Contribute to Open Source. Search issue labels to find the right project for you!

Trigonometric functions in CMSIS-DSP


CMSIS-DSP offers only few trigonometric functions (arm_sin, arm_cos and arm_sin_cos, as far as I can see). Some other functions are useful for my needs; I mean:

  • tangent - I know I can calculate it using arm_sin and arm_cos

  • atan2

  • arcsinand arccos

For the end developer it’s quite difficult to implement the inverse trigonometric functions. I’m almost interested in q31_t format.

Do you think they would be available in a future release?

Updated 14/12/2017 13:27 1 Comments

Future: LineGrabPlot enhancements


Some usability improvements that might be nice to add to the plot in future:

  • [ ] Add an inset plot to the LineGrabPlotter that shows the users location when zoomed in significantly. For reference: Like the left side plot, but actually inset into the main Axes.
  • [ ] Enable users to set/control the graphical decimation ranges (i.e. set the % where a specific slicer is applied, and specify step parameter for the slicer).
  • [ ] Allow users to enable/disable decimation with a checkbox or radio button on the chart area.
  • [ ] Display the current decimation level somewhere on the chart area.
  • [ ] Investigate using a simple filter function as opposed to the simple slice/step to sample the array for graphical display.

These are all very low priority, and I just want to record some ideas for future use as I worked through the recent plot code improvements.

Updated 12/12/2017 20:34

Improve namespacing


These changes replace the isSVG flag with a namespace string. Namespacing is automatically triggered by using svg or math tags, and inherited by children. Namespacing can be explicitly triggered by adding a namespace to the props object passed to h. This is useful for e.g. embedding HTML inside SVG via the foreignObject tag.

Defined namespace URIs are added as a property of the h function, for convenience - saves having to look up and type them when doing explicit namespacing.

The other change is to always createElementNS even for html tags. There is no performance advantage to createElement

NB: my package.json and typescript definitions file are out of sync with the master - ignore please.

Updated 14/12/2017 13:57 10 Comments

Update to Babel 7.x



This is my very firs PR which includes Babel 7.x related updates. Please let me know if you will be interested in such contribution so I can make further updates in that matter.

Updated 13/12/2017 17:28 6 Comments

search - async query running

  • querySearch generates three queries for a given string, and returns the results for the first query that returns more than 0 results. This creates issues where the first query succeeds, but only gives one result.

  • change the search app to accept a minSearchResults param. Asynchronously query pc using each of the three queries until minSearchResults is returned. i.e do not stop fetching search results using the generated queries until you have minSearchResults

Updated 07/12/2017 20:18

Looking ahead: Feature additions for 1.1 through 2.0


In no particular order, just some features I’d like to implement in future releases:

  • [ ] Store colours like this
  • [ ] Dark/Light mode, dark by default
  • [ ] Look into adding UI themes
  • [ ] Look into extended functionality
  • [ ] Look into AppleWatch functionality
  • [ ] Look into iCloud saving
  • [ ] User lookup for farming level to auto recommend seeds
  • [ ] Option to schedule seeds / plan farming runs
  • [ ] Widget integration for upcoming timers
  • [ ] SiriKit integration for setting/reading timers
Updated 08/12/2017 20:18

Debug sequences inadequate for writing flashloaders/handlers


I’ve been looking at implementing flash loaders using debug sequences.

Q1. To get reasonable performance out of the loader, we’d need a couple more primitives: Q1A: Read/WriteBLOCK , and Q1B: PollBits. The first one is self explanatory, the second one would save USB roundtrips when polling for a ready or error bit. Q2: Also missing are definitions for the erase/write sequences including chip erase . Something like “FlashProgramPage” “FlashEraseRow” . Q3: Is it possible to give arguments to a debug sequence ? Page address and payload would be required for page writes.

Updated 06/12/2017 14:19 2 Comments

Story Arc


Hi Titus, Thank you for your work on retextjs. I used your simplicity plugin in an online editor. I’m planning to use more of your great plugins.

