| | <<<Back 1 day (to 2015/07/20) | 20150721 |
| Robin_Watts | chrisl: And another commit that moves that stuff into windows. | 00:11.12 |
| henrys | wow customer list with dollars for the top ten | 00:56.51 |
| mvrhel_laptop | blah. will need to talk about some ideas for the names of the icc profile device parameters tomorrow. All my names are a little long. Shortening them makes them very cryptic. | 05:02.30 |
| | done for the night | 05:02.36 |
| pavan | Hi all , how can i get winmd file for mupdf ? | 05:37.53 |
| | for my winodws store app | 05:38.06 |
| ManDay | Hello, I've been trying to do impose different cropboxes on different pages of a pdf in /EndPage. According to pstack, /EndPage is, throughout the main part of the run, only called with stacks 0 0 0 and 2 1 2 ! Can someone explain this? According to p427 of the spec, /EndPage should be called with exactly 2 integers, one of which indicate the number of previous showpage executions (which, I hope, amounts | 05:56.42 |
| | to the number of pages) | 05:56.44 |
| | Oh, I think I can suspect what's going on. The 2 1 2 is probably indicating device deactivation, which resets the showpage counter | 05:59.08 |
| | Why is this happening and how should I cound pages? | 05:59.26 |
| coelebs | how to convert fonts to outline while creating .eps? | 05:59.56 |
| ManDay | Anyone here who knows how to get the current page number? I'm despairing over this... | 06:33.36 |
| | I tried to keep track of it manually in the userdict, but I seem not to figure out how to use global dictionaries. my combinations of userdict, load and def just produce errors | 06:34.24 |
| | I also tried things like systemdict /PageCount get but it also gives me an /undefined error | 06:34.47 |
| chrisl | ManDay: I would guess that the PDF(s) you are processing have non-standard page sizes, and thus the PDF interpreter does a setpagedevice for each page, thus causing the device to be reset - that might be why you see device deactivation happening | 06:38.54 |
| | coelebs: if you are using Ghostscript 9.15 or later you should be able to use the command line option "-dNoOutputFonts" to degenerate fonts to outlines in ps2write | 06:41.05 |
| ManDay | ty chrisl , could you give me a hint as to how to fix my workarroud? | 06:41.34 |
| chrisl | ManDay: My only suggestion would be to track the page count yourself. I'd have thought keeping the value in userdict would work.... | 06:42.38 |
| ManDay | My "manual counting pages" looked like an initial userdict begin 0 \MyCount def end and then userdict \MyCount get ... in the /EndPage | 06:43.11 |
| | But I get an /undefined error when I try to to get /MyCount | 06:43.22 |
| | / not \ | 06:43.35 |
| chrisl | I guess there much be a save/restore happening somewhere in ther | 06:43.48 |
| ManDay | chrisl: I didn't even want to store anything. just reading it out the page number without incrementing it already produced the error | 06:44.16 |
| | i.e. userdit /MyCount get gives an error | 06:44.26 |
| chrisl | Well, if you haven't defined it, you will get an error - because it's undefined | 06:44.51 |
| ManDay | chrisl: But I did define it, as I said userdict begin 0 /MyCount def end | 06:45.17 |
| | Here is the code: | 06:45.20 |
| chrisl | You just said "I didn't even want to store anything. just reading it out" | 06:45.41 |
| ManDay | http://dpaste.com/0HYHV5M | 06:46.17 |
| | chrisl: Yes, but initially I stored it, of course | 06:46.26 |
| | I mean I'm not incrementing it even, yet | 06:46.33 |
| chrisl | Your parameters are the wrong way round. You're defining a key of '0' with the value /MyCount | 06:47.38 |
| | "userdict begin /MyCount 0 def end" | 06:48.24 |
| ManDay | oh man, how silly of me | 06:49.05 |
| | thanks | 06:49.07 |
| chrisl | ManDay: But it would probably be neater to use the "known" operator, so you handle everything in the EndPage proc | 06:49.41 |
| ManDay | neater? i don't think so. initializing the variable outside of endpage sounds reasonable i think | 06:50.18 |
