| <<<Back 1 day (to 2015/07/21) | 20150722 |
rayjj | hmm... is the cluster honked up ? My job shows as "running" but all the nodes (those that ate up) show "Aborting run" or "Connection interrupted" | 01:04.40 |
| oh, maybe just a timing issue. Now they show "idle" | 01:05.26 |
| marcosw: I am trying to do a cluster run from my sandbox, and it fails. I haven't yet updated this to the new directory structure, but it looks like the cluster is broken for the old stuff. | 01:15.42 |
| marcosw: please see my log: http://www.ghostscript.com/regression/cgi-bin/clustermonitor.cgi?report=ray | 01:16.08 |
| I've updated my tree to the new structure (build_consolidation) and am trying again.... | 01:23.28 |
| guess I'm done for the day. A bunch of errors from the build: | 01:38.37 |
| cp: cannot stat `/home/marcos/cluster/ghostpdl/gs/libpng/scripts/pnglibconf.h.prebuilt': No such file or directory | 01:38.38 |
| touch: cannot touch `/home/marcos/cluster/ghostpdl/gs/zlib/gzio.c': No such file or directory | 01:38.40 |
| mkdir: cannot create directory `/home/marcos/cluster/gs': File exists | 01:38.41 |
| mkdir: cannot create directory `/home/marcos/cluster/gs/bin': File exists | 01:38.43 |
| make: *** No rule to make target `install-gs'. Stop. | 01:38.44 |
| cp: cannot stat `/home/marcos/cluster/users/ray/ghostpdl/urwfonts': No such file or directory | 01:38.46 |
| -- I don't know which of these are benign, but the last one w.r.t. urwfonts doesn't look good | 01:38.47 |
| note that I _do_ seem to have pcl/urwfonts | 01:47.01 |
| maybe it's because I'm only building 'gs' -- trying without that subset | 01:48.02 |
| well, it got further -- into the "Starting jobs" step (all except for peeves that still had compile fail) | 01:53.36 |
| I disabled peeves and actually got a useful run :-) Tomorrow I'll look into what's up with peeves (probably needs something updated) | 02:10.17 |
mvrhel_laptop | ok. the christ apostle wig is too funny | 04:29.39 |
Robin_Watts | chrisl_away: Do you have admin access to bugzilla? | 11:00.20 |
chrisl | Robin_Watts: no, IIRC, marcosw is the (only?) one with full bugzilla admin rights | 11:31.09 |
Robin_Watts | chrisl: Ah, fair enough. I've sent an email anyway. Thanks. | 11:31.28 |
chrisl | Yeh, saw the mail, just now | 11:31.54 |
| Robin_Watts: has there been any discussion about flights for Chicago? I think I saw a brief one between you and Ken a while back (when I was on vacation), but I don't recall any decisions | 11:32.57 |
Robin_Watts | chrisl: We discussed possibilities, but no firm conclusions were drawn. | 11:33.19 |
chrisl | Okay, we should try to remember early next week, when Ken is back | 11:33.46 |
Robin_Watts | I guess we'll do it all again when Miles gets back to us (which should hopefully be today/tomorrow as I want to get it booked before I go away( | 11:33.50 |
chrisl | I thought Chicago was all confirmed? | 11:34.14 |
Robin_Watts | chicago's dates were confirmed, but Miles didn't say "book your flights" | 11:34.31 |
chrisl | Oh, okay | 11:34.45 |
jhabjan | hi, some person raised an issue that Ghostscript.NET doesn't printing Turkish characters properly. Can somebody please take a look at this file: https://www.dropbox.com/s/h0nuufcsenkps8h/turkish_characters_sample.pdf?dl=0 | 13:12.47 |
| I also did a test with display device (gs 9.16) and it's a same problem... | 13:13.38 |
| ...just noticed .. printing = print, sorry | 13:14.24 |
henrys | jhabjan: please report the problem at bugs.ghostscript.com with a gs command line and test file. | 13:20.18 |
| Robin_Watts: I can change bugzilla if marcosw doesn't get to it soon. | 13:26.25 |
Robin_Watts | henrys: Thanks. | 13:26.38 |
chrisl | jhabjan: the fonts aren't embedded, therefore you get the best guess we can manage. Especially with non-latin fonts, you *must* have the original font available to Ghostscript to have a any chance of consistent/correct output. And the best way to manage that is to embed the font in the PDF - as is *strongly* recommended by the PDF spec. Basically, the answer is "yes" | 13:26.41 |
