0:00:09 | right me |
---|
0:00:11 | okay |
---|
0:00:12 | right i'm alex also and |
---|
0:00:15 | and i |
---|
0:00:19 | i am and so we can might recognise this is from glottic two thousand virus |
---|
0:00:25 | and then in that i know for a long time i initially got there because |
---|
0:00:31 | i'm right diameters and i ran an sauce |
---|
0:00:35 | since then i moved on since two thousand one word for red hat just a |
---|
0:00:39 | group |
---|
0:00:40 | you mean random stuff all over the place |
---|
0:00:44 | recently i've been working on the overall high dpi stuff for |
---|
0:00:50 | i know desktop and the linux desktop in general |
---|
0:00:54 | i want to start with a |
---|
0:00:57 | very specific |
---|
0:00:58 | problem definition this is a from book pixel |
---|
0:01:04 | typically runs from list but it's very easy to kind of easy to install any |
---|
0:01:09 | linux operating system on it and this is |
---|
0:01:13 | muscles and or nineteen which is |
---|
0:01:17 | it supported every kind of harvard ask a sickly |
---|
0:01:22 | except |
---|
0:01:24 | why when you read it looks like this |
---|
0:01:26 | i don't it doesn't actually look like this it's much high resolution but |
---|
0:01:30 | scaled down to fit on the display |
---|
0:01:33 | right so |
---|
0:01:35 | it's not really useful even though it's taking we will support it |
---|
0:01:39 | is completely useless |
---|
0:01:42 | because you can even in stuff |
---|
0:01:45 | normal it's so you know about the do you stuff |
---|
0:01:48 | and you can use that operate in sure |
---|
0:01:51 | for wanna taurus generally how many pixels per |
---|
0:01:54 | in some you wanna start |
---|
0:01:57 | and |
---|
0:01:58 | historically all like |
---|
0:02:00 | normal monitors i mean be a projected doesn't have it but any random normal wanna |
---|
0:02:05 | so we have around hundred |
---|
0:02:07 | nineteen eighty or a hundred twenty or something but not |
---|
0:02:11 | but the pixel two hundred thirty nine |
---|
0:02:14 | right so |
---|
0:02:16 | doesn't really work |
---|
0:02:18 | so we had a setting in X next we actually |
---|
0:02:22 | which for on various reasons is always set to ninety six by know |
---|
0:02:26 | but there is a there's a property |
---|
0:02:29 | take scale factor i think is that so this had to we get that |
---|
0:02:34 | and it's now |
---|
0:02:37 | juice able and some so you can actually read stuff |
---|
0:02:41 | but |
---|
0:02:43 | it really only change the text side |
---|
0:02:46 | or someone's mother is just to so |
---|
0:02:49 | one is that why did like scale and it right |
---|
0:02:55 | it turns out that |
---|
0:02:57 | during that we went traditional density monitors would be a bad thing |
---|
0:03:04 | and this monitors be maybe not a pixel but on the other ones |
---|
0:03:08 | a pixel is like invisible thing |
---|
0:03:10 | you better not enjoying one point five pixels because that looks fancy we're |
---|
0:03:16 | and so everything you mentor has actually matched the target |
---|
0:03:20 | pixel great you even if you're sources an S P G or whatever invented wrong |
---|
0:03:25 | one point zero pixels or things will |
---|
0:03:28 | really bad |
---|
0:03:29 | also lot of content that you see in a typical you why |
---|
0:03:32 | you have the i can't swich are changing use or can choose a bunch a |
---|
0:03:37 | pen about our pictures whatever |
---|
0:03:39 | cursors are also pixels |
---|
0:03:43 | and scaling those for just doesn't look very well |
---|
0:03:47 | there's an example of this is |
---|
0:03:49 | i mean this is a scale version and then all those |
---|
0:03:53 | and put it down on which is scale so they can see all the pixels |
---|
0:03:57 | and if you were to render this is a hundred twenty percent |
---|
0:04:01 | say to compensate for your like to guys change you get like |
---|
0:04:06 | kind of hard to see that |
---|
0:04:08 | some of the lines would get in the middle of some pixels evening been fuzzy |
---|
0:04:12 | summer sharper and the probably once or less |
---|
0:04:16 | re to get this |
---|
0:04:19 | and the pixels are scale on they only i can so |
---|
0:04:23 | around |
---|
0:04:25 | those are ceases images and there's the ganges at the core so their scale |
---|
0:04:31 | also |
---|
0:04:33 | turns out that it's not really a problem right |
---|
0:04:36 | for instance this |
---|
0:04:37 | thank but i have |
---|
0:04:40 | are you still a twenty three inch monitor at work |
---|
0:04:44 | and i'm not sure a button on it |
---|
0:04:47 | and it's like six millimetres on the laptop in eight on the |
---|
0:04:51 | so this way and it's not really a problem |
---|
0:04:54 | and i could |
---|
0:04:56 | happy to the size and everything to match the deep i spent will be stupid |
---|
0:05:01 | because you wouldn't be anything but you get all this bad looking stuff |
---|
0:05:06 | the |
---|
0:05:08 | we were getting it from the time they were solution the pixel have not care |
---|
0:05:14 | the routine and just lay stuff and something it doesn't work |
---|
0:05:19 | so we had to do something |
---|
0:05:22 | sell |
---|
0:05:24 | it's not looking at some of the things that wrong |
---|
0:05:27 | i don't obviously a pixel they so they're not scale and also the way too |
---|
0:05:31 | small can see you know they look kind of weird if you look at the |
