...

Has anyone come up with a quick workaround for it ??

Yes, I'll working on it (help for R >=2.10 and R < 2.10) if I finished a very large and urgent R project (using 2.9.2). Thus please be patient.

Cool, thanks!

Since I haven't installed R 2.10 yet, I have some tiny question:

1. What does happen if you execute "?mean" ?

R starts a built-in Web server that generates HTML help pages dynamically (and handles things like searches through the Web interface). It uses a randomly selected local port, which you should easily be able to figure out in the R process.

2. Are there any tex, man pages available?

(in /Library/Frameworks/R.framework/Resources/library/PACKAGE_NAME/ help_or_latex)

Nope. Just a binary database (*.rdb / *.rdx) in the help/ subdirectory.

You can ask for static HTML pages during package installation (from source), but most users will get the pre-built binaries from CRAN without any "readable" version of the help pages.

Best, Stefan

Stefan beat me to it... I switched back to 2.9 immediately, when I realized how much I adapted my workflow to this *amazing* bundle!

Why the R team introduced this (quite radical) change I do not know....

Daniel

On Nov 9, 2009, at 11:11 PM, Hans-Jörg Bibiko wrote:

On 09.11.2009, at 23:03, Stefan Evert wrote:

...

R starts a built-in Web server that generates HTML help pages dynamically (and handles things like searches through the Web interface). It uses a randomly selected local port, which you should easily be able to figure out in the R process.

OK.

...

...

2. Are there any tex, man pages available?

(in /Library/Frameworks/R.framework/Resources/library/PACKAGE_NAME/ help_or_latex)

Nope. Just a binary database (*.rdb / *.rdx) in the help/ subdirectory.

You can ask for static HTML pages during package installation (from source), but most users will get the pre-built binaries from CRAN without any "readable" version of the help pages.

I see. This and next week I've to travel a lot, ie much time for adapting hopefully.

--Hans

---

textmate mailing list textmate@lists.macromates.com http://lists.macromates.com/listinfo/textmate

Hey Hans,

awesome work, thanks a million!

I noted some small quirk: Previously invoking help on a keyword (CTRL-H) , say "recode" yielded (in my case) four packages that contain that function (Hmisc, car, depmix, memisc).

But now I get a list of packages -- for exampel 5 hits in the base package -- which as far as I can tell are unrelated to any for of "recode"

Best, Daniel

On 20 Nov 2009, at 23:14, Hans-Jörg Bibiko wrote:

Hi,

sorry for the delay but the daily work… ;)

I've just uploaded the new versions of R and Rdaemon to support the new R 2.10 help httpd server.

Could you please test them? Feedback is highly welcomed. Please note that the Package Manager wasn't updated yet and due the dynamic search a full text search is up to now not possible.

Furthermore I changed some tiny stuff. Rdaemon:

• execute selection and copy into pasteboard now makes usage of

write() or write.table() resp. for better pasting

• chooseCRANmirror now sets the repo persistently

R:

• if DIALOG2 is available the completion makes usage of it and

inserts '(|)' as snippet [use ⇥] (comments?)

• if no Rdaemon is running a first help request starts the httpd

help server in the background; if the server doesn't receive any queries for 10 minutes the server will be terminated automatically; this will have one side-effect: if you leave an Help window open for more than 10 mins any links will be broken due to the termination of the server → simply redo your help search

…and some tiny other stuff

Cheers,

--Hans

---

textmate mailing list textmate@lists.macromates.com http://lists.macromates.com/listinfo/textmate

On 21.11.2009, at 12:33, Daniel Stegmueller wrote:

Discard my last problem report. Logging out and in made it disappear...

Two things:

1) I forgot to mention that one has to rebuild the help index "R > Documentation > Create Command Help Index", which should be done automatically, but if you did a svn update without Reload Bundles the index won't be recreated.

2) Since there are no HTML/tex file stored on the hard disk I have to use "help.search('A_TERM')". Here one can discuss whether it would be nice to do it via fuzzy matching (agrep) which is the default, or not, or with the max distance of 1. I'll try to find a compromise. On the other hand I found it quite useful to have the chance to search in the entire help text, e.g. I'm looking for an example but I cannot remember the function name, but I can remember a line in the example. The only way would be to generate ALL HTML page and store them on the disk (>=13MB). I wrote such a routine but I do not know whether this would be helpful for other users as well.

Cheers, --Hans