Hi,
I use the R-TextMate Bundle and just recently run into problems when executing code via 'Run Selection in R'.
After some loop-iterations I get the following error message and have to start the calculations all over again:
/Library/Application Support/TextMate/Bundles/R.tmbundle/Support/tmR.rb:157:in `write': Invalid argument (Errno::EINVAL)
from /Library/Application Support/TextMate/Bundles/R.tmbundle/Support/tmR.rb:157:in `print'
from /Library/Application Support/TextMate/Bundles/R.tmbundle/Support/tmR.rb:157
from /Library/Application Support/TextMate/Bundles/R.tmbundle/Support/tmR.rb:144:in `each'
from /Library/Application Support/TextMate/Bundles/R.tmbundle/Support/tmR.rb:144
There seems to be no inherent problem with the code. Sometime the script runs till the end, sometimes I get these error-messages ...
Can someone explain these error messages to me?
thanks in advance!
> textmate2 wiki with ads?
> no thanks :)
No problem. :)
> Why do we need matewiki.com (http://matewiki.com) when there is already wiki.macromates.com (http://wiki.macromates.com)?
I posted the first 5 reasons that came to mind an email or two back:
> I want EVERYONE to have access to it, not just me which will lead to a broader and more updated resource. The
> macromates CSS theme is way too thin for my screen. I'm a control freak and want to be able to use it as a personal
> resource for TextMate bundle development. I'm going to put ads on it in an attempt to support my wild and not-yet-
> profitable personal projects. I'd rather figure ___ out myself then get access to someone else's ___.
BTW, yes I know everyone hates ads but I can either get a job and not work on TextMate & other personal projects or try to get my hustle on and do what I care about. I'll set up a donate link soon since people are so vocally against ads. ;)
PS If you are logged in to the Wiki the ad disappears. Registration is obviously free and simple. Happy? :)
--
Brandon Fryslie
Dear all,
I'm just playing with the latex bundle and TM2. Every time I do anything
like edit text or save TM2 freezes and takes up one core on my machine. I
this due to 'attr.save-on-deactivate' ?
my .tm_properties file looks as follows:
#######
PATH =
"/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin:/usr/texbin:/usr/bin/local:/Library/Frameworks/GTK+.framework/Resources/bin"
[ *.{R, c, cpp, m, pl, rb, html, xml}]
scopeAttributes = 'attr.save-on-deactivate'
######
Also quick 2nd question how do I get the compiled pdf to appear in a
separate window like TM1 instead of at the bottom of the window?
Cheers,
Paul
Hi All,
I think the tabulation settings were preserved across TM2 sessions,
but they are actually not on my MBP.
What are the rules about the tabulation settings? Is there some
specific action(s) to do so that soft tabs are always on?
Thanks,
Emmanuel.
Hi All,
I'm editing C files with TM2, and the auto indentation is often wrong,
and forces the cursor to move to an unexpected position when I start
writing a new line.
I'm looking for information/guide to:
* disable the whole auto-indentation feature
* configure the auto-indentation feature, if it's feasible
* cancel the last auto-indentation action: Cmd-Z removes both the
indentation and the typed character. In many others editors, Cmd-Z
cancels the automatic action, then a second hit remove the last typed
char. Is there an equivalent in TM2?
Thanks,
Emmanuel.
Is there a way to prevent the tabs from changing order?
I usually setup my workspace, with specific files in the first tabs (e.g.,
spec/test on cmd-1 and code on cmd-2). It seems that every once in awhile
it changes the order. Can I get TM2 to leave the tab order alone?
I'm trying to use Textmate2 with a big php project. I have made a huge commit in the git repository and while processing the commit Textmate starts to hang bringing the cpu at 80% and memory up to 1.25GB.