I have this dream of a tool that draws (graphically) the story arc based on text analysis. It could map your writing to the classical arcs as enumerated by Lafrance: 1. Rags to Riches [a complete rise]; 2. Riches to Rags [a fall]; 3. Man in a Hole [fall then rise]; 4. Icarus [rise then fall]; 5. Cinderella [rise then fall then rise]; 6. Oedipus [fall then rise then fall].

I know it’s a moonshot, but yeah.

Updated 05/12/2017 10:57 2 Comments

Graph diff view for two graphs


use case: - pc10 comes out and maybe the sbgn output has changed as a result, this tool should provide help in terms of porting a pc-N custom human layout to pc-N+1

design inspiration: - git diff in github/visual studio code

literature: Structural Differences Between Two Graphs through Hierarchies: - functionality: - for each node in the previous graph - report if it has the same edges or neighbours - report if it still has the same parent - report if it has the same data (class, label, statevars/units of info) - report if it has been removed from the graph (i.e doesn’t exist in the new graph)

  • for each edge in the previous graph

    • report if it still has the same src/target
    • report if it has the same data (class, label, statevars/units of info)
    • report if it has been removed from the graph (i.e doesn’t exist in the new graph)
  • for each new node/edge

    • denote that it is new via a green colour / colour that means ‘new’.
Updated 03/12/2017 16:15 1 Comments

Paint - generalized paint


use case, user wants to paint a few nodes without going through cytoscape

same table view as the current paint app, instead of expression data, each node in the cy.js graph has a row - user clicks on the row to enter value, value is the painted onto the node - configurable legend (min, max) to determine colour/opacity

Updated 06/12/2017 22:58 2 Comments

Open API 3.0


Open API 3.0 adds a few things like Bearer token. However, there is no annotation support for it as of yet so this will be postponed until the supporting technologies come up.

Updated 03/12/2017 02:44 1 Comments

Streamline repo scripts


I’m building the packages for the binary repo manually. The repo is enabled by default in pmbootstrap (disable with --mirror-pmOS="") and it is somewhat usable now and it’s a big improvement over having no binary packages at all. But it still needs improvements.

  • [ ] Change pmbootstrap aportgen to generate gcc-aarch64 not for arch=all but only for specific ones (x86_64 only for now?), otherwise it will build versions for aarch64 and armhf
  • [ ] use --strict for building
  • [ ] improve the building script: don’t use pmbootstrap directly, wrap it with a different config and work path, so the same PC can be used for quickly testing a build without messing with the repo
  • [ ] polish repo build scripts and publish them
  • [ ] properly delete old packages (right now we keep them all!)
  • [ ] use proper repo description with pmbootstrap commit hash
  • [ ] use -9 compression for gzip packages (new pmbootstrap parameter)
  • [ ] Firmware packages: download blobs after installation, so we can build firmware packages, too (right now they don’t get built) (#797)
  • [ ] Some packages will still build despite the same version being on the binary repo (#955)
  • [ ] add pmbootstrap build --all to simply build all packages?

and finally (future!): - [ ] automatize it completely: whenever a new commit comes in to master, run the build

I would tackle all of this some time after the 200d blog post, if that is okay @MartijnBraam (you host the mirrors - do we have enough space?).

Updated 02/12/2017 17:37

Add tool that takes the name of a PyPI package, searches out its docs, and pulls its objects.inv if it can find it


Long-shot. Would want this to be a function separate from the Inventory instantiation, because it might take some time (multiple web requests, etc.), and it’ll almost certainly take the use of requests and/or scrapy, or similar.

Search PyPI for documentation or source repo links; follow source repo and look for doc links there; if a ‘doc-like’ page is found, attempt loading the objects.inv from one or more likely spots.