---|
0:05:35 | i can there next to the text |
---|
0:05:38 | much more than the textbooks |
---|
0:05:40 | out of sync |
---|
0:05:43 | something on the will gents are |
---|
0:05:46 | specified in terms and |
---|
0:05:48 | and you know they're fixed pixel so there will be hard to pay |
---|
0:05:53 | trying to |
---|
0:05:54 | had a scrollbar for instance is really hard |
---|
0:06:00 | and i don't know it's okay here but if you know how they won't talk |
---|
0:06:04 | as opposed to look supposed to have a really large image background with the tech |
---|
0:06:08 | sensors on it and it just doesn't care |
---|
0:06:13 | and |
---|
0:06:14 | the full window sizes a |
---|
0:06:17 | i mean some of this so the windows look okay is there are fully specified |
---|
0:06:21 | by the size of all we distant anything that has anything that scrolls |
---|
0:06:25 | but you have it or not less |
---|
0:06:28 | we don't really know the default size to |
---|
0:06:31 | the waste typical we fixed a biased having to order of the at specify random |
---|
0:06:36 | the full |
---|
0:06:37 | size in pixels |
---|
0:06:40 | typically a not a problem but and get everything wrong for |
---|
0:06:44 | the pixel |
---|
0:06:46 | also some of the intro widgets are specified in pixels |
---|
0:06:49 | okay with of the sidebars |
---|
0:06:52 | the weight of the columns are notes |
---|
0:06:55 | so small that |
---|
0:06:57 | doesn't say it's a common labels adjust not doesn't make sense |
---|
0:07:02 | and all on around of you can see it in minutes |
---|
0:07:05 | really tiny |
---|
0:07:06 | and also go slowly because it's expiration and everything is based on the typical size |
---|
0:07:12 | is that |
---|
0:07:13 | excellent have a little monitors |
---|
0:07:16 | resizing windows really hard |
---|
0:07:19 | all the padding is really thin so it's |
---|
0:07:22 | doesn't look right |
---|
0:07:25 | and basically |
---|
0:07:27 | pixels or absolute pixels what is everywhere |
---|
0:07:30 | either directly or indirectly via pixel based icons and it is |
---|
0:07:36 | and also there's some points of code all the place that uses this |
---|
0:07:40 | but we don't wanna have everyone change older cut |
---|
0:07:46 | additionally if you |
---|
0:07:48 | and we use multiple monitors things that you know more complicated |
---|
0:07:52 | because |
---|
0:07:53 | i don't think there's any how you've got external monitors |
---|
0:07:57 | well |
---|
0:07:58 | what negatives something super expensive want been generally if you plug in |
---|
0:08:03 | extra montana and pixel you get a little you can i and how do you |
---|
0:08:08 | have monitored same time |
---|
0:08:10 | and windows the rolling back from whence going to the other |
---|
0:08:14 | well i would become hero which or tiny depending on |
---|
0:08:17 | what we invite them and if you have a minimum both look really weird |
---|
0:08:25 | so |
---|
0:08:27 | so the solution i came up and then it's not really |
---|
0:08:31 | original in the sense that number didn't before it's pretty much one that covers dustin |
---|
0:08:37 | i'm not exactly sure what |
---|
0:08:39 | they do and when there's a has something kind of similar |
---|
0:08:43 | but basically define redefine one pixels me on an old one utterance pixels new money |
---|
0:08:49 | used to do but on kind of i mean something else |
---|
0:08:53 | that at you one pixel napster multiple |
---|
0:08:57 | device pixels |
---|
0:08:59 | and then |
---|
0:09:02 | and the scaling factor that would make it an integer to vital the |
---|
0:09:06 | green alignment issues and also for technical reasons |
---|
0:09:10 | lot of stuff like X windows |
---|
0:09:13 | but when the are digit widgets can't add |
---|
0:09:17 | don't a list of sizes and clipping goes to hell if you have |
---|
0:09:23 | fractional |
---|
0:09:25 | a box of that's it |
---|
0:09:27 | and all the scaling is done automatically so the that absolutely know about it |
---|
0:09:33 | thank you render high resolution or if you draw |
---|
0:09:38 | but not based in just to get scale automatically |
---|
0:09:42 | and |
---|
0:09:42 | and then and change you actually have some |
---|
0:09:45 | X a laid out like icons or ceases images for backgrounds and borders or even |
---|
0:09:52 | like about source or external images |
---|
0:09:56 | we had a new A P I S or ways to let you specify multiple |
---|
0:10:00 | sources and then just so that you can pick the right one depending on the |
---|
0:10:04 | on the on the window |
---|
0:10:05 | okay |
---|
0:10:08 | so when that and it looks like this |
---|
0:10:13 | i mean obviously at this is like |
---|
0:10:22 | i mean obviously it just looks like any other know but if you actually lot |
---|
0:10:27 | of pixel and the to be high resolution icons and |
---|
0:10:33 | all the things where the right sizes and everything i mean |
---|
0:10:37 | it's still a tiny per se so there's some stuff like to fix but most |
---|
0:10:40 | of it works |
---|
0:10:43 | so |
---|
0:10:45 | i guess there's a lot of all percent sorry we go over all you have |
---|
0:10:49 | to change your apps mostly they shouldn't have to change that |
---|
0:10:53 | only see the ice change a bit |
---|
0:10:56 | we stopped i roll |
---|
0:10:58 | generally hire a support scaling and you can just set scale |
