| <<<Back 1 day (to 2014/04/13) | 2014/04/14 |
pedro__ | hi folks | 07:52.41 |
chrisl | Good morning | 07:56.21 |
Robin_Watts | Morning | 08:43.29 |
| pedro__: Nice break? | 08:43.45 |
pedro__ | Robin_Watts: yeah, quite nice chilled place and great food | 08:49.46 |
paulgardiner | jogux_mac: rebranded versions of most of the splash screens are on paulg/master. I haven't yet processed the GD or OE versions because I have yet to find the source for the logos. They don't seem to be in the psd files. :-( | 09:09.38 |
jogux_mac | oh dear :-S | 09:09.56 |
| OE? | 09:10.07 |
Robin_Watts | Are the GD/OE versions just tweaks on the existing ones? | 09:10.33 |
jogux_mac | I presume it's just the GD logo | 09:10.51 |
| and I don't think we care about OE | 09:10.55 |
paulgardiner | Robin_Watts: the GD ones are just tweaks, and I can do it at the bitmap level, but it would be much more sensible to base it on vector graphics versions of the logo | 09:12.37 |
| jogux_mac: Ah, I was hoping we wouldn't care about OE | 09:12.56 |
| Hmmm, actually using GIMP, I can probably do the bitmap-level tweaks really very easily now I have the main versions converted | 09:14.17 |
jogux_mac | argh. my brain is infected again. it took me several attempts to figure out why e2label was giving command not found. I've typed it as ue2label several times. | 09:24.49 |
Ziai | hi all! :) | 09:43.42 |
| i am looking for latest gsdll32.dll / gsdll64.dll what is the best source? | 09:44.23 |
kens | the Ghostscript.com downlaods site Ziai | 09:48.32 |
| The current version is 9.14 | 09:48.49 |
Robin_Watts | Quality. MS VS2010 Professional doesn't include a profiler. | 10:38.53 |
| but there is a free one you can download and install. | 10:39.07 |
| Logic--. | 10:39.12 |
kens | No, I think you need the team edition for that | 10:39.13 |
| ROFL | 10:39.25 |
| MS has always been somewhat schizophrenic about its development tools | 10:40.01 |
jogux_mac | robin_watts : uh. awesome. | 10:40.24 |
kens | Seems its not integrated unless you buy ultimate edition, they run from the command line ? | 10:42.19 |
Robin_Watts | kens: Indeed | 10:44.18 |
| So back to 2010 for me. | 10:44.25 |
kens | Yuck | 10:44.27 |
Robin_Watts | 2008 I mean. | 10:44.29 |
norbertj | Robin_Watts: I still have problems (at work) to use 'git cluster' (No supported authentication methods available: server sent: publickey).. So I use the clusterpush on peeves, but this is very tricky, when doing the edits local, push to origin, pull to peeves, clusterpush to regression . | 12:52.34 |
| Robin_Watts: at least that is my understanding of the process. | 12:52.46 |
Robin_Watts | norbertj: can you ssh to casper from your work machine? | 13:28.51 |
norbertj | Robin_Watts: yes | 13:37.07 |
Robin_Watts | norbertj: In the msys git window, do: git remote -v and tell me what it says for cluster | 13:43.20 |
| paulgardiner: Splash screens look OK to me. The Good Dynamics logo suffers a bit from a lack of sharpness, but it'll do. | 13:43.44 |
paulgardiner | Thanks Robin_Watts. Yeah, the originals lacked sharpness. I guess a bitmap-based source was used - quite possibly the one you found. | 13:45.29 |
norbertj | Robin_Watts: cluster regression@184.73.189.105:/home/regression/cluster/gitbridge/ghostpdl (fetch) | 13:45.41 |
| Robin_Watts: (same for push). | 13:45.48 |
| Robin_Watts: bmpcmp shows left = new, middle = old, right = diff? | 13:49.50 |
kens2 | Yes | 13:49.58 |
Robin_Watts | norbertj: Yes. | 13:50.08 |
| norbertj: what ssh are you using for the cluster? plink or msys ssh ? | 13:50.33 |
| (sorry, for the lag, several conversations going on at once) | 13:51.08 |
norbertj | Robin_Watts: kens2: could you have a quick look at my bmpcmp (e.g. 3 link from 0-6, job 5: tests/pcl/dcf.pcl.ppmraw.75.0) looks to me as a progression> | 13:51.14 |
