| <<<Back 1 day (to 2015/06/30) | 20150701 |
chrisl_away | git log will still work fine, git log of a directory, I dunno.... I'm not sure how that could work, given that the directory has changed | 00:01.40 |
| AIUI, git doesn't store the history of a directory, only the stuff inside it | 00:02.22 |
sebras | chrisl_away: true. you never check in a directory. | 00:04.32 |
| chrisl_away: only the files in it. | 00:04.36 |
| which means that if you remove all the files, git will also remove the empty directory. | 00:04.50 |
chrisl_away | is going to bed now.... brain shutting down...... | 00:05.21 |
Anil_ | Hi robin , how are you? | 09:43.37 |
ghostbot | Anil_: pretty good. how about you? | 09:43.37 |
Anil_ | good | 09:44.02 |
| as you told me , we can get winrt 8.1 dll? | 09:44.45 |
jogux | anil_: I'm not sure Robin_Watts is actually here just now (but I'm sure he'll be around in the next hour or two) - I don't know why ghostbot decided to reply to you :-) | 09:54.33 |
Robin_Watts | Anil_: I told you to come back later and we'd ask mvrhel :) | 10:02.47 |
Anil_ | yes | 10:04.46 |
| can we ask him now if he is available.. | 10:04.58 |
Robin_Watts | Anil_: No. He'll be here in about 8 hours. | 10:07.34 |
| He's in a different timezone. | 10:07.42 |
| I believe the answer is that the winrt code used by the mupdf windows rt app is part of gsview. | 10:08.32 |
| As such, it's not yet available under the GNU AGPL. | 10:08.56 |
| We have made it available to customers before though. | 10:09.21 |
| You said you were in the process of sorting out a commercial license? | 10:09.41 |
Anil_ | yes | 10:16.19 |
| before that i have to check each and every functionality of the mupdf | 10:17.02 |
| how smooth it is rather compare to other versions | 10:17.18 |
Robin_Watts | ok. | 10:17.27 |
| I suspect you'll need to speak to sales and explain that you're evaluating mupdf. | 10:18.24 |
| and then maybe they can get you the windows rt stuff for evaluation. | 10:18.39 |
| How much of the winrt stuff do you need? | 10:19.50 |
| Are you wanting just to get mupdf to render bitmaps, and you'll put them on screen? Or are you wanting the whole framework to display pages/scroll/zoom etc? | 10:20.32 |
b80905 | any mupdf devs here? | 10:26.56 |
Robin_Watts | b80905: Sure. If you have a question ask it, don't ask to ask it :) | 10:28.07 |
Anil_ | actually i needed to check annotations, search... those two are enough for me.... | 10:29.25 |
Robin_Watts | Anil_: Well, the easiest way to see them in action would just be to use the Android app. | 10:30.19 |
b80905 | Robin_Watts: I just added a feature to my local copy of mupdf, so that when you quit, all your marks are saved to a file and when you open the same file again the marks are available for use. What do you think? | 10:32.21 |
Robin_Watts | b80905: By all means open a bug on bugs.ghostscript.com and add a patch. | 10:33.35 |
| I can't promise to take it - I'd wantto talk to Tor about it's desirability first. | 10:34.07 |
| kens: There is a version of that commit on robin/device_subclassing that runs through cluster testing with only 60-something diffs. | 11:52.24 |
kens | Ther are some diffs expected | 11:52.40 |
Robin_Watts | Is that the same as the main device_subclassing? | 11:52.44 |
kens | I don't remember how many, let me look at the log | 11:52.58 |
| Yes those are all expected diffs | 11:53.27 |
| I wonder why its throwing you 2 compile wranings as well.... | 11:53.52 |
Robin_Watts | which ones? | 11:54.19 |
| There are 6 warnings in the cluster report. | 11:54.28 |
kens | ignoring return value of freopen | 11:54.30 |
Robin_Watts | I think they are all old ones. | 11:54.36 |
kens | I know about the pie warnigns | 11:54.38 |
Robin_Watts | But the cluster checking for 'are these warnings new' gets confused easily. | 11:55.03 |
kens | I imagine the gxclfile ones are old too, just odd that it suddenly decided to tell you | 11:55.05 |
| OK you want to tell me how to pull those 2 off your repo and onto my local copy ? | 11:57.42 |