---|
0:11:03 | and we can just set the scale |
---|
0:11:06 | but |
---|
0:11:10 | i have |
---|
0:11:11 | and it something on the my scale like this at the surface |
---|
0:11:14 | so that any higher context you create on that so we will automatically always get |
---|
0:11:19 | initialised diskette |
---|
0:11:21 | you cannot get rid of the scale accident by like resetting that matrix and whatever |
---|
0:11:27 | and additionally and maybe even more important if you if you draw the surface of |
---|
0:11:33 | the something else |
---|
0:11:34 | scaling is also applied |
---|
0:11:36 | so you have and |
---|
0:11:38 | i can as |
---|
0:11:41 | target of forty eight by forty eight abstract pick something that you have ninety six |
---|
0:11:47 | by nine six |
---|
0:11:48 | image then you set unless got that you will enjoy a non scale thing it |
---|
0:11:54 | will automatically scale down |
---|
0:11:57 | although generally all this you know don't have to care that a does that for |
---|
0:12:01 | you |
---|
0:12:03 | you can however that with changes |
---|
0:12:06 | straight window abstraction for forty K so it has all the window screen monitor need |
---|
0:12:13 | to targets |
---|
0:12:15 | those not always almost report cites in this abstract pixel space so |
---|
0:12:20 | when you thing is they were G D K window might not match the size |
---|
0:12:24 | of the corresponding X one now |
---|
0:12:27 | so the expand maybe turin pixels and detailed reports |
---|
0:12:30 | three hundred pixels |
---|
0:12:33 | and it so to compensate if you really need to do something a little you |
---|
0:12:36 | have to get scale factor for the when the |
---|
0:12:39 | or not or if you're doing something |
---|
0:12:41 | global |
---|
0:12:43 | and also the multiple sessions |
---|
0:12:45 | or like and then it |
---|
0:12:48 | reported in the abstract texas based that |
---|
0:12:51 | since time immemorial we use floats for the event so |
---|
0:12:56 | still have superstition |
---|
0:12:58 | just in a different for |
---|
0:13:01 | i had to add a couple of at articles to |
---|
0:13:05 | get the current |
---|
0:13:06 | cursor positioning floats of stuff but generally |
---|
0:13:10 | it always been using floats for women's |
---|
0:13:15 | did you integrate similar services is what you use when you create all screen images |
---|
0:13:19 | in U K three |
---|
0:13:22 | it's always been there is change now so that automatically creates a scale surface if |
---|
0:13:28 | you if you are targeting scale window |
---|
0:13:31 | you should always use this role directory in your in services because |
---|
0:13:34 | if you create a little for solution low scale double buffer for a high scale |
---|
0:13:39 | when though you get excellent results |
---|
0:13:42 | and you have to be careful that the scale when that for instance widget whatever |
---|
0:13:48 | you're working on can change over time so |
---|
0:13:52 | make sure old problem of course or |
---|
0:13:55 | we created when you |
---|
0:13:57 | in the scale changes |
---|
0:14:00 | and you have this is they wanted to decide |
---|
0:14:04 | or something about it but |
---|
0:14:06 | off you don't have to do you |
---|
0:14:08 | you just |
---|
0:14:09 | look at the scale it or something |
---|
0:14:12 | and it is different from what have you rendered it is invalidate |
---|
0:14:16 | next or |
---|
0:14:17 | but there is a thing |
---|
0:14:20 | there there's some you think very similar to prince in the surface but it creates |
---|
0:14:24 | and just |
---|
0:14:26 | in the surface which is |
---|
0:14:29 | it's a long time or service but it's you can access the pixels |
---|
0:14:33 | market expert |
---|
0:14:37 | the advantage here is that you can now specify the scale of it |
---|
0:14:40 | and then |
---|
0:14:42 | and it also lets you specify window for the target so |
---|
0:14:47 | if you're creating an |
---|
0:14:49 | in the service that you eventually will draw a next one now we can handle |
---|
0:14:53 | time okay and i'm really generally space |
---|
0:14:57 | so you always get the extra |
---|
0:15:00 | exchange |
---|
0:15:04 | also there's this |
---|
0:15:07 | have a circus create for exposed help or which is really |
---|
0:15:10 | simple wrapper of quite similar in the surface that just converts |
---|
0:15:14 | from picks before that too |
---|
0:15:17 | performance which incidentally are not exactly the same so right now |
---|
0:15:23 | we are actually allocating a new surface and converting from the |
---|
0:15:28 | to promote apply and also mode |
---|
0:15:31 | every time you draw textbook so you should use sort of this |
---|
0:15:37 | i don't know |
---|
0:15:39 | okay this is what most people actually see it technicians |
---|
0:15:43 | well |
---|
0:15:45 | the size allocation everything that related size |
---|
0:15:48 | or and this |
---|
0:15:50 | abstract well that's the pixels |
---|
0:15:52 | and weighted has a scale proper day you get the scale what you can get |
---|
0:15:56 | scale |
---|
0:15:57 | but there's also it's also proper basic and get notified on it if you need |
---|
0:16:02 | to |
---|
0:16:03 | and i wanted to catch whatever |
---|
0:16:06 | also |
---|
0:16:08 | for the reasons |
---|
0:16:10 | and this about the for hire circles that it is then you prefer way too |
