| | <<<Back 1 day (to 2018/03/26) | 20180327 |
| HenryStiles | >/quit | 00:42.25 |
| deekej | chrisl: hello :) It might be worth to add a small note into README.md that bugs should be reported directly at https://bugs.ghostscript.com/. | 08:30.31 |
| | https://github.com/ArtifexSoftware/ghostpdl-downloads/issues/8 | 08:30.43 |
| chrisl | deekej: The front page already says: "This is purely for downloads, please check the website for full information http://www.ghostscript.com" | 08:33.26 |
| | kens: This looks like a simple omission/oversight - comments? https://git.ghostscript.com/?p=user/chrisl/ghostpdl.git;a=commitdiff;h=8531185a1b | 08:36.58 |
| kens | SOrry was fetching coffee | 08:42.38 |
| | It looks reasonable to me | 08:42.57 |
| | I can't believe it was as inrtended before | 08:43.20 |
| chrisl | I'm going to also include the configure checks as I have below that change | 08:44.40 |
| kens | Makes sense | 08:45.24 |
| chrisl | Hmm, compile fail on the cluster, but not here... balls :-( | 08:45.35 |
| kens | Oh, not good :( | 08:45.45 |
| | All the nodes too | 08:46.06 |
| | undefined reference to FT_Get_Font_Format ? | 08:46.26 |
| chrisl | Yeh.... so I understand the compile fail - I don't understand why it doesn't fail here...... | 08:47.19 |
| kens | Good Karma ? | 08:47.32 |
| chrisl | Freetype have been through a file reorg.... | 08:47.52 |
| kens | I guess you tried a clean first ? | 08:47.53 |
| | Oh..... | 08:48.02 |
| | Hmm, a fair few diffs there, are they alll FT do you think ? | 09:00.40 |
| | Oh dear that first one is bad | 09:02.24 |
| | Some kind of path problem there | 09:02.38 |
| chrisl | Yeh, FT 2.9 seems to badly break something :-( | 09:03.22 |
| kens | Just outlines maybe ? | 09:03.39 |
| | ie uncached | 09:04.01 |
| chrisl | No | 09:04.03 |
| kens | Oh :( | 09:04.07 |
| | Number 9 is a bad looing breakage with pdfwrite | 09:04.43 |
| chrisl | Fairly sure these will be cached glyphs: https://ghostscript.com/~regression/chrisl/compare5.html#67 | 09:04.48 |
| kens | hmm yeah quite small sizes | 09:05.29 |
| chrisl | There was a Postscript file throwing an error, I'd guess #9 is throwing an error, too | 09:06.01 |
| kens | I'm sure it is yes | 09:06.16 |
| | Tha'ts exactly what I'd exepct to see | 09:06.23 |
| | That's a pain really, I wonder what it is.... | 09:07.30 |
| chrisl | Well, I think I'll that on the back burner, get the other libs up to date, then circle back to freetype | 09:07.31 |
| kens | Shout if you want a hand with it. | 09:07.51 |
| chrisl | I'm fairly sure the broken glyphs are hinting | 09:08.28 |
| kens | Oh, double bad :-( | 09:08.36 |
| chrisl | Hmm, except disabling hinting doesn't change the brokenness - hmm | 09:13.33 |
| kens | Odd | 09:13.46 |
| chrisl | Well, onto libpng and libtiff, for now..... | 09:14.09 |
| kens | OK | 09:14.20 |
| | I hope they turn out easier | 09:14.32 |
| Rutkay | Hi all, I couldn't found about gs 8.70 documentations. I know it is | 10:42.07 |
| | I know it is pretty old | 10:42.12 |
| | But we're using rhel6 and it has the latest version of ghostscript 8.70 :/ | 10:42.28 |
| | Yesterday I asked about font issues. Today I'm looking for my image rendering | 10:42.53 |
| | Problems | 10:42.56 |
| | Can you please help me? | 10:43.02 |
| kens | Don't ask if you cna ask a question, just ask your question | 10:43.24 |
| | FWIW you can also check out old versions of Ghostscript form our Git repository, which includes the documentation. Though the documentation *should* have been supplied with the package from the distribution. | 10:46.04 |
| Rutkay | I want to make the investigation by myself while I can. If there is a documentation, I can try to look for my problems. Unless, the original pdf https://ufile.io/nkrmw and output tiff https://ufile.io/nc2dp | 10:46.32 |
| | Okay I'm looking for the repository | 10:47.00 |
| kens | You haven't explained what you see as a problem | 10:48.45 |
| | Given that its a full page A4 form, there's a lot to look at and I can't see anythign wrong | 10:49.09 |