kens2 | Sure, one moment | 13:51.30 |
| dcf.pcl looks like a regresion, text goes missing | 13:52.15 |
| at 600 dpi | 13:52.26 |
Robin_Watts | Yes, looks good to me. | 13:52.41 |
norbertj | putty ssh . If I do a plink casper, I get a beep and strange prompt. | 13:54.06 |
kens2 | OK yes hte 75 dpi ones look much better | 13:54.48 |
norbertj | kens2: at 600 dpi you get multiple tiles for the complete bitmap, and there is, due to my plexity-fix, a shift to the right (as can be seen in the 75dpi), so leftmost tiles are empty. | 13:55.07 |
Robin_Watts | norbertj: Hit Ctrl-D to get out of the SSH. | 13:55.27 |
| Then do plink regression@184.73.189.105: | 13:55.38 |
| and I'll bet you get the same error as before. | 13:56.05 |
norbertj | Robin_Watts: that's correct. No supported authentication available. | 13:56.48 |
| Robin_Watts: plink norbert@184.73.189.105 is working (with beep and strange prompt, due to incorrect TERM setting I think). | 13:58.46 |
Robin_Watts | norbertj: OK, so start putty, and setup a connection to regression@184.73.189.105: | 13:59.19 |
| Tell it what key to use. | 13:59.25 |
| Then save that connection (as 'cluster' say). | 13:59.36 |
| Then you should find that when you plink regression@184.73.189.105: it gets the key automatically. | 13:59.59 |
| At least, I think that's how it works. | 14:00.07 |
henrys | norbertj: do you have an HP duplex reference? There are device differences. | 14:06.31 |
norbertj | Robin_Watts: from putty I can open an ssh to with user regression (with correct key), | 14:07.02 |
| henrys: I have the PCL5 Printer Language Technical Reference Manual (quite old). | 14:07.26 |
Robin_Watts | norbertj: And you saved that connection? | 14:07.45 |
| norbertj: Now try plink again? | 14:07.55 |
norbertj | Robin_Watts: If I plink -ssh regression@IP then the plink window shortly pops up with user: regression and then it's gone again. | 14:09.37 |
Robin_Watts | norbertj: Perfect. | 14:09.54 |
| Now try: git cluster | 14:10.02 |
| I think you'll find the ssh attempt is going through, and you're then being kicked out because you don't have permission to execute bash on that account. | 14:10.32 |
norbertj | Robin_Watts: on my workpc gives me still the same error (No supported authentication methods available). | 14:12.17 |
Robin_Watts | norbertj: echo $GIT_SSH | 14:13.23 |
ray_laptop | I got that (with putty) when I was not set for SSH | 14:14.10 |
Robin_Watts | Maybe we need to run pageant. | 14:14.28 |
| ray_laptop: plink should default to ssh, I believe. | 14:15.13 |
ray_laptop | Robin_Watts: I run paegant, but on the putty settings under SSH I also have attempt "kyeboard-interactive" auth (SSH-2) | 14:15.21 |
Robin_Watts | ray_laptop: Why? No accounts on casper accept passwords. | 14:15.46 |
ray_laptop | Robin_Watts: it goes by the saved settings for that @_____ "host" | 14:16.14 |
| Robin_Watts: it's just what I have set | 14:16.33 |
Robin_Watts | ray_laptop: Indeed. Hence why I told norbertj to create a connection and save it. | 14:16.42 |
ray_laptop | I "load" a known working connection, then change host name or IP (or appearance ...) and save it as the new one | 14:17.45 |
norbertj | Robin_Watts: echo $GIT_SSH -> /c/users/njj/bin/plink.exe | 14:17.57 |
ray_laptop | norbertj: that looks fine | 14:18.09 |
norbertj | Robin_Watts: pageant is running, with my (norbert) keys for peeves/casper | 14:18.31 |
ray_laptop | (assuming that is actually where you have a copy of plink.exe) | 14:18.32 |
Robin_Watts | norbertj: The key you are using; is it protected with a passphrase ? | 14:18.53 |
ray_laptop | norbertj: on the "Connection" "Data" screen, you have your username ? | 14:19.35 |
norbertj | ray_laptop: Auto-login username: regression | 14:20.28 |