| chrisl | I tend to think self contained is neater..... | 06:50.38 |
| ManDay | chrisl: Could you have another look? It seems to ignore my incremening the variable. The second pstack (delimited by ###) at the end of beginpage clearly shows /MyCount 1 on the stack but when I load it in the next pass I get 0 back again: http://dpaste.com/17GC9TD | 07:08.08 |
| chrisl | ManDay: you don't ever save the new value back to the dictionary | 07:10.33 |
| ManDay | But doesn't store on l300 do that? | 07:10.55 |
| | I've also tried to replace store by def, but that didn't change anything | 07:11.05 |
| chrisl | Possibly. I consider it better practice to be explicit about which dictionary a key/value gets placed in. Also, I don't see where you actually increment the value...... | 07:13.31 |
| | FWIW, my preference would be have the entire get value, increment and store self contained at the beginning of the EndPage proc. | 07:16.03 |
| ManDay | i increment it right above where it says 1 add | 07:20.31 |
| | anyway, the pstack between ### clearly confirms i have /MyCount 1 on the stack when I call store | 07:20.55 |
| | Could you tell me what you mean by be explicit about which dict its get placed it? | 07:21.12 |
| | I thought the begin ... store end is the only possible way to modify the user dict | 07:21.29 |
| | (and it worked for the initialization) | 07:21.35 |
| chrisl | Well, "store" searches the dictionary stack for a dictionary containing the specified key, and replaces the value in that dictionary. "def" for example, defines the key/value in the *current* dictionary | 07:22.27 |
| | As I said before, there may be a save/restore happening causing the value in userdict to be set back to it's pre-save value | 07:23.28 |
| | ManDay: you can probably use your own dictionary stored in userdict in global VM to get around the save/restore messing with your value(s) | 07:27.18 |
| ManDay | chrisl: sorry, where did you say that? What do you mean by save/restore happening causing a reset? | 07:28.54 |
| chrisl | I made that comment at 07:43 and save/restore is explained in the PLRM | 07:30.09 |
| ManDay | oh i think you are right | 07:35.03 |
| | I've just checked and the store/def DOES work and retain its value within the EndPage | 07:35.25 |
| | but as soon as we get into the next EndPage, it's reset again... | 07:35.35 |
| chrisl | So, by making your own dictionary in global VM, you'll avoid the restore problem | 07:36.08 |
| ManDay | I don't see how nesting another dictionary in the user dict will help the issue | 07:36.14 |
| chrisl | *IN GLOBAL VM* is the important part of that | 07:36.40 |
| ManDay | the userdict is in global VM | 07:36.51 |
| chrisl | No it ain't | 07:37.01 |
| ManDay | oops | 07:37.06 |
| | i can use just globaldict then | 07:37.20 |
| | Yep, that works :-p | 07:37.44 |
| chrisl | Oh, great - I actually thought that wasn't writable..... | 07:37.58 |
| | So, the only thing to be careful of is that you cannot reference anything in local VM from an object in global VM | 07:38.47 |
| | So, if you do want to save any other state information in there, and any of that is a composite object (like a dictionary or an array), it also must be in global VM | 07:39.47 |
| | integers are fine because they are atomic objects | 07:40.05 |
| ManDay | Forgive me if I don't understand what this has to do with atomic objects | 07:40.46 |
| chrisl | What do you mean? | 07:40.59 |
| | A dictionary and an array (and a string and a file etc) contain references to other objects, thus their contents can change as a result of a restore. An integer is an atomic object, it does not contain other objects, therefore it does not change with a restore | 07:43.02 |
| ManDay | ah i see what you mean now | 07:55.00 |
| chrisl | ManDay: It's not important for what you have right now, but a potential mini-gotcha for the future. I mention it because I've made the mistake of forgetting a few times....... | 07:56.36 |
| ManDay | thanks chrisl . if you care, the script is http://dpaste.com/1SR1011 | 08:08.43 |