| Let me check remotes first | 11:57.52 |
| Yeah I don't have your repo as a remote, let me start with that | 11:58.14 |
| Right, done that. | 12:00.04 |
| I guess I could just git pull --rebase robin device_subclassing ? | 12:00.34 |
Robin_Watts | kens: Yes. | 12:03.04 |
kens | Right, thanks I'll go do that | 12:03.12 |
| Hmm, seems I can't 'No supported authenticatoin methods available' | 12:04.57 |
Robin_Watts | git remote -v | 12:05.18 |
kens | D'oh stupid me | 12:05.19 |
| I had kens@ instead of ken@ | 12:06.35 |
| Works now | 12:06.45 |
Robin_Watts | ha. sounds like me :) | 12:06.53 |
| b80905`: I note an absence of any malloc checks in that patch. | 13:48.53 |
| Also, we need to be sure that patch works under windows too. | 13:50.10 |
| mvrhel_laptop: Morning | 13:50.15 |
mvrhel_laptop | Good morning | 13:51.05 |
kens | The device_subclassing branch is now on 'origin' (ghostpdl.git), preparatory to Marcos testing t then squashign and merging. | 13:52.33 |
anil_ | Hi robin , Heads Up | 14:28.59 |
| now it is time for commercial license | 14:29.13 |
| Could you tell us your packages.. | 14:29.34 |
Robin_Watts | anil_: I can't tell you, no. | 14:30.58 |
| Firstly, I am an engineer, not a sales guy. I stay well away from the money side of things. | 14:31.17 |
| Secondly, we tailor our licences to the exact needs of the customers. | 14:31.39 |
| No customer wants exactly the same thing, so no two licenses are identical. | 14:31.50 |
| You should speak to Scott Sackett on sales at artifex.com. | 14:32.09 |
| He'll send you a big questionnaire back about what products you want, and how you want to use them. | 14:32.37 |
| Lots of the questions won't apply to you, so just answer them as best you can. The more details you give, the better he can tailor the package to your needs. | 14:33.09 |
| Then he'll get back to you with a licensing proposal. | 14:33.21 |
| anil_: Make sense? | 14:33.37 |
anil_ | yes | 14:36.18 |
| why don't you people keep packages on internet ? | 14:36.42 |
Robin_Watts | anil_: As I said, no 2 customers want the same package :) | 14:41.58 |
anil_ | we need all features what Mupdf is supporting | 14:44.11 |
henrys | anil_: what are you selling? Just curious. | 14:44.49 |
anil_ | henrys_ : are you from mupdf... | 14:45.38 |
Robin_Watts | anil_: It's not a question of which features you need. It's a question of how the whole package works together. | 14:45.45 |
henrys | anil_: not to detract from what Robin_Watts said about contacting scott, we really can't help here. | 14:45.46 |
anil_ | ok Thnaks guys | 14:46.09 |
Robin_Watts | henrys is the head of engineering here. | 14:46.10 |
henrys | pointy haired guy | 14:46.40 |
| I guess chris is out --follow works okay for me, I used to do 'git log .' in the pcl or pxl directory to see only pcl changes but I can live without that. | 14:48.11 |
b80905` | Robin_Watts: You mean in the get_path_to_marks() function? | 15:10.54 |
Robin_Watts | b80905`: yeah. | 15:11.10 |
| mvrhel, henrys: just a gentle reminder that the gprf stuff on robin/master (mupdf) needs to be looked at at some point. | 15:22.02 |
henrys | Robin_Watts: I'll get to it today sorry to keep you waiting. | 15:22.28 |
Robin_Watts | no worries. I'm not blocked on it or anything. | 15:22.41 |
henrys | now that I"ve spent 20 minutes figuring out why on earth the code doesn't break on gx_default_get_param when I intended to set the break on gx_default_get_params I can continue ... ugh | 15:27.39 |
| ghostpuzzle | 15:28.58 |
b80905` | Robin_Watts: I can't seem to find the windows code.. platform/win32 contains only Visual C++ project files and a .bat file | 15:36.58 |
rayjj | b80905`: look at the project file and it will tell you where the various components (source, etc.) exist | 15:44.49 |
| b80905`: the top level sources are in platform/x11 such as win_main.c, win_res.c and pdfapp.c | 15:51.10 |
| s/win_res.c/win_res.rc/ | 15:51.24 |
| the rest is built as sub-projects using the other vcproj files in platform/win32 | 15:52.36 |