---|
0:16:16 | i in that state as |
---|
0:16:19 | don't wanna use X path |
---|
0:16:22 | i mean you can just pick spot in your internal stuff attendance |
---|
0:16:25 | supposed to be drawn the screen you probably use our services that |
---|
0:16:28 | because they can have scale |
---|
0:16:30 | so when you don't and you will automatically upscale dollars gonna whatever and you can |
---|
0:16:37 | pre target and to the right |
---|
0:16:40 | type of window any and they're already in the right kind of formants something more |
---|
0:16:43 | efficient around |
---|
0:16:45 | so these are new things you should use |
---|
0:16:49 | okay and manage and just get some hundred fixed well |
---|
0:16:53 | i mean that identical except the you say from X possibly on the explode okay |
---|
0:17:02 | so well still work but you should prefer our services |
---|
0:17:07 | i can things extend as so that |
---|
0:17:10 | if you asked for a fully pixel icon and then we did a scale to |
---|
0:17:15 | we will actually ask for ninety six pixel one use that and furthermore there's also |
---|
0:17:20 | extension to the actual i can be respectable well i can be in itself to |
---|
0:17:25 | define hire is versions of |
---|
0:17:29 | a likens because in it's not always the case and it's a good idea to |
---|
0:17:33 | use to ninety six point nine six i can |
---|
0:17:35 | for the forty eight it's a case |
---|
0:17:38 | because things like that our clients or ten or is more details if there's any |
---|
0:17:43 | text in it |
---|
0:17:45 | a and it's gonna be too small if you're better than like this excellent forty |
---|
0:17:49 | eight size |
---|
0:17:52 | there's some extensions it in the evening system to like we specify |
---|
0:17:56 | alternative images |
---|
0:17:58 | if you use this make sure they have the same |
---|
0:18:02 | size some scaling factor because layer layout is affected by the size |
---|
0:18:08 | and we're only going to load the one that |
---|
0:18:11 | is used |
---|
0:18:14 | so i don't |
---|
0:18:16 | something to the current state is layers that was on the stack |
---|
0:18:20 | i had a bunch of protocol additions for will and landed in one point two |
---|
0:18:27 | and in one but you only |
---|
0:18:30 | it's that's out and service we really means monitoring window |
---|
0:18:35 | that's the we'll in terms |
---|
0:18:37 | well the positions and sizes are in this abstract pixels |
---|
0:18:41 | and then the compulsory can shows the actual scale for every monitor |
---|
0:18:45 | and you can be in it as a property on a little bit but well |
---|
0:18:48 | the scale is so |
---|
0:18:50 | if you push a |
---|
0:18:53 | okay long window on the on a scale to output it will automatically of scale |
---|
0:18:58 | it all drawing it |
---|
0:18:59 | but declined can see this and pretty allocate a slice risque to buffer voiced more |
---|
0:19:06 | pixels |
---|
0:19:07 | so then the compulsory doesn't have to |
---|
0:19:09 | right |
---|
0:19:11 | and it also what's correct handling a mixed mode monitor so one drawing a single |
---|
0:19:17 | buffer it can upscale online and not scale you know |
---|
0:19:22 | even scale it difference |
---|
0:19:24 | slices |
---|
0:19:27 | it came the backend waylon will automatically take the maximum scale of all the windows |
---|
0:19:33 | that are cover or all the models are there are covered by the windows so |
---|
0:19:38 | if you drag your window we automatically get resized where we can you scale which |
---|
0:19:43 | is one of the reasons you have to take care about a double offers |
---|
0:19:48 | can change by using that you know in the |
---|
0:19:52 | implementation size really in simple |
---|
0:19:58 | waylon and indicate currently is uses inner surfaces |
---|
0:20:03 | so we used at the device going that and this works and this is |
---|
0:20:08 | huge and this is what we wanna go it has every support for the parts |
---|
0:20:12 | a little bit is it does correctly support the next guy wanna for |
---|
0:20:16 | and |
---|
0:20:19 | and the scaling stuff this is really |
---|
0:20:22 | part of the native system now we can't be on X |
---|
0:20:28 | a bunch of us don't use expert |
---|
0:20:31 | so we have sometimes for that you |
---|
0:20:34 | however we only support a single scale for all |
---|
0:20:38 | roll the screen actually wanted that you can technically open the books place and different |
---|
0:20:43 | scales i guess but practise everything's gonna be the same scale |
---|
0:20:48 | works fine for laptop maybe not if you have lots great thing |
---|
0:20:53 | indicate like fakes all the states i systems of |
---|
0:20:58 | so it scale is enable little i think the size of the windows and wanna |
---|
0:21:02 | sorenson screens one not the coordinates of the excellence are divided by the scale we |
---|
0:21:08 | were contradictory events |
---|
0:21:11 | there is this environment variable unique a scale you can set |
---|
0:21:14 | so really |
---|
0:21:16 | very useful for the pointing if you're actually don't have to hide if you know |
---|
0:21:19 | why she decided and running around and see a and you can see if it |
---|
0:21:25 | works |
---|
0:21:27 | the weights actually meant to work is that we had |
---|
0:21:30 | to next settings that again the settings tingle |
---|