| pipitas | Originelles Weihnachtsgeschenk für Geeks, Nerds + Mathe-Freaks schon heute bestellen: https://www.kickstarter.com/projects/fbz/knityak-custom-mathematical-knit-scarves . Fabienne braucht noch 89 Schal-Bestellungen in 57 Stunden, und ihr Kickstarter-Projekt war erfolgreich! | 08:18.19 |
| | Ooops⦠sorry. Wrong tab. | 08:18.45 |
| Robin_Watts | chrisl: Morning | 09:11.59 |
| chrisl | morning | 09:12.06 |
| Robin_Watts | Various commits on robin/master | 09:12.24 |
| | I now have GhostPDL.sln and co in "windows" | 09:12.47 |
| | And I have 5 projects; All, ghostpcl, ghostpdl, ghostscript and ghostxps | 09:13.06 |
| chrisl | The All project commit is wrong | 09:13.26 |
| Robin_Watts | In order to make All work, I had to add some targets to the makefile... ah, ok, what's wrong with it? | 09:13.39 |
| chrisl | Well, first, having two "all" targets in makefiles is bad | 09:13.58 |
| Robin_Watts | oops. Did I do that? | 09:14.15 |
| chrisl | Well, you added an all target when there already is one | 09:14.31 |
| Robin_Watts | Where is the other one defined? Let me grep. | 09:15.00 |
| chrisl | gs.mak | 09:15.49 |
| | Otherwise, how would the make invocation have worked before? | 09:16.20 |
| | Also, what you've got will break the build if the pcl and xps sources aren't available | 09:16.58 |
| | The alldebug, all-memento-target and all-profile-target are somewhat redundant | 09:17.50 |
| | And I don't see why the "all" project contains references to loads of files, but not all the files | 09:19.19 |
| Robin_Watts | chrisl: I just pushed a fix for the all project to remove all the files from it. | 09:19.47 |
| chrisl | But why does it reference any files at all? It's purely something to make the builds work | 09:20.51 |
| Robin_Watts | chrisl: As I say, with my new fix it now does not reference any files. | 09:21.14 |
| | I made 'All.vcproj' by copying and modifying ghostscript.vcproj, and I forgot to remove the files. | 09:21.42 |
| | If makefile targets are redundant, then I will happily remove them. | 09:22.17 |
| | I convinced myself last night that there is no way we can do a proper 'all' build using just Visual Studio dependencies. | 09:22.53 |
| chrisl | This approach was specifically to *avoid* using VS dependencies | 09:23.43 |
| | Previously we just recursively called nmake with the appropriate flags and the existing "all" target - I figured we'd continue to do that | 09:24.07 |
| Robin_Watts | chrisl: Ok, sorry, I'm getting confused here. | 09:24.29 |
| | Your plan is to have an 'All' project in VS that when invoked, will just do an nmake call on an 'all' target ? | 09:25.08 |
| chrisl | Yes | 09:25.15 |
| Robin_Watts | Ok, that sounds like exactly what I was aiming for. | 09:25.29 |
| | So it sounds like we were headed to the same place, and I just got the implementation wrong. | 09:25.51 |
| chrisl | And since "all" is the default target, we don't need to have it explicitly on the command line | 09:25.51 |
| Robin_Watts | chrisl: Right. | 09:26.03 |
| | So I can happily remove the 'all' target, and the calls on that. | 09:26.21 |
| chrisl | So, I have now concluded that Visual Studio (2005 at least) is one of the 5 most evil things on the planet........ | 09:26.34 |
| Robin_Watts | now tell me why alldebug, all-memento-target and all-profile-target are redundant... | 09:26.46 |
| chrisl | Tell me why they are needed? | 09:27.06 |
| Robin_Watts | Ah. Is alldebug, effectively the same as 'debug' ? | 09:27.25 |
| chrisl | Yes | 09:27.31 |
| Robin_Watts | Gotcha. I knew there was something I was missing in my understanding. | 09:27.43 |
| | OK, Let me revise the commit. | 09:27.50 |
| chrisl | BTW, all these problems with the project dependencies being still there, the configurations being messed up, and the nmake invocation being wrong.... I fixed (almost?) all of them yesterday morning, and for some reason VS didn't save my changes. It saved *something* because the .sln and .vcproj files changed, but not everything I'd done | 09:30.01 |