| Actually, " you *must* have the original font available to any PDF consumer" | 13:27.15 |
jhabjan | chrisl: got it, thanks | 13:27.57 |
chrisl | jhabjan, and, in fact, the Ghostscript backchannel prints loads of stuff about substituting CIDFonts, and referencing where in the documentation this is discussed | 13:29.23 |
jhabjan | chrisl: ok | 13:30.56 |
henrys | chrisl: are you up on the url clickable issue ken was looking or maybe I should say "not looking at"? | 13:40.30 |
chrisl | henrys: there is no issue - Ken has no plans to look at it further. | 13:41.08 |
henrys | chrisl: I wanted to understand if the url is only recognized for WinAnsi or is it just recognized because it is known encoding vs. a custom encoding. | 13:42.30 |
chrisl | henrys: it's only recognised when in WinAnsi | 13:43.09 |
henrys | chrisl: that's odd | 13:43.47 |
chrisl | henrys: This is Acrobat we're talking about...... | 13:44.08 |
henrys | chrisl: I know but you'd expect if acrobat recognized the characters it would just do it... it's a top ten customer so we will bend over backwards if it can't be fixed so be it. But I want to make sure there is no way of addressing this. | 13:46.11 |
| money is good all that. | 13:48.14 |
chrisl | henrys: it can't be fixed - Acrobat and Distiller do undocumented and unfathomable things, and we've never had any luck working out exactly what it does. Unless Adobe to document it, we're pretty much stuck. Ken spent quite some time a couple of years ago investigating it, and couldn't make progress. | 13:48.27 |
| henrys: Obviously, Ken would be able to give you more details...... | 13:49.19 |
henrys | chrisl: yeah I'll talk to him some more when he's back. thanks | 13:49.48 |
chrisl | NP | 13:49.56 |
| Heading off for a couple of hours...... | 13:50.22 |
rayjj | I've built the current tree on peeves and didn't have any problems, so maybe it was a fluke last night. I've reenabled peeves. If anyone has a problem, feel free to disable it until we can figure it out | 14:56.54 |
| I did a test cluster run and peeves seems to work OK. I guess it was still in a confused state from trying to use the 'gs' option. I'll try that later | 15:16.00 |
Robin_Watts | rayjj: Michael had problems with the cluster. | 15:16.32 |
| He said he had to have run an autogen.sh on it before it would work, I think. | 15:16.46 |
| (I pushed from windows, where I have never run autogen.sh, and that worked too) | 15:17.01 |
henrys | sees there are a few mupdf's on github | 16:45.57 |
chrisl | henrys: only one official one....... | 16:47.08 |
henrys | chrisl: right | 16:48.24 |
hyper_ch | hello there, I have a strange issue. I'm trying to convert a .pdf file to .ps using. However not output file is created and the whole process seems rather to be slow: https://paste.debian.net/284912/ | 16:48.26 |
chrisl | hyper_ch: -sDEVICE not -sDevice | 16:49.23 |
hyper_ch | chrisl: much faster and works :) | 16:49.48 |
chrisl | hyper_ch: the reason it's faster is because what you had there was the default bbox (bounding box) device (hence the printed output), which runs at *very* high resolution in order to get accurate numbers out | 16:50.45 |
hyper_ch | chrisl: no idea what you just said :) | 16:51.21 |
chrisl | hyper_ch: not really important..... | 16:51.48 |
hyper_ch | thx for the help | 16:52.05 |
chrisl | NP | 16:52.17 |
henrys | oh I do wish we had proper errors for those things. | 16:53.35 |
chrisl | That's simply not possible..... | 16:53.56 |
henrys | I know just moaning | 16:54.48 |
chrisl | We could possibly make oft used strings like DEVICE more tolerant of cap/lower case combinations | 16:55.17 |
hyper_ch | no the question is why doesn't it work in the script.... | 16:55.21 |
henrys | chrisl: yeah I'll put it on the list | 16:55.51 |
chrisl | hyper_ch: eh, huh? | 16:56.00 |
Robin_Watts | chrisl: I suspect hyper_ch meant "now, the question is..." | 16:56.31 |
hyper_ch | chrisl: I switched to nixos and now all my bash scripts need upgrading :) | 16:56.36 |
| and yes, that's what I meant :) | 16:56.41 |
chrisl | Oh, okay. I can't help with that...... | 16:57.24 |
