github/shiny
1
1
Fork 0
mirror of https://github.com/rstudio/shiny.git synced 2026-02-08 13:45:28 -05:00
Code Issues Packages Projects Releases Wiki Activity
3,653 Commits 139 Branches 79 Tags
6e44915e08e22271b1a91c74ad0135e217226324
Commit Graph

3653 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Joe Cheng
6e44915e08 Merge pull request #1984 from rstudio/joe/feature/faster-deep-stacks 
Lazily format stack traces
 2018-03-27 10:31:00 -07:00
Barbara Borges Ribeiro
66501dac97 Add new autoclose = TRUE param to dateInput() and dateRangeInput (#1987) 
* Add new `autoclose = TRUE` param to both dateInput() and dateRangeInput()

* added NEWS item
 2018-03-23 09:40:19 -07:00
Joe Cheng
bc7cd21c13 Update NEWS.md 2018-03-20 16:39:52 -07:00
Carson Sievert
0555cbdd28 relay offsetWidth/offsetHeight of htmlwidgets to clientData (#1981) 
Addresses #1980
 2018-03-20 16:37:43 -07:00
Joe Cheng
97498451bb Lazily format stack traces 
With deep stack traces enabled, whenever then() is called, we need
to grab the current stack, just in case a downstream callback throws
an error and we need to form a deep stack trace.

Previously, we were calling formatStackTrace at the time that we
grab the current stack (i.e. no error has happened yet) because I
wasn't sure whether holding a reference to sys.calls() for a long
time was a good idea from a garbage collection perspective; would it
prevent the stack frame environments from being collected? But the
answer is no, sys.calls() is just calls, which can be confirmed with
.Internal(inspect(sys.calls()).

By deferring the formatStackTrace call to when we actually need to
print the stack trace, we save ourselves a ton of work--it turns out
it's quite expensive to format the stack traces, much more expensive
than sys.calls() alone.
 2018-03-20 12:35:17 -07:00
Joe Cheng
2e0d9b5475 Bump httpuv dependency version 2018-03-18 19:07:43 -07:00
Barbara Borges Ribeiro
62395f3103 Improve error handling when addResourcePath() fails (especially for runtime: shiny_prerendered documents) (#1968) 
* A copy of yihui's PR for rmarkdown (https://github.com/rstudio/rmarkdown/pull/1171/) to avoid to error "Error in normalizePath: path[1]="": No such file or directory" when running any tutorial