| | Oh, and you haven't told us what command line you used to create the TIFF file | 10:49.57 |
| Rutkay | The problem which I thinking about is pointy lines while the original has straight lines and also barely not seen red writings | 10:56.16 |
| kens | Where do you see 'pointy' lines ? | 10:56.34 |
| Rutkay | I'm using : gs -q -dNOPAUSE -sDEVICE=tiffg4 -r204x196 -sOutputFile=$outfile $tmpfile | 10:56.39 |
| kens | The red text is hard to see for several reasons. The frst is that you are creating a low resolution output file. | 10:56.57 |
| | On top of that, you are going to 1-bit monochrome output | 10:57.08 |
| Rutkay | All boxes has pointy lines :/ | 10:57.11 |
| kens | So the result is halftone screened | 10:57.16 |
| Rutkay | What do you mean by 1-bit monochrome output | 10:57.45 |
| kens | Ummm, exactly what I said ? | 10:58.36 |
| | You started with a colour PDFD file. | 10:58.49 |
| | G4 fx is monochrome (black and white) | 10:59.02 |
| | And its one bit per pixel | 10:59.09 |
| | So colour needs to be converted into black and white. | 10:59.18 |
| | THat means that colour is first converted into a shade of grey | 10:59.47 |
| | THen the grey is represented by applying a halftone screen. | 10:59.59 |
| | You cna think of this as 'dithering' | 11:00.07 |
| | So that when viewed from a distance, teh individual black and white pixels merge to appear as a grey colour. | 11:00.31 |
| | Now if an object is drawn in pure black or white, then it will be unaffected by the halftone. | 11:01.23 |
| | However, you appear to have started with a JPEG image of the page. | 11:01.34 |
| | Which means that even the 'black' areas are not actually black, they are just a very dark grey | 11:01.48 |
| | So they too are halftone screened | 11:01.55 |
| | With the result that they are not composed entirely of black pixels, but include white pixels as well, hence what I **think** you are referring to as 'pointy' lines. | 11:02.25 |
| | If you have a newspaper with photographs in it to hand, look closely at the pictures and you will see they are composed of many small dots. That's how halftone screens work. | 11:03.10 |
| Rutkay | So If I increase the resolution, there will still be my "pointy" lines while we can't see because of the high resolution? | 11:03.15 |
| kens | However the newspaper is probably printed at 1000 dpi, so the screen cells are smaller nad the effect is less visible. | 11:03.43 |
| | If yo uincrease the resolution then yes, the effect is less noticeable | 11:03.59 |
| | But Fax is a really low resolution device. | 11:04.09 |
| Rutkay | Thank you Kens. I'm trying to increase the res and try with it | 11:04.28 |
| | I try -r1000x1000 and the result is https://ufile.io/urye7 . Lines are gotten better but red writings are disappeared | 11:08.55 |
| | Do you have any comment about this :/ | 11:09.08 |
| kens | You're using an old version, so no not really | 11:10.25 |
| Rutkay | I cannot migrate to a new version because of the rhel6 support :( | 11:10.52 |
| kens | Of course you can. | 11:11.12 |
| | Install Git, clone our repository, build from source. | 11:11.29 |
| | Tried it with current code and it works fine. | 11:11.51 |
| Rutkay | It generates an error about glibc 1.4 but I have 1.2. These are huge migrations to us .. | 11:12.11 |
| | It is good to hear it though | 11:12.19 |
| kens | Well I don't really see what you expect me to do about it | 11:12.26 |
| | I can't support an 8 year old version of the code. | 11:12.44 |
| Rutkay | Nothing, only complaining :d | 11:12.55 |
| kens | And even if I were to find a bug and fix it for you, what would you do ? You can't install the code. | 11:12.59 |
| | You could complain to Red Hat | 11:13.11 |
| Rutkay | Okay don't be mad about me :d | 11:13.20 |
| kens | Well, don't complain to me then :_) | 11:13.30 |
| Rutkay | I'm appreciated to your support | 11:13.30 |
| | :) | 11:13.37 |
| | Have a nice day, It seems I'm okay for today, too :D | 11:14.10 |
| | Thanks again, see you :-) | 11:14.18 |
| kens | BB | 11:14.23 |