Updated 02/12/2017 14:57

Very large image not supported (area > 2^32)

  • OpenCV => 3.3.1
  • Operating system => Windows 10 Pro (64bit)

Related forum link

Detailed description Large images with an area greater than 232 cause crashes or do not work in some functions. A simple example is the cv::threshold method with Triangle method flag: The operation size.width *= size.height; is done (in file imgproc/src/thresh.cpp, line 1099), but in the case of large images, this overflows the “small” integer int.

Moreover, images with area greater than 230 cannot be opened (see imgcodecs/src/loadsave.cpp, line 65). As a workaround, I just modified the mentioned line with #define CV_IO_MAX_IMAGE_PIXELS((uint64)1 << 40, but this can’t be a long term fix.

Suggestions Maybe using other types (like int64 or size_t) for size-related (or maybe index-related) code may be a good idea. Maybe this can be activated via a preprocessor directive.

Updated 04/12/2017 15:46 4 Comments

Preserve copyright comments in the output


Hi Folks! Quick question, “How do we preserve JSDoc-style copyright comments in the output?”


 * @license 

 * Component.js 2.1.5
 * (c) 2017 ACME, INC
 * License agreement...
 * By using this software, you agree that you have read, understood, and will comply with the license terms and conditions.

 * Please visit to learn more.
var _0x78sgf=['u1/as9raasd=','asd8asd3=' . . .

Thinking something similar to the uglify-js --comment flag: Preserve copyright comments in the output. By default this works like Google Closure, keeping JSDoc-style comments that contain "@license" or "@preserve".

Thanks and fantastic project!

Updated 12/12/2017 21:25 3 Comments

[2.0] Future


LegoFlow 对外服务以来,应对的是更加复杂的 PC 环境,感谢各位的反馈,让应用的稳定性有更大的提升,功能上也得到了补充。

而其核心依赖的模块经过这一两年的发展,也有了重大的更新,为了适配这些更有效率的特性,LegoFlow 将会升级到 2.0 版本。


  • [ ] Webpack 3x or 4x ( IE8 is no longer supported )
  • [ ] Babel 7
  • [ ] 支持 ES next
  • [ ] UI & 交互优化 ( 主要涉及 项目列表 与 通知 两个栏目 )
  • [ ] 功能减法 ( 一些额外功能交由自定义编辑 Shell 脚本完成 & 衡量核心功能点 )
  • [ ] 构建优化 ( override 通过模块自动完成 )
  • [ ] Open source or version cli
  • [ ] 构建 -> 部署
  • [ ] Shell Store
  • [ ] ? PWA 脚手架
  • [ ] ? 线上 构建


Updated 07/12/2017 02:53 2 Comments

Feature Phone UI


(This seems to be a recurring topic in the channel, so I’m writing it down in an issue.)

There was this idea floating around of making a feature phone UI, that makes any touchscreen smartphone look and behave like one of these old nokia phones. Possibly with extra big buttons and text.

It could be useful for old people, who are used to such devices and don’t want to learn a touch screen based UI. And in general it wouldn’t eat up much resources.

If someone is interested in this, go ahead and make something! UI could be based on libweston for example, that way it would work with all devices where weston is already working.

Updated 29/11/2017 08:26 7 Comments

Better exception for Source.detach() on detached sources


Could this get its own exception type? I want to give customers the option to remove a source from their payment methods, and right now I have to do something like this:

except NotImplementedError:

Catching notimpl is weird, and this isn’t really the appropriate exception type I think anyway.

Updated 28/11/2017 09:55 3 Comments

Request retries on failure



I was wondering if a request is retried in case of failure. I read about idempotency recommendations :

The Stripe Ruby library retries on failure automatically with an idempotency key using increasing backoff times and jitter

Does stripe-php has a similar feature ? If not, would it be interesting to implement it ?

Updated 24/11/2017 15:52 1 Comments

Improve player spritesheets


Ideally I would like to implement a character editor, but the MVP of this would be high quality or vector spritesheets.