Looking at the Console this is what I've found
30/01/2012 16:28:23 TextMate[24252] TextMate(24252,0xa059f540) malloc: *** mmap(size=872415232) failed (error code=12)
*** error: can't allocate region
*** set a breakpoint in malloc_error_break to debug
30/01/2012 16:28:23 [0x0-0x1e91e9].com.macromates.TextMate.preview[24252] TextMate(24252,0xa059f540) malloc: *** mmap(size=872415232) failed (error code=12)
30/01/2012 16:28:23 [0x0-0x1e91e9].com.macromates.TextMate.preview[24252] *** error: can't allocate region
30/01/2012 16:28:23 [0x0-0x1e91e9].com.macromates.TextMate.preview[24252] *** set a breakpoint in malloc_error_break to debug
30/01/2012 16:28:23 TextMate[24252] Attempt to allocate 872415232 bytes for NS/CFData failed
30/01/2012 16:29:15 fseventsd[56] SLOWDOWN: client 0x10082ca00 (pid 24252) sleeping due to too many errors (num usleeps 23)
30/01/2012 16:31:43 fseventsd[56] SLOWDOWN: client 0x10082ca00 (pid 24252) sleeping due to too many errors (num usleeps 35)
30/01/2012 16:34:30 fseventsd[56] SLOWDOWN: client 0x10082ca00 (pid 24252) sleeping due to too many errors (num usleeps 47)
30/01/2012 16:36:35 fseventsd[56] SLOWDOWN: client 0x10082ca00 (pid 24252) sleeping due to too many errors (num usleeps 59)
I have Textmate2 Version 2.0 (9005) installed.
Patrick Polloni
Hi,
my sources live in a network file system behind some complex automount
scheme, let's say /foo/bar/workarea/user/project. Whenever I open some
directory there inside TM2, I get tons of messages like these in the
system log:
> 30.01.12 11:11:34 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .git failed
> 30.01.12 11:11:34 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .hg failed
> 30.01.12 11:11:35 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .p4config failed
> 30.01.12 11:11:35 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .svn failed
> 30.01.12 11:11:36 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .git failed
> 30.01.12 11:11:36 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .hg failed
> 30.01.12 11:11:37 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .p4config failed
> 30.01.12 11:11:37 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .svn failed
> 30.01.12 11:11:37 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .git failed
> 30.01.12 11:11:38 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .hg failed
> 30.01.12 11:11:38 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .p4config failed
> 30.01.12 11:11:39 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .svn failed
> 30.01.12 11:11:39 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .git failed
> 30.01.12 11:11:40 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .hg failed
> 30.01.12 11:11:40 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .p4config failed
> 30.01.12 11:11:40 automountd[3079] parse_entry: getmapent for map sfs.worldwide.somewhere.workarea, key .svn failed
> 30.01.12 11:11:41 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .git failed
> 30.01.12 11:11:41 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .hg failed
> 30.01.12 11:11:42 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .p4config failed
> 30.01.12 11:11:42 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.bar, key .svn failed
> 30.01.12 11:11:43 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .git failed
> 30.01.12 11:11:43 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .hg failed
> 30.01.12 11:11:43 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .p4config failed
> 30.01.12 11:11:44 automountd[3079] parse_entry: getmapent for map auto.sfs.sol2.foo, key .svn failed
It seems that TM2 tries finding the config files all the way to the root
directory for each file or directory it sees in the directory tree. I
know TM2 development is not yet at performance tuning, I just wanted to
point out that there is a probably low-hanging fruit here.
On a related note, we are using perforce, but tend to have config files
named .perforce. Is the file name TM is looking for configurable (yet)?
Christopher
In my .tm_properties file I have the following:
[ text.html ]
tabSize = 4
[ source ]
tabSize = 2
If I edit a javascript file then I correctly see tabs taking up two spaces.
If I edit an html file then tabs are seen as four spaces. If I embed
javascript in the html file the tabs within the javascript code are now
taking up 4 spaces. Pressing Cmd-Shift-P shows the following scope:
text.html.basic
source.embedded.js.html
attr.rev-path.xhtml.html.about.misc.views.templates.WEB-INF.webapp.myproject.curt.Users
attr.os-version.10.8.6
attr.scm.svn
attr.scm.status.H
So it appears that the text.html.basic scope is taking precedence over the
source.embedded.js.html scope. Is there a way to have this behave in the
expected manner so that my javascript settings are honored when embedded in
html?
Thanks,
Curt
TM2 (at least in Ruby) seems to treat words as what the compiler sees,
rather than what the developer sees.
