[Rd] Writing in R // R 4.2.0 on Windows doesn't work with Dasher

Paulo Barata pb @end|ng |rom |n|o||nk@com@br
Wed May 18 13:41:19 CEST 2022


Dear Tomas,

Today, I installed R version 4.2.0 Patched (2022-05-17 r82371 ucrt) on 
my Windows 10 Professional. Dasher works perfectly well with that 
version of R.

Again, using Dasher, I tried special symbols, letters (including 
accented ones), digits, space, the "Enter" character and the "Backspace" 
movement, and I wrote a range of commands in the R language, everything 
worked well.

I could install some packages normally in that version of R, no problems.

I think that the Dasher issue has been solved. I will use this R 4.2.0 
Patched r82371 from now on.

Again, thank you very much, and all the other people who took part in 
this discussion, for all your attention.

Best regards,

Paulo

----------------------------------------------------------

On 17/05/2022 10:45, Tomas Kalibera wrote:
> Dear Paulo,
> 
> On 5/17/22 15:04, Paulo Barata wrote:
>>
>> Dear Tomas,
>>
>> Great news! I downloaded R-devel 2022-05-16 r82368 for Windows and 
>> installed it on my Windows 10 Professional. It works perfectly well 
>> with Dasher 5.0 beta!
>>
>> In Dasher, I tried all special characters like $ & | *, many small and 
>> capital letters, including some accented ones like á À ê ç that exist 
>> in Portuguese (and other languages, e.g. French (there are versions of 
>> Dasher for many languages)), all digits, space, the Enter character 
>> and the "Backspace" movement (to erase previously written characters), 
>> everything works well.
>>
>> Now, we can fully use Dasher in the R console.
> I am happy it works again, thanks for the testing and the report. I've 
> ported the fix to R-patched, which means it will appear in R 4.2.1.
>> It seems that we cannot yet normally install packages in this R-devel 
>> r82368. So I look forward to the day when a stable R 4.3.0 will be 
>> available. Meanwhile, for my ecological work, I can use R perfectly 
>> well with the versions 4.1.3 and 4.2.0 that I have here, so, from my 
>> viewpoint, this R 4.3.0 can take its time to come to light.
> 
> R 4.3.0 will be released probably around next April, but R 4.2.1 much 
> sooner (expect in 1-3 months, though this hasn't been scheduled, yet). 
> Before that, you should be able to use an R-patched snapshot build 
> available from
> 
> https://cran.r-project.org/bin/windows/base/rpatched.html
> 
> but only after the build is of release >= 82371 (which should happen by 
> tomorrow - the page shows the revision number). R-patched is R 4.2.0 
> with updates that will eventually become 4.2.1, so should be more stable 
> than R-devel is now (to become 4.3.0 around next April).
> 
> If you have any problems with installing packages in either version of 
> R, please report, so that someone can help or indeed if there is a bug, 
> it could be fixed.
> 
>> I can only thank you very much, and all the other people who took part 
>> in this discussion, for the quick fix to this issue.
> 
> Thank you for the report. Without it, the problem wouldn't have been fixed.
> 
> Best
> Tomas
> 
>>
>> Best regards,
>>
>> Paulo Barata
>>
>> (Rio de Janeiro - Brazil)
>>
>> ----------------------------------------------------------
>>
>> On 16/05/2022 14:08, Tomas Kalibera wrote:
>>>
>>> On 5/15/22 16:10, Paulo Barata wrote:
>>>>
>>>> To the R developers,
>>>>
>>>> I noticed today that R 4.2.0 responds to the "Enter" character of 
>>>> Dasher version 5.0.0 beta in "Direct Entry" mode. An "Enter" in 
>>>> Dasher causes the R prompt to jump to the next line, as an "Enter" 
>>>> should do.
>>>>
>>>> No other characters (letters, digits, special characters like + = & 
>>>> ~ | # $) are accepted by R 4.2.0 through Dasher.
>>>>
>>>> The "Enter" shows that, somehow, the communication between Dasher 
>>>> and R 4.2.0 through the Windows system is not completely curtailed.
>>>
>>> Dear Paulo,
>>>
>>> thanks for the report and thanks to all for the useful comments in 
>>> the followup discussion. I've committed a fix to R-devel, 82368, and 
>>> I would be grateful if you could test whether it fixes the problem 
>>> for you. I've tested it only on a small example which is using the 
>>> same method to insert keys to Rgui as Dasher does.
>>>
>>> The problem is caused by a limitation in GraphApp, a library used by 
>>> Rgui and included in the R distribution. GraphApp has two modes: 
>>> Unicode windows (with this limitation) and non-Unicode windows 
>>> (without it). GraphApp uses Unicode windows when running in a 
>>> multi-byte locale, and as of R 4.2, this is the case for all systems 
>>> on recent Windows where the native encoding is UTF-8. This means that 
>>> most likely R 4.1.3 didn't work with Dasher on systems running in a 
>>> multi-byte locale and that most likely R 4.2.0 would work with Dasher 
>>> on old Windows systems (where UTF-8 cannot be used as native encoding).
>>>
>>> Dasher uses SendInput/KEYEVENTF_UNICODE to inject text into Rgui, 
>>> which is fine as far as I can tell and should work. It didn't work, 
>>> because the Unicode windows didn't handle WM_CHAR nor 
>>> WM_KEYDOWN/WM_KEYUP for VK_PACKET. By design, Unicode windows get 
>>> input via WM_IME_COMPOSITION and WM_KEYDOWN, so I've added support 
>>> for VK_PACKET to WM_KEYDOWN (taking the logic in gdk as an example).
>>>
>>> I am cc'ing Jose as Tinn-R has a similar problem, though this patch 
