| | <<<Back 1 day (to 2019/12/09) | Fwd 1 day (to 2019/12/11) >>> | 20191210 |
| ta | 2001 Coated. | 12:03.43 |
| kens | Is that a question ? | 12:04.26 |
| ta | yes. | 12:04.42 |
| kens | Then I do not understand it | 12:04.51 |
| leot | Hello folks! | 15:15.22 |
| | A tiny patch for jbig2dec: https://www.NetBSD.org/~leot/tmp/0001-Adjust-JBIG2VERSIONGREP-to-avoid-not-supported-chara.patch | 15:15.50 |
| | (I have noticed that when updating graphics/jbig2dec package on pkgsrc, VERSION wasn't properly substituted because in BRE the `+' character couldn't be used, but we can easily convert the [0-9]+ (in possible ERE) to just [0-9][0-9]*) | 15:17.04 |
| kens | I've asked the relevant developer to step over and look, because I don't know enough to have an informed opinion. But it may take a little time | 15:17.43 |
| leot | kens: Don't worry! If you need any possible further rationale or if something isn't clear do not hesitate to ask! | 15:18.11 |
| chrisl | kens: Hm, well, the "right" developer is away just now, but I don't see a problem with the change | 15:30.48 |
| kens | Well I was thinking of you rather than sebras, since its a build thing | 15:31.09 |
| | I know its not me :-) | 15:31.47 |
| chrisl | Yeh, but that was all added by sebras for the last release (or so) | 15:31.51 |
| kens | Oh, didn't realise that. Maybe I should send him an email. | 15:32.11 |
| chrisl | Or, just don't build jbig2dec on its own!! | 15:32.18 |
| kens | Alternatively, leot could you raise a bug report ? | 15:32.20 |
| | It'll be next week before sebras looks at this, since he's in New York right now | 15:33.21 |
| chrisl | Also.... I think it's been done | 15:33.57 |
| kens | goes to look | 15:34.13 |
| | Not in the version I'm seeing | 15:35.03 |
| | Still has a + there | 15:35.10 |
| chrisl | VERSIONGREP="sed -e s/.*[^0-9\.]\([0-9][0-9]*\.[0-9][0-9]*\).*/\1/" | 15:35.34 |
| | Oh, Isee | 15:35.49 |
| kens | Yeah that's not what I see, but I'm looking a the Ghostscript version | 15:35.50 |
| | Its JBIG2VERSIONGREP in my autogen.sh | 15:36.15 |
| chrisl | Yeh, I don't know where there are two in there | 15:36.29 |
| kens | Oh, VERSIONGREP is higher | 15:36.30 |
| chrisl | Anyway, I'm quite willing to make the change, and take the blame | 15:36.57 |
| kens | :-) | 15:37.05 |
| | I'll leave it to you then | 15:37.10 |
| chrisl | leot: Commit pushed | 15:42.07 |
| kens | Thanks chrisl | 15:42.18 |
| leot | Thanks chrisl! | 17:46.23 |
| | chrisl: The patch was against the jbig2dec repository - sorry, I didn't known that ghostpdl one was the main(?) one... Are jbig2dec release generated from there or the jbig2dec repository? | 17:47.51 |
| | If that's the latter, can it be applied to the jbig2dec repository too? (it should apply cleanly there too) | 17:48.35 |
| chrisl | leot: The jbigdec repo is generated automatically from the ghostpdl one (I think once a day), then jbig2dec releases are (usually) created from the jbig2dec repo | 17:49.20 |
| leot | Oh, that's great, I'll keep that in mind for sharing possible further patches, thanks! | 17:49.44 |
| | I have another little one for ghostscript -- https://www.NetBSD.org/~leot/tmp/patch-Resource_Init_pdf__main.ps.patch | 17:53.43 |
| | The real problem was: when printing a $RANDOM PDF file via cups (and a printer that uses ghostscript, IIRC gdi device probably) 1 page was successfully printed, then I had a page with ghostscript error, another page printed and then another page of error, and so on. | 17:55.17 |
| chrisl | Well, I'd guess the reason that's printing to stdout is that, given we (by default) actually try to (and usually succeed in) continuing the interpretation of the file, it's more of a warning than an "error" | 17:57.10 |
| leot | Yes, however IMHO it's okay when using ghostscript interactively (e.g. to convert a file) that the error is printed on stdout (however stderr it's probably better to be used), it's not okay when printing to ends up printing a paper with just one line of error | 17:58.50 |
| chrisl | For that, you'll really have to open a bug at https://bugs.ghostscript.com/ so it gets seen by the relevant developer | 17:59.55 |
| leot | OK! Thanks! | 18:00.01 |
| chrisl | FWIW, I'm not saying printing to stdout is correct, I'm just saying that *may* be the reason it currently works that way | 18:00.45 |
| | <<<Back 1 day (to 2019/12/09) | Forward 1 day (to 2019/12/11)>>> | |
Log of #ghostscript at irc.freenode.net.