For example, in the following ruby code...
def some_variable
@so
end
With the cursor after '@so', when you hit escape, nothing happens. Given
that (from the developers point of view), the accessor method for
'some_variable' and the instance variable name for the same are the same
things, shouldn't that complete properly?
The same goes with symbols. This code has a similar problem...
before_filter :some_method
def some<cursor>
So far, this has probably been one of the frustrating issues in TM2 for me.
Is this going to be fixed/changed?
Hi,
I've got favorites for “lib” and “LIB” set, pointing to two different
locations. That works fine when I switch the file browser to the
favorites (with the heart shape icon), but in the command-shift-O
dialog, I only see one of them.
I know I can rename the favorites and probably will do that, but
shouldn't these two views into the favorites list behave similarly, at
least when I have no subfolders?
Christopher
Is it possible to change the key combination used for completion? Escape is a very far reach given how commonly I use completion. If not, I'll just have to resort to keyremap4macbook.
Thanks,
Connor
On Jan 30, 2012, at 6:00 AM, textmate-request(a)lists.macromates.com wrote:
> Date: Mon, 30 Jan 2012 11:10:32 +0000
> From: "ld(a)ldaley.com" <ld(a)ldaley.com>
> To: "TextMate users" <textmate(a)lists.macromates.com>
> Subject: [TxMt] Re: TM Bundle for Java - Improved Java Properties
> grammar
> Message-ID: <6ACE76A4-AF40-4943-AA54-0805C2990BDA(a)ldaley.com>
> Content-Type: text/plain; format=flowed
>
> On 29 Jan 2012, at 6:15, John Joyce wrote:
>
>> Hello List,
>> After ages of consternation with viewing Java Properties files that
>> had multiline values in KVPs not properly syntax colored, I have
>> created a better language grammar for Java Properties.
>>
>> It does several things...
>> First it captures comments of course.
>> Then single-line KVPs
>> Then KVPs with dangling spaces or tabs between the key and the = sign.
>> (These stand out awesomely during fast scrolling to find where they
>> are and fix them)
>> Then multi-line-value KVPs.
>>
>> One can never be 100% sure of these things, but I feel fairly
>> confident with it because the multi-line-value KVPs I have dealt with
>> are actually HTML emails stored in KVPs. :)
>>
>> By the way, it looks best in Sunburst them.
>>
>> Please consider it for inclusion.
>
> If you submit a pull request for
> https://github.com/textmate/java.tmbundle I'll happily include it.
>
> Thanks!
>
> LD.
Submitted pull request.
Thanks,
John Joyce
Oddity:
My cursor ist placed after the "h" in the tag that should become "h1". The tag has been produced using "Insert open/close tag with current word" but without anything selected. I just typed the "h".
when typing the "1" the snippet inserts weird stuff underneath, where the text "Headline 1" is (see second screenshot).
could someone with a little more expertise have a look at the snippet and possibly fix it? I have no idea why that happens.
Thanks,
Thomas
Hello List,
After ages of consternation with viewing Java Properties files that had multiline values in KVPs not properly syntax colored, I have created a better language grammar for Java Properties.
It does several things...
First it captures comments of course.
Then single-line KVPs
Then KVPs with dangling spaces or tabs between the key and the = sign. (These stand out awesomely during fast scrolling to find where they are and fix them)
Then multi-line-value KVPs.
One can never be 100% sure of these things, but I feel fairly confident with it because the multi-line-value KVPs I have dealt with are actually HTML emails stored in KVPs. :)
By the way, it looks best in Sunburst them.
Please consider it for inclusion.
zipped for your unzipping pleasure.
Thanks,
John Joyce
Sorry if this has been covered, but I couldn’t find a mention of it.
Is there a “Filter Through Command” in TextMate 2 (8998)? Was it re-implimented as a bundle item or something? I can’t find it and the old shortcut doesn’t seem to bring it up.
--
Rob McBroom
<http://www.skurfer.com/>
Dear Sir or Madam:
http://jpfly.com/woaint.php?fuducampID=38
Thu, 26 Jan 2012 8:40:39
______________
" At the commencement of winter, Artichoke plants should be protected against frost, which sometimes destroys them in our climate" (c) Kailene vtk5945
On 24/01/2012, at 20.54, Dirk Günther wrote:
> I use the LaTeX grammar and I have defined my own shortverb
> command [...] What should I do that |c| in \tabular{...} is not defined as
> verbatim?