0:21:34 | set for the chain will |
---|
0:21:37 | hopefully eventually i have some really bad patches but |
---|
0:21:41 | and the one on the tech the actually on the screen and just pick the |
---|
0:21:45 | right scale |
---|
0:21:47 | and it telex |
---|
0:21:49 | or is tell detaining just use |
---|
0:21:53 | there's also new settings |
---|
0:21:55 | related to scale with just on scaly got because we still wanna is |
---|
0:22:00 | this don't wanna step and in you know actually lords to support the school apps |
---|
0:22:05 | like to take a two or whatever |
---|
0:22:08 | nobody uses that anymore |
---|
0:22:11 | and it but |
---|
0:22:12 | if we set a really high tech you know i |
---|
0:22:16 | false and the scaling and then digits we have the dark to again did fallen |
---|
0:22:20 | so we have this new on scale late you guys heading that scale where you |
---|
0:22:25 | use |
---|
0:22:27 | also very simple |
---|
0:22:30 | implementation we use used normal tire likes stuff that the wide scale |
---|
0:22:37 | you have to be a bit more careful since the scaling stuff is kind of |
---|
0:22:40 | done extra lead to X |
---|
0:22:42 | X itself has no idea of the scaling so if you're mixing X stuff and |
---|
0:22:46 | it a stuff like |
---|
0:22:48 | using X video or opengl or something you have to |
---|
0:22:52 | what correctly get the scale factor a lot why sizes and stuff most of the |
---|
0:22:57 | time if you create stuff from G ditch a in just consume it and read |
---|
0:23:03 | back the sizes from X |
---|
0:23:05 | things should work but |
---|
0:23:08 | you may need to take care at least |
---|
0:23:11 | we also have a about sex implementation which is |
---|
0:23:15 | well this trivial would just uses all the ports |
---|
0:23:18 | stuff just basically identical to our stuff |
---|
0:23:21 | we do is read back the scale factor from of X |
---|
0:23:26 | right now there's some bargain on your over |
---|
0:23:28 | or create similar those and it's |
---|
0:23:31 | create a scale it sort of |
---|
0:23:33 | scale words to mention whatever it's cool so all the ski all the scrolling and |
---|
0:23:38 | everything that's double buffer |
---|
0:23:40 | currently look kind of excellent a routine this place |
---|
0:23:48 | i think said when there's eight has some kind of high depressed or i haven't |
---|
0:23:52 | really looked at it maybe we should |
---|
0:23:55 | maybe some want to do that |
---|
0:23:58 | need course are scaling with you have |
---|
0:24:01 | we need tiny cursor it's hard to see |
---|
0:24:04 | and into actually finalised i can make changes |
---|
0:24:08 | didn't get any comments on it something which is can it |
---|
0:24:12 | when i need the i versions of something like this may be able maybe is |
---|
0:24:18 | some of them i haven't extra too much detail we have to look at that |
---|
0:24:23 | details are shipped a bunch of something like those we want probably want iris versions |
---|
0:24:28 | of those |
---|
0:24:29 | and we want full integration of |
---|
0:24:32 | you know detection of dpi and everything you know |
---|
0:24:37 | i have some |
---|
0:24:38 | how to hack patches that |
---|
0:24:41 | well |
---|
0:24:44 | mostly work |
---|
0:24:45 | and it somewhere |
---|
0:24:48 | also on the future i might or might not depends on how women go so |
---|
0:24:53 | how much for content and maybe |
---|
0:24:56 | when it maybe it's possible to do we want to monitor support ready X |
---|
0:25:00 | at least you have a compulsive |
---|
0:25:03 | most or because then |
---|
0:25:05 | compulsory control which window different event |
---|
0:25:09 | well the working |
---|
0:25:11 | no it's worth it depends on a classroom goes away when |
---|
0:25:16 | really |
---|
0:25:18 | and also when we do you shot out to run remember we use randomly turned |
---|
0:25:23 | up on the list as i have this pixel doesn't wanna so they can do |
---|
0:25:28 | i did work |
---|
0:25:30 | and |
---|
0:25:32 | thank you |
---|
0:25:44 | it sounds like i think you said that we win you get to separate circuses |
---|
0:25:49 | at different scales and drive them independently i get some of these i mean it |
---|
0:25:54 | that way multimode to works and X and where were somewhere one is completely different |
---|
0:26:00 | and an X you L okay basically was should service that covers everything and they |
---|
0:26:05 | stand out different parts of it so it's really hard to do we know in |
---|
0:26:09 | a different sizes and of model to |
---|
0:26:13 | doesn't work but in whale and |
---|
0:26:15 | every output scans from it some buffer and it's up to the composite usual window |
---|
0:26:20 | in both the buffers but i mean you centre and you scale it how everyone's |
---|
0:26:24 | that if you're an application |
---|
0:26:26 | with your own application you get you get just select |
---|
0:26:30 | the scale of your window |
---|
0:26:33 | but could you possibly say i want you to scale version and the ones known |
---|
0:26:39 | that has this discussion about that but it's not have i don't i don't think |
---|
0:26:42 | it's really important because |
---|
0:26:45 | first was problematic because things like way out in terms of |
---|
0:26:50 | text is different it's like independent different sizes |
---|
0:26:55 | what you to hinting else also for related retina coming and also |