This is marked as future as I can start developing the game and add this later.

Updated 24/11/2017 06:57

tagging layouts to distinguish between human created and alogrithmic


Is it possible when submitting a layout to the server to ‘tag’ it with a category.

use case: - someone runs any algorithmic layout in /edit for any uri - when it is submitted to the server, it is tagged ‘algorithm’ or something that denotes that it was computed by an algorithm - when someone loads that same uri, query the db for all layouts for that uri that have a tag name of algorithm, or get the latest algorithm layout

Another idea is adding query params to layout requests:

e.g /getLayout&uri=…&name=Force Directed

Updated 04/12/2017 03:24 1 Comments

Restructure Overlay Attribution Credits


Currently all overlay data supplied by LCRA and the first round of lakes will all be in Region K with overlay data supplied by LCRA so the layer ‘attributes’ credit is hard coded for LCRA.

Need to update this to a dynamic system when the lake content expands more statewide (outside Region K) and overlay data is sourced from other entities.

Updated 20/11/2017 16:11

Theme discussion


Hi @tobiasalthoff! Regarding on your concept of separating widget colors, we can’t implement this for now because vscode doesn’t supports custom theme CSS stylesheets :slightly_frowning_face: (see vscode#459) Even so I’d like to have less contrast widget without colored border. What about you @tobiasalthoff? which one do you like: bordered.. 2017-11-20 02-12-30 or without border? 2017-11-20 02-14-57

Updated 24/11/2017 19:16 2 Comments

Is it possible to implement an onStopTrackingTouch?



Is it possible to add support for the function onStopTrackingTouch present in the standard SeekBar ? (i.e. it is called when the user stops changing the progress of the bar). It really helps when we need to change an attribute only when the user sets the value “as final”, without having to write to a complex object each time the range changes.

Thanks in advance!

Updated 14/12/2017 02:08 1 Comments

Add hop-to-image-location lat/long


Add an associated lat/long option for each Story Content image which would be populated with the location the image was taken in reality. Then we can add a ‘hop-to’ button for those images with a lat/long, when clicked will zoom and center the map to that location.

This might be a larger scale change and require restructuring the handling of images in general with association to Story Content. (Currently, the Story Content form feels cumbersome and adding additional ‘photo latitude’ and ‘photo longitude’ fields for every image in that form simply doesn’t feel right). Should images be in a separate table?… if so, how are they strategically and deliberately placed in specific places amongst the story text?

Updated 16/11/2017 17:25 1 Comments

Intact data


We should review what pombe interactions are curated at some point:

IntAct: access IMEx as a single, unified dataset

We are happy to announce that the IMEx Consortium is making molecular-interaction data more accessible than ever before. Users are now able to access all datasets through a unified source, as IMEx partners DIP, IntAct, MINT, UniProt, MatrixDB, HPIDB and others have merged their offerings. Users can download the data in several formats: PSI-MI XML, MITAB and XGMML.

We will announce improvements to the IMEx website and a new common query interface for IMEx (powered by PSICQUIC, the Proteomics Standard Initiative Common QUery InterfaCe) in the coming weeks.

Updated 22/11/2017 17:59

Triaging Functional Test Failures


<!– Thank you for contributing to Dojo 2.

Our issue tracker is for bugs for Dojo 2.

Please make sure you have read our Contributing Guidelines available at:

For general questions and discussion, join us on at: –>

Enhancement <!– delete as appropriate –>

A number of functional tests in widgets either employ workarounds or skip specific browsers to allow them to run. Known driver bugs are also skipped by targeting browsers. We should try to avoid targeting browsers and instead of using the available APIs to do feature detection, etc.

Skips and workarounds

Numbers in (parens) refer to bugs below.

