37 |
| However, it also fails on 9.27, 9.25.... | 19:53.51 |
| With 9.05 it fails in a slightly different way giving me an error /invalidfont in definefont | 19:54.09 |
| The font in question is Emmentaler-18 | 19:54.19 |
| I wonder if your font has been corrupted ? | 19:54.25 |
| Or, possibly, updated ? | 19:54.34 |
chrisl | An SID of -71 seems bad..... | 19:55.31 |
kens | Ah, you've gone further than me | 19:55.44 |
albertus1 | hmm, I don't really know; all I can say is I'm on Gentoo linux (where the Python2 EOL recently kicked in) | 19:56.00 |
| what is an SID in this context? | 19:56.30 |
chrisl | "String ID", IIRC - but don't hold me to that! | 19:57.21 |
| "SID 0 – 64999 2-byte string identifier" | 19:58.41 |
albertus1 | I realize Emmentaler is the font for musical symbols coming from lilypond | 19:58.42 |
chrisl | So, -71 is definitely *not* a valid value.... now, the question is, are we overflowing something | 19:59.24 |
| Nope, as far I can see, we are genuinely decoding negative numbers from the stream | 20:01.47 |
kens | Yeah, my suspicion is the font is somehow corrupted. Whether on disk or in transit into the PostScript file | 20:02.14 |
chrisl | Well, fontforge freezes and then segfault trying to read the font.... | 20:08.40 |
| Frankly, the CFF data looks like complete garbage | 20:10.40 |
| FWIW, gs reads Emmentaler-18 from disk just fine, so I really think the embedded font is corrupted | 20:18.09 |
| albertus1: Yes, I'm afraid the font data is corrupted. I can't be totally sure without quite a lot more digging, but I *think* it's okay until nearly the end of the "CharSet" data, and somewhere towards the end of that, things fall apart, and never recover. | 20:24.18 |
| (FWIW, Adobe throws an invalidfont error on StartData - which is really what Ghostscript should be reporting, too) | 20:25.57 |
| albertus1: I need to go now, too. I'm not sure what more we can tell you, but kens and/or I are usually around here during UK office hours (approximately!) | 20:27.35 |
albertus1 | this helped quite a bit, thanks! | 20:39.39 |
| <<<Back 1 day (to 2020/02/09) | Forward 1 day (to 2020/02/11)>>> | |