| <<<Back 1 day (to 2015/07/14) | 20150715 |
Mirza | Hi Good morning | 06:52.25 |
| Error6error LNK2001: unresolved external symbol _fz_new_context_impE:\MuPDF\pdfreader\PDFReader\PDFReader\MuPDF.WinRT\Win8\MuPDFDoc.objMuPDFWinRT | 06:52.35 |
| could anybody tell me the solution for above error ? | 06:52.52 |
chrisl | Mirza: I don't recognise that path as being part of our mupdf distro | 07:03.01 |
Robin_Watts | Mirza: hi. What version of MuPDF are you using? | 09:07.57 |
kens2 | Back in a couple of hours | 09:19.55 |
Robin_Watts | chrisl: hey | 12:16.11 |
| I am tempted to hack a quick gs lib build together for what I need, while you sort the rest of your commit out. | 12:16.47 |
| chrisl: pcl/pl/pl.mak still has a FIXME: in it (near pconf.h) | 12:24.42 |
chrisl | Robin_Watts: I think if you can hack something up for your needs that might be best - the Windows stuff is more involved than I'd hoped. And I think the Android thing is going to be a PITA, too..... | 12:30.29 |
Robin_Watts | chrisl: So we've lost gs/jpeg/install-sh, and we've gained libpng/install-sh | 12:41.14 |
chrisl | Robin_Watts: Do we care? | 12:47.27 |
Robin_Watts | chrisl: I just thought it was odd, and possibly an indication of a file that had been moved to the wrong place. | 12:47.50 |
| but on closer examination, they are clearly different versions of the same script, but that script has been imported from somewhere else. | 12:48.14 |
chrisl | They ship with the libraries | 12:48.47 |
Robin_Watts | Did you change versions of any of these things? I'm confused as to why we might get new files in what was essentially a reorganisation. | 12:48.55 |
chrisl | Because the versions were updated on master, it meant some messing around to get it all working. I didn't worry too much about retaining the history for the third party libs | 12:49.53 |
Robin_Watts | chrisl: I've got all your changes boiled down to a single commit here. The net effect is that we lose one install-sh and gain another. Are either of those files actually used? | 12:51.00 |
chrisl | Not ever | 12:51.13 |
Robin_Watts | So we could remove those changes from the commit? | 12:51.44 |
chrisl | We could. I'd rather sort them out, though. I'm trying to keep the third party libs as pristine as we possibly can | 12:52.35 |
Robin_Watts | Did I mention the FIXME in winlib.mak (the one that says "these should be conditional" about jpegxr and expat?) | 12:59.24 |
| I think I did, and you said it should just have the FIXME removed. | 12:59.36 |
| There are lots of targets on pcl/pcl_top.mak that seem to have been commented out. Perhaps they should just be removed? | 13:02.49 |
| Ok, that's really me finished poking the commit with sticks, honest. | 13:04.44 |
| lunchtime. | 13:04.47 |
henrys | images of kens with his laptop riding the horse ;-) | 13:20.08 |
kens | :-) | 13:20.32 |
| Erratic WiFi though | 13:20.41 |
chrisl | Back in a couple of hours..... | 13:51.20 |
kens | I think Marcos has broken the cluster | 13:54.05 |
| I keep getting compilefail on gxps yet the gcc log shows no error. | 13:54.32 |
| I tried going back to the HEAD and it fails too, which it didn't do on the last commit when I pushed it yesterday | 13:54.56 |
| (const gs_font_base *)base_font)->nearest_encoding_index(const gs_font_base *)base_font)->nearest_encoding_index(const gs_font_base *)base_font)->nearest_encoding_index(const gs_font_base *)base_font)->nearest_encoding_index(const gs_font_base *)base_font)->nearest_encoding_index(const gs_font_base *)base_font)->nearest_encoding_indexOK so if I just test gs it works. | 14:02.10 |
| Bah | 14:02.13 |
| Stupid insert key | 14:02.19 |
| As I was saying... If I just test gs then the cluster works. | 14:02.30 |
| Hmm, but throws 1230 errors on xpswrite, presumably because it can't run gxps | 14:07.13 |
marcosw | kens: yeah, my fault. give me a sec. | 14:16.15 |