(1) Calendar: Edge driver does not handle mouseup on click. (8?) Calendar: Skipped if either supportsKeysCommand is false or the browser is Safari “Arrow keys must be supported” The focusing click event happens on a td element (2) Calendar: Edge driver does not handle focus on click

(3, 4) ComboBox: FirefoxDriver and SafariDriver update the input value with non-printable characters. (5) ComboBox: SafariDriver does not move focus with tab key.

(6, 7) Dialog: Edge and Firefox are also having trouble with tab key focus. (8) Dialog: SafariDriver doesn’t recognize focus on divs

(9) SlidePane: Edge driver does not handle mouse movements correctly.

Slider: All tests skipped in Edge (“example page doesn’t work in Edge”) (8?, 10) Slider: “Click to focus” tests skipped in Firefox and Safari (11) Slider: mouse movements doesn’t work in IE. (12, 13) Slider: “slider should be slidable with mouse” tests skipped in Firefox and Safari (14, 15) Slider: pressKeys with arrow keys doesn’t work in iphone and IE.

(16) SplitPane: All tests skipped in Safari, Firefox, and Edge (StaleElement, undefined target)

(17) Textarea: getVisibleText includes the textarea placeholder in IE (18) Textarea: Firefox is not able to locate the input. (19) Textarea: Classes are not being updated in Safari 9

(20) Timepicker: Firefox throws an error when clicking on the disabled input. Timepicker: Test does not work on Internet Explorer

(21) ALL: return document.activeElement === document.querySelector('#example-s1 .${css.root} input'); is used in place of APIs


  1. Edge driver doesn’t send mouseup event on click
  2. Edge driver does not handle focus on click
  3. FirefoxDriver updates the input value with non-printable characters
  4. SafariDriver updates the input value with non-printable characters (
  5. SafariDriver does not move focus with tab key ( Potentially fixed
  6. Edge does not move focus with tab key Other tab focus tests seem to pass
  7. FirefoxDriver sends actual charcodes to the input
  8. SafariDriver doesn’t recognize focus on divs
  9. Edge driver does not handle mouse movements correctly (
  10. Click to focus not working in Firefox
  11. Mouse movements not working in IE
  12. Mouse movements not working in Firefox
  13. Mouse movements not working in Safari
  14. Arrow keys don’t work in mobile Safari
  15. Arrow keys don’t work in IE
  16. Stale/undefined targets encountered during tests
  17. getVisibleText includes the textarea placeholder in IE
  18. Firefox not locating nodes
  19. Safari is not detecting node class updates
  20. Firefox throws an error when clicking a disabled input
  21. findByCssSelector followed by .getActiveElement() may result in different elements
Updated 29/11/2017 10:43

P0829 and freestanding implementations


P0829 proposes to update the requirements for freestanding implmentations of the C++ standard, with notably a plea to make most of the header <algorithm> available to those implementations. Since cpp-sort mostly provides algorithms, it would be interesting to make most of it available for freestanding implementations too once the committee agrees in what should be available. We could introduce a simple CPPSORT_FREESTANDING switch to require that cpp-sort builds with freestanding implementations.

A few things wouldn’t be available anyway: * A few algorithms such a merge_sort allocate dynamic memory: those should either be excluded or be made to fallback to non-allocating versions, which might still be desirable. It would be a good reason to replace the deprecated std::get_temporary_buffer by a safer custom solution that would simply return nullptr in freestanding mode. * Algorithms specifically dealing with floating point should be excluded because freestanding implementations wouldn’t be required to support floating point types. * The evolutions proposed in #22 wouldn’t be available since they would rely on threads and standard execution policies. * A few math functions might have to be replaced by hand since most of them can set the thread_local variable std::errno.

On the other hand, we could introduce more compile time switches through macros to have more granularity: no floating point numbers, no memory allocation, no threads, etc…

I will keep track of the progress of P0829 and probably start working when the committee starts to get close to a new specification for freestanding implementations.

Updated 13/11/2017 17:52

Fork me on GitHub