| | Also, I renamed the "test" folder to be "hidden" and marked it as hidden, and VS didn't save those changes either. It properly pisses me off! | 09:31.19 |
| Robin_Watts | chrisl: I think you need to close the solution to force it to save. | 09:34.19 |
| chrisl | Robin_Watts: There is an explicit "Save .sln" in the file menu which I used *and* closed down VS | 09:35.10 |
| Robin_Watts | So, how do we solve the "it won't build if we only have the gs sources" problem? | 09:37.38 |
| chrisl | If you use the existing "all" target, that is solved already | 09:38.02 |
| Robin_Watts | The existing all target did not work for me. | 09:38.20 |
| | but maybe that was finger trouble. | 09:38.30 |
| chrisl | I just built both release and debug on the command line, so the make end of things should work | 09:39.44 |
| Robin_Watts | chrisl: So the makefiles cope with the sources being missing? | 09:41.01 |
| chrisl | Yes | 09:41.10 |
| | Robin_Watts: but there was a problem with the targets being incorrectly set, I updated that yesterday | 09:41.22 |
| Robin_Watts | If the makefiles cope, then why won't my solution code? | 09:41.53 |
| chrisl | The targets you added explicitly referenced the targets for pcl and xps | 09:42.32 |
| | So, the existing all target uses the macro "PCL_XPS_TARGETS" and that is set based on whether the respective sources are available - if they are not, then "PCL_XPS_TARGETS" is just empty | 09:45.10 |
| Robin_Watts | I see. | 09:45.20 |
| | So, do you have a fixed version of the makefiles? | 09:45.30 |
| | solution, sorry. | 09:45.41 |
| chrisl | No, my changes were lost - and I only realised when you mentioned your problems yesterday | 09:46.21 |
| | By that time you were fiddling with it, so I thought it better to avoid clashes | 09:46.45 |
| Robin_Watts | OK, let me tidy this up, and then you can critique it again. | 09:46.50 |
| chrisl | But if you pull from golden, you should get the makefile fix | 09:47.16 |
| Robin_Watts | will do that. | 09:47.30 |
| | Are ghostscript-ufst and ghostscript_rt.vcxproj supposed to be there ? | 09:48.01 |
| | Just not linked into the solution? | 09:48.13 |
| chrisl | Yes | 09:48.16 |
| | They probably should be updated, though | 09:48.35 |
| | As things stand, they will likely end up trying to build all the products and, particularly ghostscript_rt.vcxproj probably won't work for anything other than Ghostcript | 09:49.32 |
| | The winrt project, at least, should explicitly use the "gs" target. | 09:52.31 |
| Robin_Watts | Ok, updated versions on robin/master | 09:57.49 |
| | I've updated those so they should have moved properly into the windows dir. | 09:58.02 |
| | but I haven't tweaked the usage of targets etc. | 09:58.11 |
| chrisl | Okay, I'll pull those over.... give me a minute | 10:00.28 |
| | 4 | 10:00.30 |
| | Robin_Watts: The other thing is, I specifically didn't bother with individual profile targets as I figured we don't need profile builds *that* often, so build everything on the occasions we do need it wouldn't be a huge pain - but I did want to solicit other opinions on that | 10:08.31 |
| Robin_Watts | 'individual profile targets' ? | 10:11.12 |
| chrisl | gs-profile, gpcl6-profile, gxps-profile | 10:11.30 |
| Robin_Watts | For consistency it would be nice for those to all be there, I feel. | 10:11.48 |
| chrisl | Okay, I'm quite happy to do that | 10:12.26 |
| | Robin_Watts: I think the All project commit is now okay. I'm still wary about moving the files without soliciting wider opinions | 10:15.23 |
| Robin_Watts | chrisl: Did you see Michaels comments last night? | 10:15.56 |
| | He said (basically) "ew, the solution/projects are in the root". | 10:16.15 |
| | but yes, we should bring it up at the meeting. | 10:16.31 |
| chrisl | Well, they always have been for the Ghostscript release, so...... | 10:16.33 |
| Robin_Watts | chrisl: The single gs project file, was yes. | 10:16.48 |
| chrisl | Actually the three project files..... | 10:17.08 |