---|
0:27:02 | windows spanning multiple once or is that a colour to is |
---|
0:27:07 | didn't wanna have extra complexity of all over the place or something |
---|
0:27:12 | so i guess you just pick scale |
---|
0:27:16 | i don't want to actually happen |
---|
0:27:20 | but on our sex if you if you do it as you plug in something |
---|
0:27:24 | well gonna next i wanna ask you whether you want the best |
---|
0:27:30 | what can best on the extra always where looking best only internal and |
---|
0:27:35 | just fix that |
---|
0:27:38 | right now i'm always pick |
---|
0:27:41 | the want to not have to |
---|
0:27:42 | one of |
---|
0:27:44 | the want to almost or not but that the when the covers i pick the |
---|
0:27:48 | wanna have to heights |
---|
0:27:50 | but it's really a just |
---|
0:27:53 | attention to detail and indicate can do |
---|
0:27:57 | it's up to the at we need to select |
---|
0:28:00 | or maybe we wanna support it just seems dpi that you haven't you is really |
---|
0:28:05 | fourteen this idea |
---|
0:28:07 | we have a year |
---|
0:28:08 | one |
---|
0:28:09 | scale ribbon |
---|
0:28:11 | and that it we in this people would be we have in the future we |
---|
0:28:14 | will be able you i gonna be |
---|
0:28:17 | that's also some complexities in terms of the male being different and |
---|
0:28:22 | and the way like |
---|
0:28:24 | offscreen just for passion scrolly bottles now almost complexity for me |
---|
0:28:31 | not very important use case |
---|
0:28:33 | you can do it |
---|
0:28:38 | i think i think it i think it's |
---|
0:28:41 | hard enough that it so |
---|
0:28:46 | okay |
---|
0:28:47 | the question |
---|
0:28:51 | you can sort of come up and play with the actual pixels later but |
---|
0:28:56 | you do each buildings differently or use like demo |
---|
0:29:00 | don't know what you were you were doing like a filled earlier it's like i |
---|
0:29:04 | have a |
---|
0:29:06 | the |
---|
0:29:07 | randomly but |
---|
0:29:10 | right or something |
---|
0:29:12 | i think about a batteries explore the |
---|
0:29:16 | system petition or something so the chrome stuff but that it wouldn't boot so how |
---|
0:29:21 | to be everything that i do have some it failed |
---|
0:29:24 | and will later it's |
---|
0:29:27 | minute |
---|
0:29:29 | it looks like you know |
---|
0:29:31 | stuff like you have to be really close up to look at that and we |
---|
0:29:34 | think about later also |
---|
0:29:39 | question here and we will the composite or |
---|
0:29:43 | draw itself and multiple |
---|
0:29:45 | resolutions |
---|
0:29:49 | because applications don't |
---|
0:29:51 | don't have in the different monitors a resolution if you mean in composite turns that |
---|
0:29:57 | we having that's right all over the place right i do union X server |
---|
0:30:01 | both |
---|
0:30:02 | but i mean the way works and presses waylon is that |
---|
0:30:06 | you just give your buffer with that's to |
---|
0:30:10 | composite or and you playing that this is that scale one |
---|
0:30:14 | but what about the shell it's what's so we show was the composer |
---|
0:30:19 | so it's a so if the shell as the way one composite are saying |
---|
0:30:23 | and some apps |
---|
0:30:24 | at hands that a |
---|
0:30:28 | a hundred one hundred buffering plane this is that scale one |
---|
0:30:31 | and your actual window or your actual models or is it scale to |
---|
0:30:36 | it's up to the composite or the scale it by two |
---|
0:30:39 | what and i can come are small monitors can need to do it what i |
---|
0:30:44 | mean do you why of all of the composite or like them shot of barn |
---|
0:30:47 | no i know mostly uses it a for stuff so |
---|
0:30:53 | that's not the rest of what is a lot of a lot was also you |
---|
0:30:56 | see and it are is do you know it's easy to set scale factor you |
---|
0:31:00 | and i think is currently down that it shouldn't be that |
---|
0:31:06 | money always fixing it |
---|
0:31:09 | and but in and basically it's a lot adding it to somewhere in a matrix |
---|
0:31:13 | that get and it's not really |
---|
0:31:19 | and then there are some parts harder maybe like |
---|
0:31:23 | getting |
---|
0:31:25 | it is they're sending pixel based in is to have multiple |
---|
0:31:29 | multiple sources of pixel a |
---|
0:31:32 | i think for |
---|
0:31:34 | for the shallow so that i us are either symbolic i have the terrace at |
---|
0:31:38 | or |
---|
0:31:39 | or just a file which are these space really arts scale are still look good |
---|
0:31:47 | okay so |
---|
0:31:51 | mike |
---|
0:31:53 | my question be about the you have a motion |
---|
0:31:59 | events |
---|
0:32:00 | you use you were saying that you do you subpixel |
---|
0:32:05 | subpixel cool didn't we some hard ways actually a problem because they keep sending |
---|
0:32:11 | they keep sending events with more new changes like as it subpixel motions and we |
---|
0:32:19 | end up breaking pretty much the whole stack |
---|
0:32:22 | well we can we will run into |
---|
0:32:26 | we already do this |
---|
0:32:28 | well it everywhere |
---|
0:32:30 | and he indeed is actually causing problems and on wondering whether that's gonna have problems |
---|
0:32:37 | okay i'm generally you never gonna hit the scale other than one or two so |
---|
0:32:42 | it's |