b80905` | Why is the directory named x11 if the files in it are not only for X11? | 15:56.43 |
henrys | rayjj: do you know the history to MarginsHWResolution? Why would we need a different resolution for the margins? | 16:00.21 |
rayjj | henrys: sorry, I don't recall. I suspect that it's so that margins can be set precisely to a device pixel | 16:01.21 |
Robin_Watts | b80905`: Sorry, missed your comments. | 16:02.37 |
| Historical reasons. | 16:02.50 |
rayjj | henrys: the only clue is that it defaults to x_dpi, y_dpi (device resolution) but may for devices that want to operate at, say, 600 still want to set margins at a native engine 1200 ??? | 16:03.33 |
| Robin_Watts: ancient history | 16:03.46 |
Robin_Watts | rayjj: I was talking about mupdf, not gs, but yes :) | 16:04.17 |
rayjj | mupdf doesn't have any ancient history compared to gs :-) | 16:04.44 |
henrys | rayjj: I'm trying to add Margins support to pcl to debug something and I'm tripping over that value being set to a crazy number and the translation isn't coming out right. | 16:05.36 |
| I'll track it down. | 16:05.49 |
rayjj | henrys: well, the gxdevice.h std_device_part2_ sets it to x_dpi, y_dpi so maybe when the device is copied it isn't being handled ? | 16:07.21 |
henrys | rayjj: don't I'm getting 1152 for the so called Margin Resolution values | 16:09.59 |
| there's no docs either maybe we should give this the boot | 16:10.12 |
| rayjj: the problem seems to only happen with X11 | 16:23.54 |
rayjj | henrys: istr that x11 does funky things with resolution based on the 'geometry' properties | 16:26.50 |
| henrys: 1152.0 (float) or 1152 integer ? | 16:30.49 |
| 0 == 0.0 so maybe it's been setting it to an integer all along and we never noticed | 16:31.39 |
henrys | rayjj: it's set in the x11 prototype device through one of the hellish macros, I'm looking now, the data type is float | 16:32.33 |
rayjj | because 0x00000480 as a float would be a very strange value | 16:32.38 |
henrys | christ | 16:32.54 |
| I just want to do this very simple thing ... wow | 16:33.19 |
rayjj | macro hell | 16:33.45 |
chrisl | henrys: if you need the history of directory from before it (or its contents) were moved, you can do: 'git log <SHA> -- <directory>' - so on my branch currently, for example (in the ghostpdl directory): 'git log 32fc208587 -- ./pl' would get you the history of 'pl' before I started moving things around. | 16:45.24 |
henrys | chrisl: thanks | 16:46.04 |
| chrisl: question should I start tinkering with PCL API for language switching now or after you check this beast in? | 16:46.32 |
chrisl | henrys: I thought that was part of the language switching project.... | 16:47.05 |
henrys | chrisl: right but if I start twiddling in Robin_Watts branch or my code we have a big merge right? Whereas if you commit ... | 16:48.00 |
| chrisl: I could also work in your branch ... | 16:48.39 |
Robin_Watts | If we're going to use my stuff, there needs to be a big merge done. | 16:49.50 |
| It may be easiest to just redo my work on chrisl's branch. | 16:50.09 |
chrisl | henrys: TBH, I don't think it really matters, if you want to just do it on master - I haven't changed many source files (relatively) so merging source changes isn't a big deal. | 16:50.12 |
Robin_Watts | Let me just check I've pushed my latest stuff... | 16:50.49 |
henrys | chrisl: I'd lean toward getting your stuff in as fast as we can and everyone working from that to keep the changes to master small and incremental as possible. | 16:51.46 |
| but I guess I can go either way I don't feel strongly about it. | 16:52.16 |
Robin_Watts | henrys: Getting chrisl's stuff in without further delay would seem smart to me too. | 16:52.38 |
chrisl | henrys: as soon Ken's stuff is on master, I'm ready to get my branch in, too | 16:52.44 |
Robin_Watts | OK, I think the language branch is about as new as I had it. | 16:56.55 |
chrisl | I assume the list of API methods is just a starting point...... | 16:59.15 |
| And I think we'll ignore the visual tracer stuff | 17:00.34 |