| Robin_Watts | oh, ufst and ? _rt ? | 10:18.20 |
| chrisl | Yeh | 10:18.27 |
| Robin_Watts | Had there been 3 when I made it originally, they might not have been in the root :) | 10:18.45 |
| chrisl | And, as I said, I'd really wanted the "All" project hidden :-( | 10:21.43 |
| | Robin_Watts: so, if you want to push the commits up to and including the "All" project one, that's fine with me | 10:24.56 |
| Robin_Watts | chrisl: OK. | 10:33.18 |
| | done. | 10:33.49 |
| chrisl | Cool. I can't imagine anyone objecting to putting the VS files in their own directory, so hopefully that can go in later, too | 10:34.28 |
| Robin_Watts | runs for a bit. | 10:34.50 |
| henrys | chrisl is the man of the week! we might have other monikors for him in a few days. | 14:30.47 |
| | chrisl: is the cluster stuff resolved. | 14:31.15 |
| | ? | 14:31.18 |
| chrisl | henrys: I don't know for sure, but it does appear to be working now..... | 14:31.35 |
| | We'll need the word from marcosw to be sure | 14:31.47 |
| marcosw | morning | 14:32.37 |
| henrys | possibly naive question for marcosw -- this big testing run we do before release. Why doesn't that just run continuosly in the background. Presumably we can fix bugs easier if you find them closer to when they were committed? | 14:32.55 |
| Robin_Watts | Well, the cluster results seem reasonable. | 14:33.39 |
| henrys | some news that Miles talked a bit with Tor and he said he's be back early but didn't specify exactly when, I know I told Robin_Watts but not sure about others. | 14:34.10 |
| chrisl | henrys: AIUI, it's run weekly - hence the "weekly regression report" emails | 14:34.42 |
| marcosw | henrys: we do. the reason we do a final test before release is that there are usually last minute commits. we could just wait a week after the last commit and that would run all the tests in the backgorund | 14:34.55 |
| rayjj | chrisl: that's what I thought as well | 14:35.01 |
| henrys | chrisl: oh sorry I thought you were doing something larger before the release. | 14:35.14 |
| | okay nvm | 14:35.23 |
| chrisl | henrys: the pre-release testing is, at least partly, so we don't have to freeze master during the release process | 14:36.52 |
| henrys | Should we talk about gproof? Robin_Watts is going on vacation and the deadline is drawing near, but I can't think of much to say beyond the obvious. | 14:37.07 |
| Robin_Watts | henrys: When is the deadline? | 14:37.26 |
| henrys | Robin_Watts: it's for the chicago show 2nd week of september | 14:37.52 |
| Robin_Watts | The mupdf changes are (all except for the last few tweaks) committed. | 14:38.04 |
| chrisl | I am rapidly coming to the conclusion that building gs for android will be nothing more than minor enhancements to Robin's "hacky" solution...... | 14:38.21 |
| Robin_Watts | The gs stuff is updated on top of Chris's new branch. | 14:38.31 |
| mvrhel_laptop | What do we want to call the device parameter for setting the post rendering ICC profile which will be used by the gsproof device to create the "proof" RGB | 14:38.55 |
| Robin_Watts | I could probably disable the gprf device from the default build and then commit the device to gs. | 14:39.09 |
| | ProofingProfile ? | 14:39.33 |
| mvrhel_laptop | we already have one of those... | 14:39.39 |
| Robin_Watts | GSProofProfile | 14:39.46 |
| | PostRenderProfile ? | 14:40.06 |
| henrys | I didn't follow that, are you adding some CM to mupdf for that mvrhel_laptop ? | 14:40.13 |
| Robin_Watts | henrys: No, this is within gs. | 14:40.28 |
| mvrhel_laptop | henrys: no gs will be doing some post rendering cm in the device code | 14:40.36 |
| | PostRenderICCProfile was what I was thinking, to go with the other profile names | 14:41.00 |
| henrys | mvrhel_laptop: oh great it was beginning to sound ambitious for September | 14:41.03 |
| mvrhel_laptop | all of which end in ICCProfile | 14:41.12 |
| Robin_Watts | Sounds reasonable then. | 14:41.21 |
| mvrhel_laptop | Robin_Watts: for now we will just have the one | 14:41.36 |