ray_laptop | no, you have to log in as yourself | 14:20.41 |
Robin_Watts | ray_laptop: NO. | 14:20.46 |
ray_laptop | Robin_Watts: regression has a special unique key on casper (for the servers to use) | 14:21.22 |
Robin_Watts | ray_laptop: For the git push script, he needs to be setup to log in as regression. | 14:21.40 |
| His git remote setup for cluster is correct. | 14:21.56 |
| norbertj: In putty for your saved connection, the "hostname" entry on the "Session" tab should be "regression@184.73.189.105" | 14:23.17 |
norbertj | Robin_Watts: when adding keys to pageant I have to enter the passphrase. I changed hostname, but still the No supported auth... | 14:25.46 |
ray_laptop | Robin_Watts: I use gitpush.sh all the time, and I don't use "regression" as the user. | 14:26.39 |
Robin_Watts | ray_laptop: Well, I have it set up here with regression@ :) | 14:27.13 |
ray_laptop | and echo $CLUSTER_USER is "ray" | 14:27.35 |
Robin_Watts | norbertj: Is there a reason you are using an IP address rather than a host name ? | 14:28.36 |
norbertj | Robin_Watts: company firewall only knows the ip-address. | 14:29.02 |
Robin_Watts | norbertj: Firewall knows to let you through to 184.73.189.105, but presumably they haven't broken DNS? | 14:29.47 |
| Can you ping ghostscript.com ? | 14:29.55 |
norbertj | Robin_Watts: no pinging to outside world. | 14:30.39 |
| We have to go through a proxy, and for me they opened the peeves/casper ip-addresses. | 14:31.05 |
| henrys: did you have a look at my bmpcmp ? Looking at the 75dpi is easier to grasp. | 14:33.40 |
henrys | norbertj: no I'll look. I am concerned you don't have a reference. I distinctly remember bug fixes about duplexing that supported devices contrary to the spec. I'm afraid we are going to undo those changes. They are somewhere in bugzilla. | 14:36.06 |
| norbertj: this networking problem should result in an error your system administrator can see in the logs if it isn't on our side. Yes? | 14:36.54 |
| I assume Robin_Watts has already looked in /var/log/* for problems on our end. | 14:37.44 |
norbertj | henrys: I think so. | 14:38.04 |
| henrys: first Q | 14:38.16 |
Robin_Watts | henrys: i have not. | 14:39.01 |
norbertj | henrys: where can I get the HP plexity reference? | 14:39.04 |
| henrys: when I print the dcf.pcl file it looks ok (with my plexity mod). | 14:43.42 |
henrys | norbertj: I've always used the hp manual but then we had a bug fix where the device clearly wasn't in line with the spec and we accepted it. I need to search for that. | 14:44.40 |
| you missed one brace in the patch other than that I'm fine with it norberj | 14:45.06 |
| norbertj: ^^^ | 14:45.15 |
norbertj | henrys: if (pcs->paper_source != i) { I'll correct) it. | 14:46.38 |
henrys | Robin_Watts: well I see lots of stuff for user "regression" in auth.log | 14:46.39 |
| somebody is repeatedly submitting a public key for regression | 14:50.18 |
Robin_Watts | henrys: Sorry, I can't give this my full attention at the moment. I have a customer rush on, and stuff is not playing nicely. | 14:51.03 |
henrys | Robin_Watts: no problem. | 14:51.15 |
norbertj | henrys: just pushed the plexity to golden. | 14:55.39 |
| henrys: going home now. | 14:56.01 |
henrys | norbertj: okay thanks | 14:56.08 |
ray_laptop | henrys: you'll see lots of user "regression" in the log from the cluster nodes as they get jobs from the cluster master (they all log in as "regression") | 15:22.19 |
henrys | ray_laptop: no somebody is trying a regression login constantly I don't think it has anything to do with norbert - comcast probably marcosw with some cron thing awry | 15:23.32 |
mvrhel_laptop | good morning | 15:23.34 |
henrys | mvrhel_laptop: welcome back | 15:23.50 |
kens2 | Morning mvrhel_laptop | 15:23.52 |