Robin_Watts | chrisl: It's all just a starting point :) | 17:02.37 |
rayjj | Robin_Watts: I thought git could track history across mv operations | 17:06.37 |
| without having to remember which directory files used to be in | 17:07.13 |
Robin_Watts | rayjj: It can. | 17:07.24 |
| git diff --follow new_file | 17:07.34 |
| or git log or git blame etc. | 17:07.45 |
| What it can't do is cope with whole directories of files in a single operation. | 17:08.01 |
rayjj | so why does chrisl say: do: 'git log <SHA> -- <directory>' | 17:08.02 |
Robin_Watts | So to see what happened to everything in the 'new_pcl' directory (which moved from the 'pcl') directory, you can't do: git log new_pcl and see the history from before the move. | 17:08.45 |
| You have to do "git log new_pcl" to see everything back to the move point, then "git log move_sha~1 -- pcl" to see the history before then. | 17:09.30 |
rayjj | Robin_Watts: but you could say git log new_pl/plmain.c and that works ? | 17:09.40 |
Robin_Watts | Yes. | 17:09.45 |
| For single files, it's fine, AIUI. | 17:09.51 |
chrisl | rayjj: it's because git tracks *files* but doesn't track *directories* | 17:10.15 |
henrys | now that I've generated all the preprocessing output I see 1052 comes from 16 * 72 in gdevx.h (FAKE_RES) ... | 17:11.27 |
chrisl | henrys: I take it you aren't using something gdb based for debugging? | 17:12.05 |
henrys | I think by default we should generate the .i files in gs ;-) | 17:12.06 |
| chrisl: I'm using gdb | 17:12.55 |
chrisl | Well, "macro expand" will do what it says | 17:13.22 |
henrys | chrisl: yup but I want them all at once so I can search | 17:13.42 |
chrisl | Well, I won't trying to get the build system to produce .i files any time soon...... | 17:14.24 |
Robin_Watts | make CC=expand.sh | 17:14.47 |
| where expand.sh does gcc twice, once with -E, and once without ? | 17:15.04 |
henrys | chrisl: it's very easy to do that, I was joking of course XCFLAGS="-save-temp=obj" is all you need | 17:15.15 |
Robin_Watts | ooh, nice trick. | 17:15.31 |
henrys | that will generate .o, .i and .s | 17:15.36 |
Robin_Watts | Right hand click on the use of the macro and "goto definition". | 17:16.10 |
| oh, wait... :) | 17:16.28 |
chrisl | Robin_Watts: the nice (and bad) thing about the gdb way is that it "flattens" it completely, so macros that call macros etc all get expanded | 17:17.03 |
henrys | but I still don't understand MarginsHWResolution and don't know if I ever will.. | 17:17.04 |
chrisl | henrys: assign it to hintak?? | 17:17.36 |
henrys | chrisl: nearly fell off my treadmill ... | 17:17.52 |
chrisl | Just a thought ;-) | 17:18.08 |
| Robin_Watts: (re language_switch API) presumably if we make the lib_ctx directly accessible from the gs_memory_t then we'd avoid having to include gslang.h all through the graphics llibrary - in the graphics library "lang_inst" could be an opaque pointer? | 17:21.14 |
Robin_Watts | henrys: I really hope you don't use the treadmill all day: http://www.sciencedaily.com/releases/2015/06/150616093646.htm | 17:21.23 |
| chrisl: I hope that's what the latest version has? | 17:21.41 |
| If not, then I've lost some changes. | 17:22.04 |
| I changed it after your comments at the staff meeting. | 17:22.15 |
| D'Oh! I looked at the wrong bloody laptop. just a mo. | 17:22.27 |
| OK, I've pushed the latest changes to robin/language | 17:23.40 |
| Sorry about that! | 17:23.43 |
henrys | Robin_Watts: I might be in the extreme exercise camp without the treadmill ... | 17:25.33 |
chrisl | Okay, I like that *much* better! So, with that, we can pull (nearly) all the includes of gslang.h in the graphics lib | 17:25.36 |
Robin_Watts | henrys: yeah. | 17:25.48 |
henrys | Robin_Watts: haven't done any 24 hour stuff but I have lots of friends who do and they are always pushing me to do it. | 17:28.16 |
Robin_Watts | It's a cult, I tell you. | 17:29.04 |
| (Not that "runners" aren't a cult already, of course) | 17:29.38 |