* first try

* limited the scope of the `tryCatch` wrapper to the one important line that needed it; added news item
 2018-03-16 15:36:12 -07:00
Winston Chang
6b31cd6aee Merge pull request #1965 from rstudio/joe/bugfix/plot-dim-error 
Fix #1964: renderPlot cache breaks when width/height throw
 2018-03-01 11:12:56 -06:00
Joe Cheng
e67a8ba369 Fix #1964: renderPlot cache breaks when width/height throw 
Fixed by moving the isolate(getDims()) call into the (effectively)
try/catch that does a non-isolated getDims() if an error occurs
 2018-02-28 15:40:55 -08:00
Joe Cheng
133d301925 Merge pull request #1961 from rstudio/fix-date-sliders-bookmarking 
Fix URL-encoded bookmarking with date/date-time sliders
 2018-02-28 14:26:50 -08:00
Joe Cheng
17c40a5d1d Merge pull request #1960 from rstudio/slider-formatting 
Don't show commas after decimal mark in sliderInput
 2018-02-28 13:01:20 -08:00
Winston Chang
042211e5f6 Grunt 2018-02-28 14:34:09 -06:00
Winston Chang
d12830d700 sliderInput: don't show commas after decimal 2018-02-28 14:33:42 -06:00
Winston Chang
b411c70280 Fix URL-encoded bookmarking with date/date-time sliders 2018-02-27 20:42:11 -06:00
Winston Chang
2bc22cc7d5 Merge pull request #1955 from rstudio/update-slider 
Update ion.rangeSlider to 2.2.0
 2018-02-27 13:21:09 -06:00
Joe Cheng
b4c189c89b Merge pull request #1956 from rstudio/fix-slider-rounding 
Avoid rounding errors in sliderInput
 2018-02-27 10:55:11 -08:00
Winston Chang
fe3f351a2d Avoid rounding errors from pretty(). Fixes #1006 2018-02-27 10:50:36 -06:00
Winston Chang
076be9cba7 Remove unused keyboard_step parameter 
keyboard_step was removed in ion.rangeSlider 2.2.0.
 2018-02-26 15:57:30 -06:00
Winston Chang
f28dcd85fb Update to ion.rangeSlider 2.2.0 2018-02-26 15:57:30 -06:00
Joe Cheng
8e0f17c9d7 Merge pull request #1954 from rstudio/fix-bookmark-restore 
Look for restore context associated with session
 2018-02-26 09:25:33 -08:00
Winston Chang
d73817a0db Look for restore context associated with session. Fixes #1948 2018-02-26 10:55:38 -06:00
Joe Cheng
11874db825 Remove background-thread branch from httpuv remote 2018-02-16 15:41:32 -08:00
Joe Cheng
5d5a43ce90 Merge pull request #1932 from rstudio/async 
Async
 2018-02-16 07:41:36 -08:00
Winston Chang
75e548caab For installation of Rtools on Appveyor 2018-02-16 09:15:40 -06:00
Joe Cheng
c901e7ba06 Update TODO-promises.md 2018-02-13 15:23:06 -08:00
Joe Cheng
b1dc3dfca1 Restore label to plotObj reactive 2018-02-09 11:12:46 -08:00
Joe Cheng
ce4ed20c69 Fix remaining failing tests 2018-02-09 11:12:46 -08:00
Joe Cheng
d44df7f860 Stack traces were being lost inside hybrid_chain 2018-02-09 11:12:46 -08:00
Joe Cheng
54353e0e1f Fix coordmap unit tests 2018-02-09 11:12:46 -08:00
Joe Cheng
1c042b6efb Add to .Rbuildignore 2018-02-09 11:12:46 -08:00
Joe Cheng
b8df1f29c4 Remove unused function 2018-02-09 11:12:46 -08:00
Joe Cheng
18252f5b03 Use later >= 0.7.1 2018-02-09 11:12:46 -08:00
Joe Cheng
881370f284 Remove extraneous comments 2018-02-09 11:12:46 -08:00
Joe Cheng
35d1747bc3 Don't allow invalidation from a child process 2018-02-09 11:12:46 -08:00
Joe Cheng
91ac89a54e Update TODOs 2018-02-09 11:12:46 -08:00
Joe Cheng
3c694d9bd9 More robust process identification (thanks @HenrikBengtsson) 2018-02-09 11:12:46 -08:00
Joe Cheng
6a78e9df77 Detect (probably inadvertent) attempts to inherit reactive contexts across processes 
Example (we want this to fail):

library(shiny)
library(future)
plan(multicore)

r <- reactiveVal(TRUE)
isolate({
  f <- future({
    r()
  })
  value(f)
})
 2018-02-09 11:12:45 -08:00
Joe Cheng
078c6eb30a Add TODO 2018-02-09 11:12:45 -08:00
Joe Cheng
d35c6002a6 Respect pixelratio (retina) when redrawing cached images 2018-02-09 11:12:45 -08:00
Joe Cheng
f23fc3beaa Plots were not respecting pixel ratio (retina) 2018-02-09 11:12:45 -08:00
Joe Cheng
5a352e5ace Update TODOs 2018-02-09 11:12:45 -08:00
Joe Cheng
27cae0065e Fix bug where req(cancelOutput=TRUE) would leave things grey 
This was introduced by some changes to shinyapp.js that were necessary
before async outputs and sync outputs were held/flushed together. Now
that async/sync outputs are held/flushed together, these changes are
not necessary and removing them fixes the problem.

The test app is in shiny-examples/205-async-req. I also moved a test
app from manualtests/async/download.R to shiny-examples/204-async-download.
 2018-02-09 11:12:45 -08:00
Joe Cheng
50be2993fa Add TODO 2018-02-09 11:12:45 -08:00
Joe Cheng
d9ea15e9bc Update TODOs 2018-02-09 11:12:45 -08:00
Joe Cheng
03b1d45d7e Make deep stack traces opt-in; fix imports 2018-02-09 11:12:45 -08:00
Joe Cheng
e48d6878c4 Update TODOs 2018-02-09 11:12:45 -08:00
Joe Cheng
1a3b255848 Update TODOs 2018-02-09 11:12:45 -08:00
Joe Cheng
f00aa94d7e Suspend session during async download operation 2018-02-09 11:12:45 -08:00
Joe Cheng
f7980b19f4 Update TODOs 2018-02-09 11:12:45 -08:00
Joe Cheng
6a1f9677a5 Support async downloadHandler content functions 
If a downloadHandler content function returns a promise (or future)
then Shiny will wait for the promise to resolve before serving up
the file download.
 2018-02-09 11:12:45 -08:00