| henrys | so the big thing this week is to hammer on chrisl's build | 14:41.37 |
| chrisl | henrys, rayjj: we also wanted to ask everyone: Robin_Watts, mvrhel_laptop and myself have a preference for moving the Visual Studio solution/project files out of the repo root directory and into a "windows" directory, but I was a little wary about doing so without checking with at least rayjj | 14:41.43 |
| henrys | if you guys can do all the different targets that you use that would be useful. | 14:41.54 |
| | chrisl: that's the way it's usually done | 14:42.35 |
| chrisl | henrys: up until now, Ghostscript has had the project file(s) in the root directory of the release archive | 14:43.14 |
| Robin_Watts | (GhostPDL has not, but GhostScript has) | 14:43.28 |
| henrys | chrisl: not for me ;-) | 14:43.30 |
| rayjj | chrisl: fine with me | 14:43.42 |
| henrys | chrisl: I always use the win32 directory in ghostpdl to build stuff | 14:43.56 |
| rayjj | chrisl: the autogen.sh is still in the root, though, right ? | 14:44.17 |
| chrisl | rayjj: yes | 14:44.24 |
| | henrys: that doesn't exist in Ghostscript release archives | 14:44.29 |
| mvrhel_laptop | a windows directory would be great | 14:44.55 |
| chrisl | As I said, I felt it was right, but also felt it best to check - and make sure everyone will be aware | 14:45.04 |
| Robin_Watts | chrisl: Shall I push that then? | 14:45.23 |
| rayjj | chrisl: so windows build stuff get bumped down and linux stuff doesn't ? | 14:45.25 |
| chrisl | Robin_Watts: yes, I think so | 14:45.37 |
| henrys | chrisl: and presumably all the release scripts have been updated? | 14:45.52 |
| chrisl | rayjj: yes, just because that's conventionally how it is | 14:45.59 |
| | henrys: I haven't updated the commercial release script yet. It's on my list | 14:46.16 |
| | But everything else *should* be okay | 14:46.31 |
| Robin_Watts | chrisl: Crap. conflict city when I pulled in the latest changes from golden. Will resolve them. | 14:46.41 |
| rayjj | chrisl: but it would be hard to decide what to call the linux build directory since it is used for all unix like platforms. I like having windows tucked away, however | 14:46.50 |
| chrisl | Robin_Watts: strange, I only tweaked the makefiles..... | 14:47.17 |
| Robin_Watts | chrisl: No, you changed the projects too. | 14:47.48 |
| chrisl | Oh, thought you had that change earlier - sorry | 14:48.11 |
| Robin_Watts | no worries. Not a big deal. | 14:48.22 |
| henrys | the other thing I had for the meeting is the bugs we're now at 10 | 14:48.27 |
| | which sets of alarms at the home office | 14:48.41 |
| | s/of/off | 14:48.47 |
| | I have X11 that I need to check in with hintak about and another P1 which I'm about to mark Later unless something amazing happens | 14:49.58 |
| | rayjj: I thought we agree the performance drop could be dispensed with quickly? | 14:50.39 |
| mvrhel_laptop | 696101 is an interesting one | 14:51.45 |
| marcosw | I wasn't sure who to assign 101 to. | 14:52.25 |
| chrisl | mvrhel_laptop: that's probably one any of us could tackle | 14:52.27 |
| henrys | mvrhel_laptop: chrisl ? | 14:52.48 |
| chrisl | It seems to me we should fix the calling code rather than post-process the mode string..... | 14:53.07 |
| | Sure, punt it to me | 14:53.16 |
| henrys | and fredross-perry, were you able to identify a few mupdf problems you can work on as time is available? | 14:53.58 |
| chrisl | Bug #696018: everything I read about that suggests a bug in the MSVC run-time, but nothing really solid about the problem :-( | 14:54.22 |
| fredross-perry | In yesterdayâs duscussion we thought Iâd look at the android viewer parts of gsproof. | 14:54.41 |
| Robin_Watts | henrys: I've passed the android side of the gproof stuff over to fred. | 14:54.41 |
| | That should keep him busy :) | 14:54.51 |
| | There are some other mupdf Android UI bugs out there too. | 14:55.50 |
| henrys | fredross-perry, Robin_Watts cool! Normally we do a project and a few bugs on the side but you are part time so I think that's enough | 14:55.50 |