kens | Ah hi Marcosw | 14:16.23 |
marcosw | kens: give that a go. | 14:16.49 |
kens | Give me a minute to get the results from the current run | 14:17.05 |
| I told it to do gs and pcl | 14:17.20 |
| I'll put another one through in a minute | 14:17.45 |
| Looks like that's OK now marcosw, thanks! | 14:23.47 |
Robin_Watts | chrisl: OOOOh! There is a gs.a target that builds a lib. | 14:52.42 |
henrys | Robin_Watts: thanks for that Robin_Watts I was struggling with the diff | 14:59.15 |
Robin_Watts | henrys: Yeah, that made it managable. | 14:59.40 |
henrys | Robin_Watts: running the orginal I get: warning: inexact rename detection was skipped due to too many files. did you see that also? | 15:05.58 |
Robin_Watts | henrys: Yeah. | 15:06.40 |
| I upped the config variable to 10000 :) | 15:06.50 |
henrys | Robin_Watts: I'm doing the diff between branch tips so I need to regenerate your script | 15:10.36 |
Robin_Watts | henrys: Run the command I gave (possibly use -M80 rather than -M90), then sort the result. | 15:11.34 |
| Then you can easily see any files that should be renames but appear as D/A ones. | 15:12.10 |
| henrys: https://www.indiegogo.com/projects/bseen-this-led-pack-keeps-you-your-phone-alive#/story | 15:17.25 |
henrys | interesting | 15:19.27 |
| chrisl_away, Robin_Watts I wonder how long it will take me to stop cd'ing old directory style. | 15:43.03 |
Robin_Watts | henrys: "ln -s . gs" :) | 15:43.23 |
henrys | chrisl_away: why do we need to set the languagelevel to 2 in gs_pdfwr.ps with the rearrangement of files? | 16:00.56 |
Robin_Watts | chrisl: I wonder if gs.a doesn't include gsromfsX.0 | 16:04.35 |
| chrisl: I wonder if gs.a doesn't include gsromfsX.o | 16:04.40 |
kens | Goodnight all | 16:19.48 |
chrisl | henrys: the for setting languagelevel in gs_pdfwr.ps is because the order of the init files has changed slightly - tbh, the fact it didn't do it before was, IMO, wrong | 16:33.07 |
| Robin_Watts: I thought the gs.a target was only for testing.... | 16:33.35 |
Robin_Watts | chrisl: That's good enough for me :) | 16:37.29 |
chrisl | What I mean is, I thought it wasn't complete, maintained etc etc...... | 16:37.56 |
Robin_Watts | chrisl: Well, it seems to work well enough to give me a library that links and doesn't complain of missing functions etc. | 16:38.23 |
chrisl | Oh well, shows what I know...... ;-) | 16:38.44 |
Robin_Watts | I disabled every device I could and it was missing a few symbols. I fixed the dependencies on my new device (ahem) and added the romfs back in, and it seems to have sorted the linking issues. | 16:39.28 |
| Of course, the MuPDF exe is now 20+Meg rather than 5... | 16:39.43 |
| haha. I was thinking "It's taking ages to install, this must be MUCH bigger than before". | 16:40.20 |
| Then I realised I'd unplugged my phone. | 16:40.27 |
chrisl | Hmm, still a bit more than I would expect.... | 16:40.38 |
Robin_Watts | SEGV. | 16:41.54 |
| but that may be that I have some build flags wrong. | 16:42.08 |
chrisl | Robin_Watts: This Windows stuff is proving a huge headache :-( | 16:54.49 |
Robin_Watts | chrisl: Sorry. | 16:55.02 |
| If it's impossible (or unfeasible) then fair enough, we should drop it. | 16:55.17 |
| cos it probably means it'll be hard to maintain. | 16:55.25 |
chrisl | Well, it all works *except* the .pdb file, and that looks like it might be the deal breaker | 16:56.03 |
henrys | chrisl: I'm about half way through but I'm okay with an early commit and review after, I think there will be feedback from the staff once they pull it and it might be the easiest way to get that feedback. Up to you. | 17:02.58 |
chrisl | henrys: Well, it's going to be a few days before the cluster is up and running, and marcos has already spotted a problem, which I'm looking into..... | 17:04.11 |
henrys | chrisl: oh okay | 17:04.37 |