ray_laptop | mvrhel_laptop: hi | 15:24.12 |
mvrhel_laptop | thanks. | 15:24.13 |
ray_laptop | mvrhel_laptop: good to have you back. | 15:24.21 |
henrys | ray_laptop: sudo tail -f /var/log/auth.log | 15:24.29 |
mvrhel_laptop | ray_laptop: I was actually near you the whole time, but did not turn on my computer once | 15:24.43 |
| we were in disney | 15:24.52 |
ray_laptop | mvrhel_laptop: great ! Hope you had a good time. The weather was (characteristically) good | 15:25.16 |
Robin_Watts | henrys: Every minute, each regression node ssh's in to touch a file. | 15:25.27 |
mvrhel_laptop | ray_laptop: yes we melted the first couple days | 15:25.39 |
ray_laptop | mvrhel_laptop: I hope nobody got too sunburned | 15:25.43 |
mvrhel_laptop | no we are heavy sunscreen users. had to ride a lot of the water rides... | 15:26.12 |
ray_laptop | Robin_Watts: right -- that's how they maintain "up" status | 15:26.23 |
Robin_Watts | ray_laptop: Indeed. | 15:26.37 |
mvrhel_laptop | I was amazed at what they did in cars land. That was not there the last time I went | 15:26.40 |
henrys | Robin_Watts: how does that explain the tail -f report? | 15:27.17 |
ray_laptop | we haven't gone to disney in over a year. We have season passes at knotts and it is closer and we have a parking pass (also we have the "soak city" pass) | 15:27.39 |
henrys | I see repeated logins from the same ip | 15:28.08 |
tor8 | chrisl: I've uploaded mupdf release downloads to casper, do you want to link to them on the downloads page? | 15:28.42 |
Robin_Watts | 98.207.109.212 is apparently in fremont CA. | 15:28.58 |
| So I'd guess it's the borg in Marcosw's garage | 15:29.14 |
| yes marcos.ghostscript.com is 98.207.109.212 | 15:29.40 |
chrisl | tor8: sure, I'll do that now, but I'll leave the "commercial" release until tomorrow | 15:29.48 |
henrys | Robin_Watts: oh that makes sens sort of ⦠we should be getting these auth.log errors | 15:29.57 |
Robin_Watts | Marcos has how many cluster nodes in his garage now? 12 ? | 15:29.59 |
tor8 | chrisl: no rush | 15:30.00 |
henrys | "Bad prime descriptionâ¦" | 15:30.28 |
chrisl | tor8: other that "public image", I'm not sure there's any point to the mupdf commercial releases...... | 15:30.33 |
| s/that/than | 15:31.04 |
ray_laptop | and who is 50.240.204.222 | 15:31.15 |
Robin_Watts | Montery CA. | 15:31.43 |
| Monterey CA. | 15:31.51 |
ray_laptop | yeah, but that's not marcos (and keeps flagging POSSIBLE BREAK-IN ATTEMPT) | 15:32.52 |
tor8 | chrisl: yeah. hence my "no rush" :) | 15:33.08 |
| henrys: Robin_Watts: okay, so the 1.4 windows binary and source tarball are live. | 15:33.55 |
ray_laptop | Robin_Watts: I didn't think he had 12 servers, but a couple of his servers are BIG | 15:33.58 |
tor8 | question: how dead do we kill the google code page now that we don't use it for downloads anymore? | 15:34.06 |
Robin_Watts | ray_laptop: Each uber server is 4 nodes. | 15:34.53 |
ray_laptop | Robin_Watts: I didn't realize he had them divided up. | 15:36.00 |
chrisl | tor8: I couldn't find out how to edit the google code project page to say "this is dead, go here....." :-( | 15:36.27 |
Robin_Watts | AIUI, each one of those 2U servers is actually 4 separate ubuntu instances, each with 12 cores or something. | 15:36.52 |
tor8 | chrisl: the best thing I've found is to point the "tabs" to wiki pages that say "go here instead" | 15:39.50 |
chrisl | tor8: Hmm, I don't think I could find a way to edit *anything* other than the file uploads.... | 15:41.21 |
tor8 | do you have an Administer tab when you visit the project? | 15:41.51 |
chrisl | Ugh, I'll have to sign in..... | 15:42.28 |
tor8 | I just changed the mupdf project's download page to point to mupdf.com instead of showing the list | 15:42.49 |