As Allan suggested, I think that you need to have a rule in the LaTeX
grammar that will match the column specification part of the tabular
environment. Instead of writing a completely new rule, I tried changing the
relevant existing rule, copied below. The only difference is in the begin
match. I also tried adding your rule for scoping anything between pipes as
verbatim, and this change to the tabular rules seems to do what you want -
anything in the column spec is not scoped as verbatim. Hope that helps!
-Daniel Grady
{ begin = '(?x)
(?:\s*) # Optional whitespace
((\\)begin) # Marker - Function
(\{) # Open Bracket
(array|tabular[xy*]?)
(\}) # Close Bracket
(\s*\{.*\})? # Column specification
(\s*\n)? # Match to end of line absent of content
';
captures = {
1 = { name = 'support.function.be.latex'; };
2 = { name = 'punctuation.definition.function.latex'; };
3 = { name = 'punctuation.definition.arguments.begin.latex'; };
4 = { name = 'variable.parameter.function.latex'; };
5 = { name = 'punctuation.definition.arguments.end.latex'; };
6 = { name = 'variable.columnspec.tabular.latex'; };
};
contentName = 'meta.data.environment.tabular.latex';
end = '(?x)
(?:\s*) # Optional whitespace
((\\)end) # Marker - Function
(\{) # Open Bracket
(\4) # Previous capture from begin
(\}) # Close Bracket
(?:\s*\n)? # Match to end of line absent of content
';
name = 'meta.function.environment.tabular.latex';
patterns = (
{ match = '\\';
name = 'punctuation.definition.table.row.latex';
},
{ begin = '(?:^|(?<=\\\\))(?!\\\\|\s*\\end\{(?:tabular|array))';
end = '(?=\\\\|\s*\\end\{(?:tabular|array))';
name = 'meta.row.environment.tabular.latex';
patterns = (
{ match = '&';
name = 'punctuation.definition.table.cell.latex';
},
{ begin = '(?:^|(?<=&))((?!&|\\\\|$))';
end = '(?=&|\\\\|\s*\\end\{(?:tabular|array))';
name = 'meta.cell.environment.tabular.latex';
patterns = ( { include = '$base'; } );
},
{ include = '$base'; },
);
},
{ include = '$base'; },
);
}
Setup:
Some text here {Some text here {some more text here}} Some text here
1. Place cursor after the first } but before the second }
2. Press Ctrl-Up.
My understanding is that the cursor should jump to just after the first {,
instead it lands just before the second {.
Cheers,
David.
The default settings for TextMate 2 have soft wrapping turned on for text
scopes. When I'm editing a LaTeX document (text.tex.latex), this is fine,
except for tables. Inside of tabular environments I've usually got the
columns lined up so it's easier to read the source file, but when soft
wrapping is turned on this just becomes a huge mess. I thought it would be
possible to turn soft wrap off inside tables through the .tm_properties
file, and I tried adding this rule
[ meta.function.environment.tabular.latex ]
softWrap = false
but this doesn't seem to do anything. Do scope selectors work the same way
in .tm_properties as elsewhere? Does anyone have a suggestion for this,
short of altering the LaTeX bundle? Thanks for any help!
-Daniel Grady
The scope of a tabular environment is
text.tex.latex
meta.function.environment.general.latex
meta.function.environment.tabular.latex
meta.data.environment.tabular.latex
meta.row.environment.tabular.latex
meta.cell.environment.tabular.latex
attr.rev-path.tex.example.Documents.dgrady.Users
attr.os-version.10.6.8
Here is a short example that shows what I'm trying to accomplish:
\documentclass{article}
\begin{document}
This line should soft-wrap: Lorem ipsum dolor sit amet, consectetur
adipisicing elit, sed do eiusmod tempor.
\begin{table}
\begin{tabular}{c}
Lines inside the tabular environment should not soft wrap: Lorem ipsum
dolor sit amet, consectetur.
