| | <<<Back 1 day (to 2015/11/04) | 20151105 |
| rayjj | I think I have the SMask None clist issue fixed. | 00:17.42 |
| | I had to use a SAME_AS_PUSHCROP_BUTNOPUSH, which interestingly was implemented in the clist_create_compositor, but was never used (it may have been in the past and it got removed???) | 00:21.10 |
| | ah, I see that it used to be used by PDF14_SET_BLEND_PARAMS but was changed to ALLBANDS by mvrhel's commit 096e2102 | 00:58.03 |
| | mvrhel: can you have a look at the fix for bug 696324? http://git.ghostscript.com/?p=user/ray/ghostpdl.git;a=commitdiff;h=e60e2d37d15e5cd6703fe65d4e266fa093a6938a | 00:59.47 |
| sebras | kens: morning! | 06:53.08 |
| SSgtSpoon | Hey all, anyone awake? | 12:58.26 |
| kens | If you have a question, just ask it | 13:04.30 |
| chrisl | kens: when you have minute, could you check over: http://git.ghostscript.com/?p=user/chrisl/ghostpdl.git;a=commitdiff;h=ec5790f2 | 15:01.15 |
| kens | One sec | 15:01.24 |
| | ell it looks sensible to me | 15:07.48 |
| chrisl | Thanks, I'll push it - make marcosw happy (or less unhappy) | 15:08.46 |
| kens | Shoulldd makke us hapy | 15:08.54 |
| | as it was wrong before | 15:09.19 |
| chrisl | True. | 15:09.52 |
| marcosw | less unhappy is an unreachable goal. | 15:10.37 |
| chrisl | I thought it was more reachable than happy | 15:10.53 |
| | I still don't understand what's going on with the piped output - it seems to defy the logic of what I see happening in the code:-( | 15:13.25 |
| mvrhel_laptop | I am going to be out for a bit this morning | 15:40.46 |
| rayjj | wow. A 128Gb usb thumb drive for $29. 120Mb/s read, 20Mb/s write. | 17:27.18 |
| Robin_Watts | would rather Samsung SM951 512GB M.2 PCI-e ACHI SSD | 17:32.41 |
| | 2150Mb/s READ, 1500Mb/s write. | 17:33.28 |
| | Not $29 though :) | 17:34.26 |
| rayjj | Robin_Watts: I agree that is better (for $337) IF you have an AHCI slot. | 17:34.31 |
| | PCIe 3.0 that is. | 17:35.18 |
| Robin_Watts | Sorry, that's MB, not Mb. | 17:35.23 |
| | Yeah, that's the main OS drive on this new machine. | 17:35.49 |
| marcosw | I'm thinking that the Samsung SM951 would make a great bcache drive. | 18:02.06 |
| Robin_Watts | marcosw: There are 2 variants of it. an NVMe one and an AHCI one. The NVMe is better. And I have the wrong one, it seems. | 18:02.52 |
| | marcosw: If your motherboard doesn't have an M2 slot, then you can use a PCI-e riser card (make sure you get one that supports PCI-e M2's not just SATA ones) | 18:03.47 |
| marcosw | it turns out my new mb does have an M2 slot and I was wondering what to put it in. | 18:04.41 |
| | afaict the sm951 is ahci and the sm950 is nvme. the nvme is about 10% faster reading but $20 cheaper on newegg. Why is the AHCI one more expensive? | 18:08.30 |
| Robin_Watts | No, there are both AHCI and NVMe variants of the SM951 | 18:21.02 |
| marcosw | okay | 18:21.19 |
| Robin_Watts | http://www.scan.co.uk/shop/computer-hardware/all/hard-drives-ssd/solid-state-drives-(ssd-pcie)128gb-512gb | 18:21.30 |
| | At least, according to scan. | 18:21.39 |
| marcosw | it looks like you are correct, the nvme model is mzVpv512hdgl and the ahci model is mzHpv512hdgl (emphasis added). Why they chose the second letter of the standard to differentiate the two is anyone's guess. | 18:31.42 |
| | the other weirdness is that the nvme model is not for sale in the US, other than ebay. Lots of websites in the uk and germany will sell you one. | 18:32.34 |
| Robin_Watts | marcosw: These are not designed as retail drives. | 18:33.21 |
| | They were designed for OEM use for building laptops etc. | 18:33.31 |
| marcosw | I'm pretty sure that the ahci model is the same but plenty of places in the us sell it (most include oem in the description, so I'm presuming both the nvme and the ahci models are oem). Having bought both oem and retail hard drives the only difference I've noticed is that the oem ones don't include screws or cables :-) | 18:38.10 |
| Robin_Watts | marcosw: Both are OEM, yes. | 18:38.57 |
| | OEM in europe = minimal warranty (1 year), anything else is down to the retailer. | 18:39.18 |
| marcosw | I did finally find the nvme one on newegg: http://www.newegg.com/Product/Product.aspx?Item=9SIA12K3FG3566&cm_re=mzVpv512hdgl-_-9SIA12K3FG3566-_-Product | 18:39.23 |
| Robin_Watts | In the US you only get 90 days anyway, right? :) | 18:39.36 |
| marcosw | warranty? we don't need no stinkin' warranty! | 18:39.59 |
| Robin_Watts | That seems expensive. | 18:40.17 |
| rayjj | no moving parts. What could possibly go wrong ;-) | 18:40.18 |
| Robin_Watts | And no battery. | 18:40.33 |
| | If they could just stop those bits toggling, they'd be golden. | 18:40.52 |
| marcosw | actually on hard drives it used to be 5 years, then it was 3, and now it's 1 or 2 for most models. But I don't think that varies with oem vs retail. If anything retail is worse (i.e. it's one year for the drives inside external usb enclosures which afaik are only sold at retail). | 18:42.42 |