| greg__ | Hi! I've to understand how works Ghostscript to print PDF directly on printer. I've c#.net app and I'm able to print directly on diffenrent printers with sdevice=pxlcolor. Sometime the result is perfect and other time it is not correct. I've also tried with postscript sdevice=ps2write and I don't get constant result on different printer. It seems I would need to use sdevice=mswinpr2 to have good result but then I will have | 13:43.51 |
| kens | You have not stated what you mean by 'not correct' when using the pxlcolor device, or what you mean by 'constant result'. You can't reasonably expect different printers to produce exactly the same output, and its up to you to work out whether a pritner requires PCL, PostScript or 'something else'. | 13:45.31 |
| | Nobody can help you with such a vague statement (without examples), and its not clear what you are asking anyway. | 13:45.59 |
| greg__ | ok. I would say if I use the printer hp laserjet p2055dn with pxlcolor then I get 630 pages with symbol instead of my 2 pages PDF. If I use ljet4 driver then I get the 2 pages PDF. But ljet4 don't manage color. Which ghostscript driver should I use to print on most laserjet printer pxlcolor is it the more reliable driver? | 13:55.50 |
| kens | Well that depends on whether the HP printer in question supports PCL-XL or PCL5 and that's not a question I can answer for you | 13:56.55 |
| | It sounds to me like the printer you tried doesn't support PCL-XL, so the pritner treats the data as text and tries to pritn it, with predictably poor results. | 13:57.25 |
| | You could instead use one of the laserjet drivers which produce (I think) PCL5 | 13:57.44 |
| | I believe there's a laserjet colour device. However, you should be aware that PCL is *much* more closely tied to the device than PostScript is. A PCL file which is prepared for one specific HP printer is not going to pritn the same (or indeed, sometimes at all) on other HP printers. PCL-XL is more device independent. | 13:59.24 |
| greg__ | ok thank for the answer! Then if I I understand the PCL file render by ghostscript can not be read by all printer? hp printer and ricoh printer has different pcl language. Is postscript is more universal. Sorry to ask this question maybe I'm out of the track | 14:59.51 |
| | the goal is to send the pdf file on different printer (different brand) without have the printer installed. I don't know if I could do it with ghostscript in all cases | 15:03.21 |
| kens | When we talk about the pxl devices we are not strictly talking about rendering | 15:05.30 |
| | The PXL devices are (at least in part) high level output devices. They don't render to a bitmap, they interpret the input, and produce high level output (that is, fonts, curves, lines etc) rather than a bitmap. | 15:06.07 |
| | The same is true for the ps2write and pdfwrite devices (though obviously those produce PostScript and PDF respectively) | 15:06.39 |
| | Every pritner has sublte (and sometimes not so subtle) differences in the way it interprets and renders high level input. | 15:07.13 |
| | To take a simple example, the unprintable margins of the devices are probably different. | 15:07.41 |
| | In addition, some printers will support features (eg Duplex) which others do not. | 15:08.00 |
| | And of course, some printers support PostScript, some support PCL, some support both. Some support XPS and some support PDF. | 15:08.37 |
| | SO realistically there's no way for me to tell you 'use this device and it will always work' because I can't know which printers you are going to be using. | 15:09.03 |
| | If you use the mswinpr2 device then that *is* a low level device. The input is interpreted and rendered to a bitmap. The bitmap is then drawn on a Windows Device Context suitable for the selected pritner. Then the device context is told to draw itself, which results in it being sent to the pritner device driver. | 15:10.16 |
| | The pritner device driver 'packages' the bitmap in a way suitable for the printer, and sends the result to the pritner, which pritns it. | 15:10.43 |