hyper_ch | wasn't expecting you to :) | 16:57.58 |
chrisl | hyper_ch: surely bash is bash, and *most* of the core utilities are common to almost all Linux variants, so I wouldn't have thought much would need to change..... | 17:00.52 |
hyper_ch | chrisl: it's the way nixos works... nixos puts stuff into nix store to enable atomic builds | 17:01.28 |
chrisl | hyper_ch: sure, I've heard about nixos (no direct experience though), but I rather than all that would be mostly hidden from normal user space.... | 17:02.40 |
hyper_ch | chrisl: well, each upgrade/rebuild will make a new profile... and symlink stuff.... so you can rollback easily | 17:03.28 |
chrisl | hyper_ch: still, I'd hope that would only possibly raise an issue during an update cycle. Except possibly for a few tools like "find" that have to be told how to handle symlinks, I suppose | 17:05.29 |
hyper_ch | well, for my script collection, gs is actually found there right now: /nix/store/mnwbk5sspmcnmac2aicxrgzaln8v2z6v-ghostscript-9.15/bin/gs | 17:07.03 |
chrisl | Oh, I see.... I'd assumed there would still be a link /usr/bin or similar..... | 17:08.07 |
hyper_ch | well, current profile is at /run/current-software/sw/bin/ | 17:09.05 |
chrisl | I guess I rarely use the full path to the executable and mostly rely on PATH being correctly set, so it would almost never be a problem for my scripts | 17:10.31 |
hyper_ch | it works a bit different on nixos... you can have multiple versions installed installed of a program and to call a specific one you have to use the according path | 17:11.21 |
| is there a way for gs to make verbose output to a file? | 17:12.03 |
chrisl | more verbose than usual? | 17:12.35 |
hyper_ch | well, I run the scripts as action scripts from dolphin | 17:13.20 |
| so I don't see output | 17:13.23 |
chrisl | I usually just pipe to a file when I need that | 17:13.43 |
hyper_ch | how? | 17:14.19 |
chrisl | Normal bash redirection | 17:14.40 |
hyper_ch | --> /nix/store/mnwbk5sspmcnmac2aicxrgzaln8v2z6v-ghostscript-9.15/bin/gs -sDEVICE=ps2write -o "${curPDF}.ps" "${curPDF}" >> "gs.txt" ? | 17:15.05 |
chrisl | If you do gs -sDEVICE=ps2write -o "${curPDF}.ps" "${curPDF}" >> "gs.txt" 2>&1 | 17:16.27 |
| that will redirect both stdout and stderr to the file | 17:16.41 |
hyper_ch | thx | 17:16.59 |
chrisl | The '2>&1' redirects stderr to stdout, and your '>> "gs.txt' redirects stdout to the file | 17:17.39 |
hyper_ch | I never understood that 2>&1 thing | 17:18.00 |
Robin_Watts | Send file descriptor 2 to file descriptor 1, AIUI. 2 is stderr, 1 is stdout. | 17:18.43 |
chrisl | Exactly as Robin_Watts says..... | 17:19.31 |
hyper_ch | no gs.txt file was created :( | 17:20.15 |
chrisl | hyper_ch: you may need to specify the full path, and make sure the directory in question is writable by the effective user when the script gets run | 17:20.59 |
hyper_ch | it is writable | 17:21.18 |
| it's a tmpdir | 17:21.21 |
chrisl | hyper_ch: is this the dolphin the PHP related thing? | 17:23.34 |
hyper_ch | no, it's kde file manager | 17:23.45 |
chrisl | Oh, okay | 17:24.00 |
| Well, I'm not sure what to suggest - it definitely works for me, because I use that construct *extensively* in my Ghostscript test scripts | 17:24.42 |
hyper_ch | chrisl: it worked on ubuntu just fine... and I think I have it now | 17:25.08 |
chrisl | From what you say, I think I'll be steering clear of Nixos.... ;-) | 17:26.04 |
hyper_ch | nah, it's just me being a noob :) | 17:26.45 |
chrisl | Well, it's also that it's an interesting idea, but it's a solution for a problem I haven't run into enough to worry about | 17:28.57 |
hyper_ch | what I like about it, is that you have a configuration file and just with that you setup the computer/server to your needs | 17:29.31 |
chrisl | It is probably worthwhile for servers, yes, I can see that | 17:30.13 |
hyper_ch | and for desktops and notebooks | 17:34.01 |
Robin_Watts | Customer asks a question. Marcos answers it. Customer "says "thanks for the quick response" and then asks exactly the same question again. | 17:37.10 |