| rayjj | btw, the spec on the Patriot USB drive was also 120MB/s (not Mb/s) | 18:46.14 |
| marcosw | it is overpriced, but it's a third party seller and they charge $429 for the ahci version, so a $20 difference. newegg sells the ahci version for $369, so the nvme version should cost $389, which is closer to what scan.co.uk sells it for (but it's still more expensive over here, especially after adding 10% sales tax in california). | 18:47.49 |
| rayjj | mvrhel: Can you review my patch (see the top of the logs for the URL). I'd like to commit it. Mainly I want to know if the commit msg is clear enough | 18:57.58 |
| mvrhel_laptop | hi rayjj | 18:59.26 |
| rayjj | mvrhel: You get the nod to review it since your fingerprints are all over that function (and it's transparency related) | 19:01.57 |
| mvrhel_laptop | rayjj: it sounds reasonable to me. | 19:03.48 |
| rayjj | mvrhel: OK, thanks. | 19:04.02 |
| mvrhel_laptop | Not sure I recall all the issues in the group pushes etc. | 19:04.13 |
| | I do recall the one issue we had was related to a mask which just had a different background value | 19:04.40 |
| rayjj | mvrhel: SMask should be local w.r.t. groups | 19:04.42 |
| mvrhel_laptop | and the background value goes beyond the size of the mask | 19:05.02 |
| | has an effect beyond the mask image | 19:05.15 |
| rayjj | mvrhel: right, that's the one I found where you changed PDF14_SET_BLEND_PARAMS to be ALLBANDS | 19:05.22 |
| mvrhel_laptop | not sure if that was the commit | 19:05.30 |
| | ok | 19:05.32 |
| rayjj | you changed it from SAME_AS_PUSHCROP_BUTNOPUSH to ALLBANDS | 19:06.02 |
| | and the comment on that sounds the same as what you just said | 19:06.25 |
| mvrhel_laptop | ok. for some reason I cant search for commits on http://git.ghostscript.com/?p=ghostpdl.git;a=summary for something like 096e2102 | 19:07.26 |
| | how do I find that commit | 19:07.35 |
| | I just wanted to see what I had done | 19:08.03 |
| | anyway back to fooling with mupdf. I wanted to ask tor a couple questions. will try to catch him tomorrow morning | 19:08.58 |
| Robin_Watts | mvrhel_laptop: Anything I can answer? | 19:13.34 |
| | http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=096e2102 | 19:14.28 |
| | Just click on any commit, then paste the sha into the URL. | 19:14.51 |
| mvrhel_laptop | Robin_Watts: maybe. Just want to make sure I am heading along the right path. I started with the command line program to create a pdf. Right now it doesn't do much except make a page and do some drawing into it and then save it out. My question is, are the text objects and images added to the page through the fitz library calls and then will that be written out by pdf_write_document or... | 19:16.37 |
| | ...am I suppose to be adding these at some different level than with fitz calls? | 19:16.39 |
| | Robin_Watts: thanks for the explanation on the sha in the URL | 19:16.51 |
| Robin_Watts | What do you mean by "through the fitz library calls" ? | 19:17.11 |
| | If you mean, by making calls at the fz_device level, then, yes, that is where they SHOULD be made. | 19:17.33 |
| mvrhel_laptop | well for example I can to fz_fill_path(ctx, dev, path, 0, &ctm, fz_device_rgb(ctx), yellow, 1); | 19:17.38 |
| Robin_Watts | Right. | 19:17.43 |
| mvrhel_laptop | ok | 19:17.48 |
| Robin_Watts | That will make the path and add it. | 19:17.52 |
| mvrhel_laptop | yes | 19:17.54 |
| | ok | 19:17.56 |
| Robin_Watts | If you do fz_fill_text() then it'll add the commands to output the text into the stream. | 19:18.12 |
| mvrhel_laptop | ok | 19:18.20 |
| Robin_Watts | BUT it will set it up to use Helvetica. | 19:18.21 |
| mvrhel_laptop | like wise | 19:18.22 |
| | fz_fill_image(ctx, doc->dev, image, &local_ctm, doc->opacity[doc->opacity_top]); | 19:18.24 |
| Robin_Watts | I *think* images properly put everything into the resources tree. | 19:18.46 |
| mvrhel_laptop | ok I understand. I need to add the font | 19:18.46 |
| | I just wanted to make sure there was not some other pdf object stuff that I was missing | 19:19.08 |
| Robin_Watts | I think fonts create a font entry in the resources tree, but it's basically empty. | 19:19.11 |
| | No, I think you're on the right track. | 19:19.23 |
| mvrhel_laptop | Robin_Watts: ok thank you | 19:19.36 |
| Robin_Watts | no worries. I wrote the PDF output device (while on holiday in Malta). | 19:19.57 |
| mvrhel_laptop | nice | 19:20.04 |
| Robin_Watts | Or at least the first version of it. Paul has hacked on it since for the appearance stream stuff. | 19:20.26 |
| | So feel free to ask questions, and I can subject you to my dodgy memory :) | 19:20.45 |
| mvrhel_laptop | ok. I am sure there will be a pile more as I move forward | 19:21.01 |
| Robin_Watts | "What were you thinking?!" | 19:22.14 |
| mvrhel_laptop | I doubt that | 19:25.21 |
| | bboaw | 20:04.07 |
| | bbiaw | 20:04.09 |
| | Forward 1 day (to 2015/11/06)>>> | |
IRC Logs