| | The key point in this chain is the Windows printer device driver. | 15:10.56 |
| | That device driver is specific to the pritner, and so it *knows* how to do all the things I've just been saying that you are guessing at. | 15:11.27 |
| | So realistically the only way to do what you seem to want to do is to use the mswinpr2 device. | 15:11.43 |
| | Alternatively, on Linux, use CUPS, which is sort of like a similar system for Linux. | 15:12.06 |
| | Please note; Ghostscript is not 'free' software, it is open source and licenced under the AGPL. You should carefully read the licence before you decide whether you can legally use it without a commercial licence. | 15:12.44 |
| | Oh, and of course, if you don't have the printer installed, you can't use the mswinpr2 device anyway | 15:14.34 |
| | You can use Ghopstscript to create either a PostScript, PCL, PCL-XL or XPS file for you, but Ghostscript has no way to know which output you need, you have to tell it. Which means you have to know what the printer expects and, especially for PCL, which exact Ghostscript device best matches the specific PCL pritner. | 15:15.54 |
| greg__ | thanks kens your explications are clear! Yes I have read to AGPL licence. Then what you are saying if a printer support postscript and I use ps2write I should have acceptable result (not talking about margin and duplex)? I think there is a lot of printer which support postscript but I've tried this driver and the output it huge specifically when the pdf contain images. | 15:48.24 |
| kens | Yes, if you use ps2write you should get decent results from any printer supporting PostScript. PostScript is generally quite device=independent, most implementations are quite sound, and the ps2write device has been hammered a bit, so its known to work with quite a few devices. | 15:49.23 |
| | Are you sure you are using ps2write ? If so your PDF files may contain transparency | 15:49.53 |
| | If they do, then they cannot be retained as 'high leve', and must instead be rendered to a bitmap (PostScript does not support PDF transparency) | 15:50.28 |
| | In which case they will be rendered at the default resolution of 720 dpi. | 15:50.41 |
| | If the printer has a lower resolution you can easily reduce that, which will result in smaller files. | 15:50.59 |
| greg__ | ok! | 15:51.09 |
| kens | For example, 150 dpi is usually sufficient for good results on a 600 dpi printer. | 15:51.17 |
| chrisl | FWIW, neither does PCL/PXL, so the same input *should* result in fairly large output from the pcl/pxl devices, too | 15:51.21 |
| kens | Yes, I was going to say that too :-) | 15:51.30 |
| | But they may have a lower default resolution | 15:51.43 |
| | I'd expect the PXL devices to default to 600 | 15:51.51 |
| chrisl | Which is why I didn't say "similarly large output" | 15:52.06 |
| greg__ | ok if I choose to use ps2write and use print server do I need to install postscript driver on this server? | 15:53.03 |
| kens | If the pritner supports PostScirpt, then no. | 15:53.25 |
| greg__ | ok | 15:53.32 |
| kens | But it rather depends on the nature of the 'print server' | 15:53.48 |
| greg__ | let say I'm sending a job from my computer on print server \\server1\printer1 and not directly on printer1 is there a difference? | 15:54.47 |
| kens | In the past a 'pritn server' woudl be a server with printers attached, when you print to the print server, you actually would send a device-neuttral (GDI in Windows terms) file to the print server, which would then do the job of running the GDI through the printer driver and sending the result to the printer | 15:54.53 |
| | OK that's not exactly a 'print server' that's just a remote printer | 15:55.16 |
| | I would assume that the remote computer would have the relevant driver installed :-) | 15:55.51 |
| | How do you propose to send the output to the printer when it isn't attached to the local computer ? IP protocol ? Something else ? | 15:56.42 |
| | Provided you can send the filef directly to the printer then it will work. If you can't do that, then I don't think it will work at all | 15:57.01 |