>>> doesn't solve it.
>>>
>>> Best
>>> Tomas
>>>
>>>>
>>>> Regards,
>>>>
>>>> Paulo Barata
>>>>
>>>> ----------------------------------------------------------
>>>>
>>>> On 13/05/2022 16:44, Paulo Barata wrote:
>>>>> To the R developers,
>>>>>
>>>>> I would like to point out an issue that occurs with the new R 4.2.0 
>>>>> running on Windows; this is possibly related to the recent wide 
>>>>> changes made to the Windows port of R. The issue has to do with 
>>>>> writing of the R window.
>>>>>
>>>>> My computer runs on Windows 10 Professional 64-bit with all the 
>>>>> latest updates, and has a 4-core 8-thread Intel Xeon processor with 
>>>>> 32 Gb of RAM.
>>>>>
>>>>> I have some slight impairment in my hands, so for typing I avoid 
>>>>> using the computer keyboard as much as possible. For most of my 
>>>>> typing, I use the software Dasher, a kind of assistive technology 
>>>>> software. Dasher was created by the Inference Group led by 
>>>>> Professor David MacKay (died in 2016) at Cambridge University in 
>>>>> England. Later, Dasher was further developed by Ada Majorek (a 
>>>>> Google engineer, recently deceased).
>>>>>
>>>>> General information on Dasher and about the Dasher Project can be 
>>>>> found on the Inference Group page:
>>>>>
>>>>> http://www.inference.org.uk/dasher/
>>>>>
>>>>> An example of Dasher in action can be seen on this 2-minute YouTube 
>>>>> video:
>>>>>
>>>>> https://www.youtube.com/watch?v=nr3s4613DX8
>>>>>
>>>>> An one-hour YouTube video of a 2007 Google Talk by Professor MacKay 
>>>>> himself explaining the Dasher concept can be found here:
>>>>>
>>>>> https://www.youtube.com/watch?v=wpOxbesRNBc
>>>>>
>>>>> In the 2-minute video (the first YouTube address above), the 
>>>>> sentence written through Dasher, for actual use, needs to be copied 
>>>>> and then pasted into another software, say a text editor. But 
>>>>> Dasher also has a "Direct Entry" mode, where words typed are 
>>>>> entered directly on the screen of the text editor or other software 
>>>>> -- this is how I use Dasher to write on the R window.
>>>>>
>>>>> Using Dasher`s "Direct Entry" mode, for all effects, to write a 
>>>>> word through Dasher works exactly like writing the word through the 
>>>>> normal computer keyboard. Somehow (I do not know about system 
>>>>> programming), in "Direct Entry" mode, the output of Dasher is 
>>>>> understood by the computer as an input of the same nature as a 
>>>>> keyboard input.
>>>>>
>>>>> But this is the issue: until R 4.1.3, R accepted normally the 
>>>>> word-typing through Dasher. But R 4.2.0 (and R 4.2.0-patched build 
>>>>> r82348, 12 May 2022) does not accept Dasher anymore.
>>>>>
>>>>> So, for now, if I wish to use R during some extended interactive 
>>>>> section (say, for a first look at a complex data set), and if I 
>>>>> wish to use Dasher for this interactive session, then I have to use 
>>>>> R 4.1.3. For using R through programming, I can use the new R 
>>>>> 4.2.0, by writing the program with Dasher in my text editor, and 
>>>>> then pasting that program into R.
>>>>>
>>>>> The latest Windows version of Dasher, version 5.0.0 beta, works 
>>>>> perfectly well with text editors (I use EditPad Pro), with 
>>>>> Microsoft-like Office packages (I use SoftMaker Office), for 
>>>>> writing e-mails (I use Mozilla Thunderbird), for writing on a web 
>>>>> page (I use mostly Mozilla Firefox). It worked perfectly well with 
>>>>> R up to version 4.1.3, but it does not work with R 4.2.0
>>>>>
>>>>> The latest Dasher 5.0.0 beta Windows installer and the source code 
>>>>> can be found on this page below (please click on "Windows 
>>>>> installer" and on "source code", just below the head "Dasher 5.0 
>>>>> beta release"):
>>>>>
>>>>> https://github.com/dasher-project/dasher/releases
>>>>>
>>>>> To set Dasher 5.0.0 beta to the "Direct Entry" mode, after 
>>>>> installing Dasher one should go to Preferences > Application > 
>>>>> Application Style.
>>>>>
>>>>> Currently, after the passing of Ada Majorek, the Dasher Project is 
>>>>> being put forward by a group of people in England, but, it seems, 
>>>>> at a very slow pace. They have announced the existence of a version 
>>>>> 6 beta, but this is only an online instance of Dasher, there is no 
>>>>> Windows installer, as far as I know. Their web page:
>>>>>
>>>>> https://dasher.acecentre.net/
>>>>>
>>>>> For me, and possibly for other users, the Dasher version 5.0.0 beta 
>>>>> for Windows that is available works perfectly well, and allows me 
>>>>> to use the computer with a proficiency that would be hardly 
>>>>> possible otherwise.
>>>>>
>>>>> Here I request that the R developers take a look at this 
>>>>> word-typing issue that happens with the new R 4.2.0 on Windows.
>>>>>
>>>>> As an R user (I work in ecological research), I can only thank the 
>>>>> R team for this really wonderful piece of statistical software that 
>>>>> you make freely available to us all.
>>>>>
>>>>> Regards,
>>>>>
>>>>> Paulo Barata
>>>>>
>>>>> (Rio de Janeiro - Brazil)
>>>>>
>>>>> ______________________________________________
>>>>> R-devel using r-project.org mailing list
>>>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>>>
>>>> ______________________________________________
>>>> R-devel using r-project.org mailing list
>>>> https://stat.ethz.ch/mailman/listinfo/r-devel



More information about the R-devel mailing list