\end{tabular}
\caption{The caption should soft wrap: Lorem ipsum dolor sit amet,
consectetur adipisicing elit, sed do eiusmod.}
\end{table}
\end{document}
Hello,
I use the LaTeX grammar and I have defined my own shortverb command: | so
that I can write |foo| and have everything enclosed in | and | treated the
same as \verb|foo|. So I'd like to mark the contents between | and | as
verbatim, but not in the tabular definition such as
\begin{tabular}{|c|c|c|c}
My attempt
{ begin = '\|';
beginCaptures = { 0 = { name = 'punctuation.definition.verb.latex'; }; };
end = '\|';
endCaptures = { 0 = { name = 'punctuation.definition.verb.latex'; }; };
name = 'markup.raw.verb.latex';
},
works for requirement #1, but not for #2 (the tabular case). What should I
do that |c| in \tabular{...} is not defined as verbatim?
Thanks
Dirk
I want to access a set of 5 files which are not all in one folder. I can get
all 5 into the file browser and work with them. Good so far. Now I want to
save the set so that next month, when I need another change, I can open the
same set of files without visiting each location individually. What's a good
way to open the set of files with one action? I don't have control over
where the original files are stored. They are part of Apache2.
I thought of making a folder with a link to each file, which is ok for 5
files, but not for a larger group.
I think that .tm_properties might be an approach, perhaps by setting up a
folder with a list of includes, but that seems kind of clumsy. I suspect
this isn't hard, but I'm still finding my way around TM2. Any hints,
guidelines, whatever?
--Lewy
--
View this message in context: http://old.nabble.com/TM2-saving-pointers-to-set-of-files-tp33185861p331858…
Sent from the textmate users mailing list archive at Nabble.com.
TextMate 2 seems very promising feature wise, but one of my biggest complaints with version 1 was the slow performance of extremely large projects (20k+ files), particularly when using the "Go to file" command. I was very disappointed then to see that the v2 alpha is even slower in this regard, unusably so. When I try to type a file name, one letter appears at a time with up to ten seconds in between, and this is on a quad core i7 with an SSD. Even using the `find` command from the terminal is faster, and it walks the entire directory tree every time.
I understand that this is an alpha, and that performance was one of the things that was left for later tuning. But I'm afraid that more than tuning is needed in this case, although I could be wrong. I would simply ask that the performance of large projects be made more of a priority than it currently is.
Thanks,
Connor McKay
Hi,
Does anyone know what might prevent the results of a an executed command to show up in the R Console? Every time I type a command the line ends up cleared and no output is displayed. It seems that the command is getting executed as I can can see the result if I open up r_out. It even shows up in the history.
Here is a simple example of what I get on newly installed Rdaemon after launching from TM_Rdaemon.app and executing:
> ls()
After hitting return the console gets cleared to:
>
In r_out I get:
WARNING: unknown option '--TMRdaemon'
ARGUMENT '2' __ignored__
> source("/Users/linaskk/Library/Application Support/Rdaemon/daemon/start.r")
> sink('/Users/linaskk/Library/Application Support/Rdaemon/r_tmp');cat(geterr
<ary/Application Support/Rdaemon/r_tmp');cat(geterrm essage());sink(file=NULL)
> sink('/Users/linaskk/Library/Application Support/Rdaemon/r_tmp');args(ls)
> sink(file=NULL)
> sink('/Users/linaskk/Library/Application Support/Rdaemon/r_tmp');cat(gsub('
<ary/Application Support/Rdaemon/r_tmp');cat(gsub('. *?/library/(.*?)/.*','\\1
<aemon/r_tmp');cat(gsub('.*?/library/(.*?)/.*','\\1' ,as.vector(help('ls',try.
<?/library/(.*?)/.*','\\1',as.vector(help('ls',try.a ll.packages=F)),perl=T),s
<as.vector(help('ls',try.all.packages=F)),perl=T),se p='\n')
> sink(file=NULL)
> ls()
[1] "jpeg" "pdf" "png"
>
I've used Rdaemon before on other systems running the same OS X 10.6.8 without a problem, so I am not sure why it doesn't work on this particular system. The same thing happens in both TM1 and TM2.
-Liisi