---|
0:32:43 | i'm gonna have closer |
---|
0:32:45 | zero point five or one |
---|
0:32:48 | no way how that was great |
---|
0:32:52 | that know that those adjudicated motion compression full things well subpixel |
---|
0:33:02 | i mean it was the right way of floating point because |
---|
0:33:05 | an is to fix it for waco talents |
---|
0:33:08 | nineteen ninety seven or whatever so |
---|
0:33:11 | it's in his work |
---|
0:33:14 | okay very no thank you |
---|
0:33:26 | so we're now if you're very high dpi screen |
---|
0:33:31 | and you start an application a |
---|
0:33:34 | would you could be |
---|
0:33:36 | does it start a with its windows |
---|
0:33:40 | who |
---|
0:33:42 | no scaling or want one or |
---|
0:33:45 | all you need to you know with about a hundred snaps you can see it |
---|
0:33:49 | it's gonna be a do you mean |
---|
0:33:52 | if you if you i mean i previously i'm recommended or i am updating an |
---|
0:33:58 | existing application |
---|
0:34:03 | what happens by the just good but you'll it's it looked as if you were |
---|
0:34:09 | on a little bit i screen and everything is as |
---|
0:34:12 | for |
---|
0:34:14 | what you mean need to looks |
---|
0:34:16 | i mean it no it's actually in terms a holly application server whatever sees everything |
---|
0:34:22 | but it will be scale by default |
---|
0:34:25 | i know it's if you if you install zero twenty whatever where this is an |
---|
0:34:30 | excellent it automatically pick up effect of those are really quite a screen but people |
---|
0:34:35 | everything will be scaled so it's |
---|
0:34:38 | approximately this line size that would be a |
---|
0:34:44 | button and hire or something |
---|
0:34:47 | right precision i mean you know maybe just works |
---|
0:34:52 | there's some cases i am where you can write it if you're for instance |
---|
0:34:59 | full at anyone and you have a bunch of ex boss but for all your |
---|
0:35:04 | project X a wanna render them |
---|
0:35:06 | you need to take care |
---|
0:35:09 | detecting that |
---|
0:35:10 | is actually more pixels |
---|
0:35:12 | i mean like spectrum ask about a slight begin so well you know if something |
---|
0:35:18 | present we need to know if it's down scaling |
---|
0:35:22 | a large textbook and it |
---|
0:35:26 | have the but to me to do you have to get this going factor to |
---|
0:35:30 | know that you need to generate more pixels when you're doing something |
---|
0:35:34 | like one i'm thinking about him and i have michael some widget problem or run |
---|
0:35:38 | team actually you know those you don't okay |
---|
0:35:44 | i mean well we don't know but i would still don't show the wrong pixels |
---|
0:35:48 | compared the device text |
---|
0:35:49 | yes |
---|
0:35:51 | also |
---|
0:35:57 | so it's not a problem because that's what it's doing okay so this is around |
---|
0:36:03 | when you have a problem is gonna scale itself but what about the came victims |
---|
0:36:08 | custom widget |
---|
0:36:10 | you mean the canvas the display of the actual image itself |
---|
0:36:14 | and that's the only place think impair you have to take care about actually having |
---|
0:36:18 | twice the amount of pixels on each axis then you see as virtual pixel so |
---|
0:36:23 | i don't know how takes cost and i think from what particle rumours that show |
---|
0:36:28 | how many pixels a but we but you should be i mean and |
---|
0:36:33 | if you have to image in one to one |
---|
0:36:36 | right so like hundred percent than you will actually show a little square per image |
---|
0:36:40 | pixel maybe because i'm just wondering how like about a picking up pixels like rulers |
---|
0:36:49 | why be a that's sure that a little numbers permission expand scale my |
---|
0:36:55 | accordingly your i that's just the front pixels like the virtual pixels you |
---|
0:37:02 | you use an answering everything else i mean the text we get smaller and to |
---|
0:37:06 | get a |
---|
0:37:07 | rendered as a vector graphics anyway so it look at risk |
---|
0:37:10 | but the ruler chicks they should be just in your normal pizzas like not the |
---|
0:37:15 | device pixels but you know from pixels |
---|
0:37:18 | and only if you scale out so you your image source has |
---|
0:37:22 | the canvas itself has actually more pixels available then you want to display then you |
---|
0:37:26 | can go hi dpi and actually show more this shall know exist and then normally |
---|
0:37:33 | but that's the only case i need to adapt to make it look |
---|
0:37:37 | everything else pictures of the box |
---|
0:37:39 | even the i consider |
---|
0:37:55 | i'm really surprised by a minimal or no in terms of you know |
---|
0:38:01 | and like the it just works for a radio readers |
---|
0:38:07 | very |
---|
0:38:09 | i haven't really tried it on absent you custom lot until an expensive of |
---|
0:38:18 | partially because most ominous for why just lot are excellent |
---|
0:38:21 | or not really doing a lot of course open G L anymore |
---|
0:38:25 | i don't know if i would be other pride but registration or |
---|
0:38:30 | expand you need to fix |
---|
0:38:33 | stuff that works but and that it should but thank you just work |
---|
0:38:39 | and it's really nice to hire |
---|
0:38:42 | right next everything is |
---|
0:38:45 | automatically scale |
---|