| fredross-perry | mvrhel_laptop: question about using a gs shared library in gsview. Youâre doing that now, right? Are you installing the library as part of your Windows installer? | 14:55.50 |
| | robin: will I find those bugs in bugzilla? Any specific ones youâre interested in? | 14:56.23 |
| henrys | mvrhel_laptop: your other customer bug looks a little scary | 14:56.32 |
| Robin_Watts | Like people want to be able to have a way of going 'back' from links they have followed in a document. I am not keen on using the back button for that as the bug requested. | 14:56.36 |
| mvrhel_laptop | fredross-perry: yes, I am installing the dll | 14:56.50 |
| Robin_Watts | Also, someone wanted a way of jumping to a specific numbered page. | 14:56.53 |
| mvrhel_laptop | henrys: yes, that is going to take some time | 14:57.03 |
| Robin_Watts | You might well be able to make improvements in the UI too. It's just kinda grown the way it is. | 14:57.15 |
| fredross-perry | yes, i am supposed to look at that too. | 14:57.29 |
| chrisl | I need to duck out - will check back in a couple of hours..... | 14:57.31 |
| henrys | mvrhel_laptop: I wonder if we should have a field in bugzilla that says this is going to take a while, which would then show up in the report. That might make the report better for Miles | 14:58.09 |
| rayjj | has anybody tried to reproduce bug 696018 ? (marcosw) | 14:58.30 |
| mvrhel_laptop | that would be a good thing | 14:58.30 |
| henrys | I'm done, anybody else have stuff? | 14:58.35 |
| fredross-perry | no thanks | 14:58.47 |
| Robin_Watts | henrys: Not that I can think of. | 14:58.59 |
| henrys | it's good to know sales@artifex.com is archived on Scott's computer ... | 16:52.07 |
| Robin_Watts | but not *just* on Scotts computer, right? | 16:52.28 |
| henrys | Robin_Watts: the first thing I notices when I joined was sales was just an alias for Scott ugh | 16:53.02 |
| | Robin_Watts: Ted and I are on it now | 16:53.24 |
| Robin_Watts | I'm sure Scott has a highly technical backup system arranged. | 16:53.30 |
| henrys | but I don't get the responses because he doesn't copy me in (forgets). We need a system that records everything that goes in and out | 16:54.37 |
| Robin_Watts | mvrhel_laptop: I have 3 commits on the master branch of my gs repo. | 17:21.24 |
| | The first of those adds the gproof device. | 17:21.35 |
| | but doesn't actually add it to the default builds. | 17:21.49 |
| | So I think it should be safe to push that now. | 17:21.57 |
| | Have you looked over the device enough to be happy to review it? | 17:22.18 |
| mvrhel_laptop | Robin_Watts: yes. please commit | 18:07.06 |
| | I am ready with my commit that changes gs. I will work on the device changes next | 18:07.29 |
| jogux | henry: the email system emobix uses (fastmail) supports archiving /everything/ (it's a requirement for various kind of businesses). I'm sure Office365 must... but I'm sure they'll want more money :-) | 18:23.51 |
| | henry: the alternative would be for sales@ to be a kind of ticketing system CRM type thing. | 18:24.11 |
| Robin_Watts | mvrhel: Pushed. Thanks. | 18:25.30 |
| | The gprf device is on master. | 18:25.39 |
| | If you want to build it on windows, you'll need the commit from robin/android_mupdf_gs_so that adds it to msvc.mak | 18:26.07 |
| henrys | jogux: the problems aren't technical - I'm probably overtepping with this and should stay out of it. | 18:44.38 |
| | Robin_Watts: well do we continue with this or drop it? | 19:00.57 |
| mvrhel_laptop | oh I need to get a chicago ticket too | 20:49.30 |
| | looks like things have gotten straightened out on the cluster for me. After running autogen here prior to the push things are fine | 21:35.26 |
| | still odd, since the cluster should just run autogen itself | 21:35.59 |
| | but with the reorg who know... | 21:36.19 |
| | s | 21:36.21 |
| | ok. general icc stuff pushed. now onto the gsproof device specific stuff | 22:34.44 |
| | Forward 1 day (to 2015/07/22)>>> | |
IRC Logs