henrys | honesly most of them are either crazy or stoners ... | 17:29.42 |
Robin_Watts | We're not mad. Become one of us, and you'll see... just let your mind go. | 17:30.01 |
chrisl | I feel like an hour of squash training with a pro player on a 36+ degree celsius squash court is probably enough for me...... | 17:30.32 |
Robin_Watts | Doesn't it take a stoner 24 hours just to make it to the corner store and back? :) | 17:30.34 |
henrys | it surprised me how much pot is used but it does relieve pain and inflammation so it sort of makes sense. | 17:31.25 |
chrisl | Does it? Or does it just make you not care? | 17:32.14 |
henrys | 1 month to ironman for me. Time to start monitoring the long term weather forecast and nightly ritual prayer for low temps | 17:32.29 |
Robin_Watts | where are you ironmanning ? | 17:33.03 |
henrys | chrisl: not sure. I can't imagine running 24 hours in the woods high... what could possibly go wrong ? | 17:33.40 |
CHris____ | Hey all | 17:33.42 |
| Im very new to trying to use Ghostscript and Im getting an error | 17:34.01 |
| Could anyone try to explain it please? | 17:34.10 |
chrisl | henrys: I can't imagine doing that not being high! | 17:34.17 |
henrys | http://www.ironman.com/triathlon/events/americas/ironman/boulder.aspx#axzz3ef7WjOra | 17:34.18 |
chrisl | CHris____: can you pastebin the (whole) error | 17:34.39 |
CHris____ | http://pastebin.com/2pUh6K3Z | 17:35.13 |
Robin_Watts | CHris____: OK, so you're using imagemagick, not ghostscript. | 17:35.50 |
CHris____ | I thought I was effectively using both? | 17:36.07 |
henrys | I can live with the heat I'm not sure I can live without the wetsuit, which you can't wear if it's hot. | 17:36.15 |
Robin_Watts | ImageMagick is calling Ghostscript in the background. | 17:36.28 |
| CHris____: It looks like imagemagick is calling gs saying "read this file", and that file doesn't exist. | 17:37.27 |
chrisl | henrys: heat, hmm - I lost nearly two litres of fluid in the hour training this afternoon - I find that's really hard to replenish | 17:37.51 |
CHris____ | Thats this c:\windows\temp file | 17:38.14 |
| ? | 17:38.14 |
Robin_Watts | chrisl: You need to identify the 3 closest pubs and do circuit training :) | 17:38.32 |
| CHris____: Indeed. | 17:38.43 |
CHris____ | Ive checked that the IIS APPPOOL has rights to do whatever in C:\Windows\Temp | 17:39.06 |
henrys | chrisl: salt tablets have been effective for me. | 17:39.07 |
CHris____ | But im seeing no files beginning with Mag at all | 17:39.16 |
Robin_Watts | CHris____: I would expect the file to be created, then gs called, then the file to be deleted. | 17:39.42 |
| So you wouldn't see the files persist there. | 17:39.51 |
CHris____ | Ah ok :/ | 17:40.00 |
chrisl | henrys: I have a full electrolyte supplement which does help, but still...... | 17:40.06 |
Robin_Watts | but it smells like a problem at the imagemagick end to me. | 17:40.15 |
chrisl | CHris____: there may be an imagemagick option to leave interim files in place, for debugging purposes | 17:40.48 |
| CHris____: there also could be some funny permissions going on, so I'd look at whether you can point imagemagick at a different temp directory, and create one in your user folder | 17:43.28 |
CHris____ | Ok I can try that :) Thank you | 17:43.46 |
henrys | I see the light - x11 uses the FAKE_RES number to indicate the resolution wasn't set on the command line so it can go figure out a good resolution, but then it never sets the MarginsHWResolution when it updates the resolution. So this setting margins should work in gs either. | 17:47.22 |
| s/margins should/margins shouldn't/ | 17:47.45 |
CHris____ | Ok making a new directory, setting the Imagemagik environment variable to use it and rebooting fixed it thank you guys :) | 17:53.31 |
| Thank you for taking the time to help me :) | 17:54.51 |
Robin_Watts | CHris____: Fab. | 17:55.23 |
chrisl | CHris____: I'm glad my stab in the dark helped ;-) | 17:55.29 |
CHris____ | You are down in my book at legengary experts :) | 17:56.19 |