| greg__ | ok then maybe it's better to send it directly on the printer as PCL or Postscript if the printer support rather then the printer server because the printer server will need the specific driver | 15:59.09 |
| | Just do know what is your title? Printer god master?? :) | 15:59.44 |
| kens | The server doesn't need a driver, but you need some way to direct it to send the file unchanged. | 15:59.46 |
| greg__ | ok | 15:59.57 |
| kens | is a Senior Software Engineer with Artifex Solutions Inc. | 16:00.03 |
| mvrhel_laptop | Solutions? | 16:00.29 |
| kens | Ins't that the company name ? | 16:00.38 |
| | I oculd be mistaken :-) | 16:00.46 |
| HenryStiles | huh? | 16:00.46 |
| mvrhel_laptop | How long have you worked here ;) | 16:00.58 |
| chrisl | Artifex Software LLC | 16:01.11 |
| kens | Long enough for the Alzheimer's to cut in obviously | 16:01.15 |
| HenryStiles | you learn the company name in the second 10 years | 16:01.17 |
| kens | Software, that's the word | 16:01.22 |
| | It begins with S, close neough | 16:01.35 |
| greg__ | if I want to use ghostscript in my business and there is no user interaction (just a service that read pdf folder and print it) do I need licence or I can use it for free? It's not to resell | 16:04.22 |
| mvrhel_laptop | Say that when you call your wife Sara | 16:04.25 |
| | kens ^^ | 16:04.35 |
| kens | greg__ : I cna't really answer that question. | 16:04.43 |
| greg__ | ok because I'm not a lawer to understand all in the APGL licence | 16:05.59 |
| chrisl | greg__: that sounds very like a SAAS case, but if you can't make your own judgement reading the license, you should get legal advice - we're just engineers | 16:05.59 |
| Robin_Watts | mvrhel_laptop: I'll leave you to look up "bronco" on the urban dictionary. | 16:06.13 |
| kens | If you are not distributing Ghostscript, nor running a software-as-a-service application whcih makes use of Ghostscript, but only using it locally in your own in-house workflow then you are probably OK. However, I am not a lawyer and I cannot offer you any indemnity. If you are in any doubt you should seek an opinion from a competent legal professional. I'm just an engineer | 16:06.18 |
| greg__ | ok thanks! | 16:06.57 |
| ray_laptop | chrisl: FYI, it is "Artifex Software Inc." (not LLC) | 16:17.09 |
| chrisl | Uh, oh, I thought it said LLC on the bank transfers..... | 16:17.57 |
| kens | Hmm tht bug report looks bogus to me | 16:18.34 |
| | Someone else reporting cppcheck warnings | 16:18.47 |
| ray_laptop | greg__: The page at https://artifex.com/licensing/ might help explain the licensing (then again, maybe not) | 16:19.15 |
| kens | Although looking at it, it does seem odd | 16:19.16 |
| | Thos do look like they should be && not || | 16:19.33 |
| | Of course, its the UFST interface..... | 16:19.50 |
| | So no use to anyone in the open source world | 16:20.01 |
| | I didn't realise we even compiled that code when UFST isn't there | 16:20.43 |
| chrisl | We don't | 16:20.50 |
| kens | :-) | 16:20.57 |
| | So he's just using a static analysis tool then | 16:21.09 |
| chrisl | I could make that check neater | 16:22.13 |
| kens | It kind of looks wrong | 16:22.27 |
| | It does look like hte avluation is always true | 16:23.04 |
| | Maybe I'm mis-reading it | 16:23.21 |
| chrisl | I'll have a look at it tomorrow - I can't remember why I didn't just use strncmp() | 16:24.08 |
| kens | Shall I assign it to you ? | 16:24.17 |
| chrisl | Sure | 16:24.20 |
| kens | Done | 16:24.44 |
| chrisl | Even if it was "right" it seems like overkill, since the files will either be .fco or .FCO | 16:26.33 |
| kens | right..... | 16:26.50 |
| HenryStiles | es | 17:41.14 |
| Robin_Watts | mvrhel_laptop: Historically we've generated such repos from gs automatically. | 18:42.13 |
| | but that probably doesn't play nice with wanting to be able to merge in changes from marti. | 18:42.30 |
| | Forward 1 day (to 2018/03/28)>>> | |
Log of #ghostscript at irc.freenode.net.