chrisl | Robin_Watts: Does this strike you as a customer we're likely to want enhancements from? ;-) | 17:37.56 |
Robin_Watts | Strikes me potentially as that being manglement rather than engineering. | 17:39.29 |
chrisl | One can hope.... | 17:39.56 |
mvrhel_laptop | hmm. the visual studio solution is screwed up a bit | 18:33.45 |
| in gs | 18:33.49 |
hyper_ch | so, no script works again just fine :) | 18:34.09 |
mvrhel_laptop | Unfortunately I dont have VS8 to fix this as easily as others | 18:34.33 |
| Robin_Watts: do you have a sec | 18:35.21 |
| to check something for me | 18:35.25 |
hyper_ch | s/no/now/ | 18:36.26 |
mvrhel_laptop | darn I just realized I have had this commit sitting here in another checkout for making sure that we honor the fast color option in the pdf14 device. I discovered that issue some time ago, had a fix and never followed through with it. Trying to rebase it but now I have delete modify conflicts... | 18:44.01 |
chrisl_away | mvrhel_laptop: what's the problem with the solution? | 18:48.51 |
mvrhel_laptop | hi chrisl_away : oddly the files in the solution located in base/color have the wrong path | 18:49.31 |
| at least in my converted one. I was just getting ready to look at the solution file bug got distracted by this other thing. | 18:50.00 |
| s/bug/but/ | 18:50.07 |
chrisl_away | mvrhel_laptop: it's probably easier to fix that by editing the XML rather than doing it in the GUI | 18:50.23 |
mvrhel_laptop | yes | 18:50.28 |
| I can do that | 18:50.30 |
| I will take care of it | 18:50.36 |
chrisl_away | OKay, thanks. | 18:50.51 |
mvrhel_laptop | np | 18:50.56 |
chrisl_away | Actually, there are loads of files haven't had the "..\" added so quite a number will have the same problem | 18:51.49 |
mvrhel_laptop | ok | 18:52.00 |
chrisl_away | So, it's up to you if you want to do it, or leave it to me or Robin | 18:52.21 |
mvrhel_laptop | if its not done when you come back tomorrow then you can get it. I may just do it now to get it taken care of | 18:52.53 |
| trying to make sure I don't lose this old work I did right now.... | 18:53.25 |
chrisl_away | Okay, I'll check in the morning. I'm disappointed that you had rebase problems, I'd hoped git would handle it okay :-( | 18:53.52 |
mvrhel_laptop | chrisl_away: it may not have been what you did. it was at least 6 months ago that I did this. that directory had been sitting there | 18:54.20 |
| or that work | 18:54.24 |
| and I completely forgot about it | 18:54.29 |
chrisl_away | mvrhel_laptop: do you have the pdf14 work in a commit? | 18:54.33 |
mvrhel_laptop | I did. and I pushed it to a patch | 18:54.44 |
| I will go back through and fix it up later | 18:54.56 |
| It wont take too long | 18:55.08 |
| But its been so long I kind of forgot what all I had done.... | 18:55.23 |
chrisl_away | What I'd do is create a branch from that commit, then you can mess up all you want on master, the original commit remains intact on the branch. When you're happy, delete the branch | 18:55.41 |
mvrhel_laptop | when I was working on stuff for the big customer for ray I had this stuff going. yes that is a good idea | 18:56.07 |
chrisl_away | I used that approach a *lot* when doing this build stuff! | 18:56.30 |
mvrhel_laptop | thanks chrisl_away : ok I am going to see if I can fix up the solution file now | 18:56.43 |
chrisl_away | Cool, and I will slink off into the background again...... | 18:57.14 |
mvrhel_laptop | :) | 18:57.18 |
| is jasper even in the project anymore? | 19:09.23 |
| henrys: do you know the answer to this? | 19:13.13 |
| I will leave it for now and commit the current path fixes for the other files | 19:16.56 |
| ok. all committed | 19:37.35 |
| lunch time now. | 19:38.12 |
henrys | mvrhel: we replaced jasper with openjpeg some time back | 21:06.53 |
mvrhel_laptop | henrys: ok that is what I though. I am going to remove the jasper references from the visual studio project then. | 21:42.34 |
| lcms needs to go out of the project too | 22:23.59 |
| as we have lcms2 | 22:24.04 |
| Forward 1 day (to 2015/07/23)>>> | |