| err as legendary experts :) | 17:57.15 |
Robin_Watts | PC unwell. Rebooting. | 18:11.20 |
henrys | is tempted to get rid of all this nonsense and make X11 default to 96 dpi. | 18:13.09 |
Robin_Watts | Is there a call on X to say "What is the real dpi?" | 18:19.44 |
| (if such a call fails, then 96dpi seems a reasonable default) | 18:20.06 |
henrys | sort of see gdevxini.c:317 where FAKE_RES is set | 18:20.51 |
| Robin_Watts: I have a fix for the problem that is less disruptive, probably best not to fool with this house of cards too much. | 18:21.54 |
| lunching... | 18:23.10 |
chrisl | x11 and x11alpha are abject lessons in the abuse of the device interface :-( | 18:33.00 |
b80905 | Robin_Watts: I have created a patch for each platform: http://pastebin.com/TsekWa17 http://pastebin.com/D4zDRJSK . What do you think? | 18:54.54 |
Robin_Watts | b80905: both win_main.c and x1_main.c use pdfapp.c | 18:55.37 |
| so any common code can go there. | 18:55.48 |
| But attach the patch to the bug, and I'll talk to Tor about it when he comes back from holiday. | 18:56.16 |
| I think that's 2 weeks time. | 18:56.22 |
| (sorry, he's away for a while) | 18:56.27 |
b80905 | Robin_Watts: I haven't tested the windows version though... | 18:57.11 |
| Robin_Watts: So I'm not sure it works. Do you guys accept patches for one platform only? | 18:59.17 |
Robin_Watts | we would much prefer a patch that works on both. | 18:59.49 |
| It is possible that your existing patch will work already, if it's just in pdfapp.c | 19:00.12 |
| but I haven't got time to test it at the moment. | 19:00.22 |
b80905 | Robin_Watts: You mean the first patch? | 19:00.46 |
Robin_Watts | yes, | 19:01.04 |
b80905 | Robin_Watts: There's no way it works on windows... | 19:01.10 |
| Robin_Watts: I mean, getenv, realpath, mkstemp.. | 19:03.17 |
Robin_Watts | getenv is on windows. | 19:03.43 |
| I even have HOME set in my install - I assume that's standard. | 19:04.00 |
scorneli | hi, I'm trying to file a private bug into bugzilla, but I don't see a checkbox or field to check in order to have it private | 19:12.49 |
Robin_Watts | scorneli: You attach it, and someone here marks it private. | 19:13.08 |
| I have to step afk now, but someone else will be along in a mo that can do it for you, I'm sure. | 19:13.28 |
| henrys maybe? | 19:13.31 |
scorneli | http://bugs.ghostscript.com/show_bug.cgi?id=696070 once this is marked private I can add details | 19:16.06 |
jogux | scorneli: I think I've done that. Hopefully that doesn't stop you seeing it :-) | 19:28.36 |
Robin_Watts | scorneli: No, you have to add the file, then we mark the file private. | 19:30.43 |
| Although, this way might work if you can still see the bug. | 19:31.02 |
scorneli | it seems like I can | 19:31.18 |
Robin_Watts | And if you log out from bugzilla, you can't? Sounds like you're sorted. | 19:33.29 |
scorneli | ok, I've added details know, I think | 19:34.37 |
jogux | certainly I can't see it when logged out | 19:34.42 |
scorneli | Users in the roles selected below can always view this bug: Reporter was checked. I should be reporter, so I think I'm good. | 19:36.03 |
jogux | scorneli: good stuff, thanks. A ghostscript person will need to look at that now (I am not one of those). | 19:38.36 |
scorneli | jogux: yeah, no worries. thank you your help | 19:39.59 |
rayjj | mvrhel: Robin_Watts: If either of you has any comments on this commit, I'd appreciate it: http://git.ghostscript.com/?p=user/ray/ghostpdl.git;a=commitdiff;h=eec6f4f50d030ab0fc2f4a707a0a7788819fe03c | 21:18.35 |
| mvrhel: Robin_Watts: This wraps up all of the serious issues for fast HT except matching the transition levels when there is a transfer function in place. There is commented out code that is supposed to do that, but I suspect it is wrong, so it will take some work to fix/replace it | 21:21.00 |
| Robin_Watts: this does *not* change the transition levels as we've discussed, but makes the 'fast HT' code match the non-fast method | 21:22.08 |
| Forward 1 day (to 2015/07/02)>>> | |