Robin_Watts | wossup with the .pdb file? | 17:04.41 |
henrys | didn't know about that | 17:04.43 |
Robin_Watts | Oh, simulaneous access from the builds? | 17:04.58 |
chrisl | Robin_Watts: Yep, exactly | 17:05.05 |
Robin_Watts | Does adding /FS help? | 17:05.26 |
chrisl | Possibly. I'll try it - didn't know about it. | 17:06.33 |
Robin_Watts | I think I had to add /FS when parallelising the FBS. | 17:08.17 |
chrisl | Seems like it should be the default these days | 17:09.20 |
Robin_Watts | Yeah. | 17:11.01 |
| It may be set by default in new solutions. | 17:11.16 |
henrys | chrisl: so the countof() operator doesn't work with io_device_table in some place so the new structure members count and size? Is the table opaque? | 17:13.24 |
| I guess it is ... nvm | 17:13.49 |
chrisl | You'll have to give me more context to answer that...... | 17:14.12 |
henrys | chrisl: you added 2 fields io_device_table_count and io_device_table_size to lib_ctx | 17:15.29 |
Robin_Watts | countof only works for foo table[10]; not for foo *table | 17:16.10 |
chrisl | henrys: Oh, right. It's because we can now have empty entries in the table. The count and size were are because I plan to make it dynamically resizeable at some point | 17:17.04 |
henrys | chrisl: I see got it | 17:17.23 |
chrisl | It became a pain because it's (badly done) in gc memory :-( | 17:18.10 |
henrys | ah NUM_RUNTIME_IODEVS... missed that | 17:19.45 |
chrisl | So there is a static table in the graphics library, and then a number of empty "slots" reserved so an interpreter can install it's own io devices (required for Postscript) | 17:21.19 |
| Robin_Watts: /FS is only available on VS2013 and later :-( | 17:38.21 |
Robin_Watts | hmm. How come this works for the FBS then when that runs on VS2010 ? | 17:38.54 |
chrisl | Well, VS2005 throws an error, and MSDN only lists it for 2013 and 2015rc | 17:39.33 |
Robin_Watts | Ok, so I add /FS for vs2013 :) | 17:45.18 |
| So either there isn't a problem for older versions, or there is a different fix. | 17:46.17 |
chrisl | One possibility is to switch /Z7 - but I don't know if the pdb format actually gains us stuff we'd then lose | 17:48.25 |
| Robin_Watts: ^^ | 17:48.29 |
Robin_Watts | yeah. /Z7 sounds like a good option. | 17:50.22 |
chrisl | My other idea was to have an explicit "all" project, and set the other to "not build" - so you can force each product to build, but if you build the solution, it only builds the "all" project" | 17:51.44 |
Robin_Watts | chrisl: Ah, that may be an ideal solution. | 17:52.14 |
| (or close enough to ideal for our purposes) | 17:52.30 |
chrisl | So, go with that for now, and possibly revisit this when I'm feeling inspired? | 17:53.46 |
Robin_Watts | chrisl: Yeah. | 17:54.36 |
chrisl | OKay, I will sort that out - thanks | 17:54.52 |
Robin_Watts | That removes the nasty "better rebuild gs before pcl" thing that worried me. | 17:54.56 |
chrisl | TBH, I was hoping there would be a way to force only one build to happen at a time, but without the dependency in there - but there isn't | 17:55.38 |
Robin_Watts | Ok, so what's -15... | 18:01.25 |
chrisl | gs_error_rangecheck | 18:02.04 |
Robin_Watts | OK, so I'm getting a rangecheck error back from gs... | 18:03.01 |
| Let's do a debug build of gs... | 18:05.12 |
fredross-perry | I just took my Android tablet apart and fixed a hardware problem with scotch tape. I thought someone should know. | 23:24.22 |
Robin_Watts | fredross-perry: touch screen ribbon cable ? | 23:28.42 |
| I had to do that with my mother in laws Nexus 7. | 23:29.02 |
fredross-perry | accelerometer cable. But probably the same thing. Stick a piece of something in there to apply pressure to the connector. | 23:29.31 |
| I shouldnât complain, it was $100 US. | 23:29.52 |
Robin_Watts | yeah, sounds like the same thing. | 23:30.21 |
| Forward 1 day (to 2015/07/16)>>> | |