chrisl | Hmm, no, I don't have an Administrator tab - but I am listed as an "Owner" | 15:43.33 |
tor8 | Administer -> Tabs and then either check the hide checkbox or add "Downloads" to Downloads: to replace the tab with a wiki page | 15:43.50 |
| chrisl: that's odd | 15:43.57 |
| might be that ray needs to grant administer rights | 15:44.12 |
chrisl | Oh, wait, I may have used the wrong google account | 15:44.14 |
| tor8: aha, got it - too many google accounts :-( | 15:44.58 |
henrys | ray_laptop, Robin_Watts : I really hate this cron stuff - it just loads the logs files so it's impossible to find stuff. I guess we can search for norbert's ip in the auth log when we next see him, if it doesn't rotate out before the grep ;-) | 15:46.47 |
Robin_Watts | henrys: Yes. It's a question of me being able to dedicate time to it. | 15:47.35 |
henrys | Robin_Watts: well I wish we could do the clusters differently. Both my machines have the same problem, very hard to notice or diagnose when something goes wrong. | 15:49.09 |
Robin_Watts | henrys: Indeed. | 15:49.22 |
| If I'd been writing the cluster, each node would have kept a permanent connection to the master. | 15:50.19 |
| but maybe I'd have hit problems that marcosw avoided. | 15:50.37 |
chrisl | That wouldn't work as the nodes and master have dynamic IP addresses | 15:50.58 |
henrys | chrisl: yea the dynamic ip is the issue without that either it could be connected or we could use broadcast when the code changes not poll every minute | 15:53.38 |
chrisl | I wouldn't have thought we'd need to log every poll, though - or possibly, have a specific "cluster.log", for it | 15:55.07 |
Robin_Watts | chrisl: The master has a static IP I believe. | 15:55.24 |
chrisl | Robin_Watts: casper is master? I thought it could change on reboots | 15:55.49 |
Robin_Watts | And even if it had a dynamic IP, we have a dynamic dns system set up. | 15:56.04 |
henrys | chrisl: I get an log for the cron every minute on my local machines then on the server auth.log gets a ssh entry | 15:56.04 |
chrisl | henrys: hmm, okay, I'm not sure there's a way round that :-( | 15:57.01 |
henrys | maybe we should at least look at static ip pricing | 16:00.45 |
chrisl | Possibly I'm wrong about casper. But rewriting the cluster would be a *big* job | 16:01.40 |
Robin_Watts | yes, not suggesting we rewrite the cluster. | 16:01.55 |
paulgardiner | tor8: I'm just wondering if I might have been misleading Raed. Do the MuPDF build files provide for building with or without openssl. A quick look has made me wonder whether now we always build with it. | 16:07.40 |
chrisl | wonders why Raed cares....... | 16:08.21 |
kens2 | Presumably we don't run a web server in MuPDF< and therefore don't respond to Heartbeat packets ? | 16:09.21 |
Robin_Watts | chrisl: because he's thinking "oh, heartbleed, I should upgrade everything that uses openssl" | 16:09.27 |
ray_laptop | chrisl: he probablt wants to know if we are susceptible to heartbleed | 16:09.37 |
chrisl | So a simple "don't worry about it" should suffice, then | 16:10.39 |
tor8 | paulgardiner: we build with openssl if it is present in the thirdparty directory or pkg-config can find libcrypto | 16:33.13 |
paulgardiner | tor8: okay thanks. | 16:33.52 |
tor8 | paulgardiner: we don't have a way of turning it off other than editing the makefiles though, we might want to add a HAVE_OPENSSL macro that can be set to "no" like our other optional features | 16:34.57 |
paulgardiner | Yeah might be best. | 16:35.49 |
tor8 | paulgardiner: https://queue.acm.org/detail.cfm?id=2602816 | 16:36.25 |
paulgardiner | tor8: quite. The trouble is, I'm not sure that creating an alternative that is less than 300,000 lines is possible. There are just too many standards all intermingled. | 16:42.02 |