0:38:50 | it does what |
---|
0:38:53 | but does it also thanks |
---|
0:38:55 | does it get the size of a X I D from X |
---|
0:38:59 | minutes or more |
---|
0:39:00 | but it should |
---|
0:39:02 | as long as it doesn't get the case size and believes that was like you |
---|
0:39:08 | just |
---|
0:39:11 | actually i think |
---|
0:39:13 | there is one issue and when can doesn't currently okay to |
---|
0:39:17 | twelve double size double but for so what okay and was asleep kind of excellent |
---|
0:39:23 | my |
---|
0:39:26 | which you can use you pictures and |
---|
0:39:30 | what do you can ace |
---|
0:39:33 | which toolkits |
---|
0:39:35 | well right now it's just |
---|
0:39:36 | okay two or three two am in three a to your whatever you whatever work |
---|
0:39:42 | because it |
---|
0:39:44 | has the full set of for X wondering it guys |
---|
0:39:49 | we don't wanna be alone |
---|
0:39:52 | it does right and you doing because although i have that so |
---|
0:39:59 | don't wanna do that |
---|
0:40:03 | it's just kind of bad but |
---|
0:40:07 | just wanted i already |
---|
0:40:10 | i was wondering how this works with the hinting this car now |
---|
0:40:16 | exactly the dpi that is gonna be |
---|
0:40:20 | and working out or is it took to a it the other the other thing |
---|
0:40:25 | i choose |
---|
0:40:26 | thinking through a little bit about the clapped aside the ten |
---|
0:40:31 | which is going to require or |
---|
0:40:35 | we're gonna need to make sure that ten |
---|
0:40:37 | when it's going to knows about the scale point to |
---|
0:40:41 | huh |
---|
0:40:42 | to do that |
---|
0:40:44 | i mean really looked at it and i just assume that in the you know |
---|
0:40:48 | that |
---|
0:40:49 | but what if we just trucks care what to do when the technicians when as |
---|
0:40:53 | you look |
---|
0:40:54 | but there should be some |
---|
0:40:56 | it's also the for matrix and everything i just have think more they share |
---|
0:41:01 | so just kind of words statistically well in and well to it |
---|
0:41:08 | slightly or my maybe give you a slightly different layout your widget sizes |
---|
0:41:14 | because maybe a label was have extra later |
---|
0:41:18 | or whatever and |
---|
0:41:21 | we're gonna pick a different part |
---|
0:41:24 | or want your phones where the metrics are slightly differently around and whatever |
---|
0:41:30 | it's not a problem right now because we only support one |
---|
0:41:34 | surface before per window so it's always to just one scale |
---|
0:41:38 | but maybe we wanna okay having some way to |
---|
0:41:43 | always get the same kind of |
---|
0:41:48 | well did you think of a turning hitting a really want to go to the |
---|
0:41:52 | like three or something |
---|
0:41:55 | then that's right out of high dpi text |
---|
0:41:59 | and various ways that those that |
---|
0:42:02 | and what does it |
---|
0:42:04 | i have not really thought about it |
---|
0:42:07 | it's something we should maybe look at |
---|
0:42:10 | maybe hunting doesn't really matter that size |
---|
0:42:15 | i think actually practise it will be disable mostly |
---|
0:42:19 | basic of origin of fonts the hands are not |
---|
0:42:22 | there for that kind of size at least like |
---|
0:42:26 | what the really details |
---|
0:42:29 | manual point handing is just not enable and |
---|
0:42:33 | forty pixel point awesome |
---|
0:42:35 | so we kind of get away |
---|
0:42:40 | as possible that we want to have |
---|
0:42:43 | we used to |
---|
0:42:46 | declare your fonts metric intense whatever one on widgets |
---|
0:42:52 | try to make in exactly the same on different scales |
---|
0:43:03 | you know |
---|
0:43:05 | and something to work well worth looking into |
---|
0:43:10 | actually but not did a lot more content for |
---|
0:43:14 | crow i guess but we can ask |
---|
0:43:18 | here |
---|
0:43:27 | more question |
---|
0:43:31 | and they brought first ask me excuse you're gonna talk another and a for firefox |
---|
0:43:39 | actually passed this |
---|
0:43:45 | kind of hidden but if we make that scale that we all the detect |
---|
0:43:52 | make that |
---|
0:43:53 | and |
---|
0:44:00 | make that available |
---|
0:44:10 | but |
---|
0:44:12 | very see okay |
---|
0:44:20 | so the this is layouts use S them pixels per pixel |
---|
0:44:25 | if you set that to choose that makes basically needs to scale |
---|
0:44:29 | every pixel by to |
---|
0:44:31 | if we just get gave the finals all people way to we the current scale |
---|
0:44:37 | should be simple melodies pick that up |
---|
0:44:41 | and it more |
---|
0:44:44 | that's it's extension you have it's not going upstream a like if you propose that |
---|
0:44:49 | in the stack or anything |
---|
0:44:51 | not really i mean if you look at how all multiple |
---|
0:44:56 | that's you know alternative things work on the web right now it's like this horrible |
---|
0:45:00 | mismatch of |
---|
0:45:02 | defined everything |
---|
0:45:04 | there's like and javascript the runs |
---|
0:45:07 | and rewrites that you're always the one |
---|
0:45:10 | problems all over the don't felix for images and places them that or do you |
---|
0:45:15 | know this way |
---|
0:45:22 | maybe i mean |
---|
0:45:28 | and it's nice a that |
---|
0:45:33 | i wanna be a |
---|
0:45:35 | and indexation this |
---|