| tor8: am I right in thinking that the final link will drop openssl if the app makes no use of the signature support? We were careful to use openssl only from the signature extensions | 16:45.16 |
henrys | kens2: is alexander's bounty work ready to commit? | 16:45.45 |
| kens:he sent in the CLA and I'm not sure if he is waiting for us to commit | 16:46.50 |
Robin_Watts | paulgardiner: If he builds mupdf and never calls the digital signature stuff, will openssl be dropped on link? | 17:06.18 |
paulgardiner | Robin_Watts: that's what I was just asking tor8. I'd imagine so | 17:31.53 |
kens2 | henrys sorry was off eating. I made one small addition to his code which frees the memory, otherwise its fine. | 18:05.00 |
| henrys I can commit it at an time, I've already cluster tested it | 18:05.19 |
Robin_Watts | Morning marcosw. Thanks for bringing fermis back up. | 18:16.53 |
| Something is unhappy still though. | 18:17.03 |
marcosw | Robin_Watts: I'm surprised it's up. I just rebooted it after enabling remote console access and didn't expect it to be reachable. | 18:17.44 |
Robin_Watts | The VM I am working on keeps giving backtraces in it's vnc console :( | 18:17.47 |
marcosw | Robin_Watts: that doesn't sound like anything I can help you with :-( | 18:19.41 |
Robin_Watts | Trying a different VM to see if it's a VM specific problem. | 18:20.05 |
mvrhel_laptop | bbiaw | 18:31.31 |
henrys | kens2: sounds good. | 19:01.41 |
kens2 | OK I'll commit first thnng tomrorw | 19:01.54 |
henrys | Robin_Watts: sorry just noticed log into Skype message - must have been this morning (my morning) | 19:06.57 |
crazy6 | So, I'm frustrated enough with Powerpoint's inability to use postscript graphics that I came here. Has anyone come across any plugins or workflows that make PS/EPS in powerpoint possible? | 19:17.57 |
| I ask here since, as I understand it, such a plugin would most like really just be calling ghostscript in the background. | 19:18.16 |
henrys | crazy6:windows mac? | 19:28.35 |
crazy6 | henrys: Oh, on Windows. I think Macs do actually support postscript at a system level, thus it works properly in PP & Word? | 19:29.22 |
| And for what it is worth, I can put a PS file into powerpoint, but it seems to render terribly. | 19:29.53 |
henrys | crazy6: one way would be to convert it to pdf then view it in acrobat and copy paste selected graphics. I don't know if you can copy past in the acroread - the free adobe product. | 19:31.35 |
| crazy6: other than that use ghostscript to convert the postscript to a format pp can digest. I don't know much about it ⦠ghostscript stack overflow may be of help | 19:32.56 |
crazy6 | henrys: ah, neat, let me try the acrobat route (we do have Acrobat X license here on campus) | 19:33.24 |
| When a PS is included in a latex file, ultimately making it's way out to a PDF, is it really just a series of translations of vector drawing languages? | 19:35.09 |
henrys | pdflatex does not support eps | 19:40.18 |
| but if you go to postscript yes it is all postscript - then you have to use ps2pdf but I think most folks use pdflatex these days | 19:42.18 |
crazy6 | Hrm, yes, I do use pdflatex via TexWorks, and a package that ends up calling epstopdf, which looks like it is bundled wth MikTeX | 19:52.34 |
| Someone should write a plugin that converts PS files in-situ in powerpoint, rasterizing them at an adjustable resoltion, so it's *like* PP can support them natively... I think that's a bit beyond me, so I better try to convince someone here to attempt it. | 19:59.32 |
fatih | hi | 20:05.39 |
| i want to ask a question about mupdf | 20:05.50 |
| i can use mupdf 1.3 my android project with android 4.42 | 20:06.21 |
| but when i wan to load project other phone with android 2.3.3 | 20:06.45 |
| Forward 1 day (to 2014/04/15)>>> | |