Thank you. Now makes sense why they aren’t consistently created. Sometimes I do edit with files on my hard drive and then sync them to the flash card. (I had started to edit on the card to avoid the added step of ensuring syncing).
I’ll think about revising my workflow, particularly since others have to look at my flash drive from time to time. I’m still adjusting to working on Windows and some other complications of the project.
Greg
>
> I think the the technical term is "AppleDouble" files.
>
> https://en.wikipedia.org/wiki/AppleSingle_and_AppleDouble_formats <https://en.wikipedia.org/wiki/AppleSingle_and_AppleDouble_formats>
>
> They get created anywhere the underlying filesystem doesn't provide resource forks, or extended attributes (which is mostly anything not HFS+). Since TextMate uses those extended attributes to stash some relevant (like the cursor location), this causes them to appear more frequently.
>
> If you're using git, you should definitely setup a global exclude file with a few of those macOS-specific patterns. See https://www.gitignore.io/ <https://www.gitignore.io/> for some helpful ones.
>
> Another way would be to edit your project on an HFS+ drive* ? you'll get those pesky .DS_Store anyway, but at least the ._ files shouldn't appear.
> * I'm only suspecting you do, since AppleDouble files are unneeded on HFS+.
>
> Regards,
> Etienne Samson
> ————
> TextMate is using extended attributes, which were introduced in 10.4 (I think) and which are also stored in doubles on filesystems that don't support them natively.
>
> -Martin
Matt
Thanks for the response. I never noticed these files before. Are they created by TextMate, because I don’t see them anywhere else. Or are they created when load the thumb drive on the Windows machine? From the linked discussion in another response I gather it’s TextMate storing some other information, but they don’t always seem to be created. Does TextMate need them or use them? I ask because maybe I don’t want to delete them.
I see them also using Transmit (an ftp app).
Greg
> On Sep 24, 2017, at 5:00 AM, textmate-request(a)lists.macromates.com wrote:
>
> This is not a TextMate question. It's just how Macs work. Those are resource forks, which Windows can't comprehend. See for example https://superuser.com/questions/851373/hide-macs-dot-underscore-files-on-a-… <https://superuser.com/questions/851373/hide-macs-dot-underscore-files-on-a-…>
>
> m.
>
>> On Sep 23, 2017, at 9:48 AM, Greg <web(a)web.knobby.ws <mailto:web@web.knobby.ws>> wrote:
>>
>> Is TextMate adding files with a leading period and underscore in front, for example, ?._styles.css"?
>>
>> I?m using TextMate for a class that mostly Windows and all these extra files show up when I?m on Windows as well as other apps which show invisible files. This doubles the apparent number of files.
Is TextMate adding files with a leading period and underscore in front, for example, “._styles.css"?
I’m using TextMate for a class that mostly Windows and all these extra files show up when I’m on Windows as well as other apps which show invisible files. This doubles the apparent number of files.
Greg
I know that you can rig TextMate 2 to trim trailing space on save by
editing the Text -> Converting / Stripping -> Trim... bundle by
changing the semantic class (or creating a whole new bundle). It works
well, but I was wondering if there are any plans to add this as an
option in the preferences (with a checkbox, etc.) for convenience?
It's a rather common one to have that's found on most other editors
I've seen, and it would be super cool to not have to mess with the
bundles for such a basic thing.
Kind regards,
- g
Hi all,
This might be a stupid question: Is there a way in TextMate 2 to cut, and then uncut(paste), the current line with 2 keystrokes? I.e, similar to the nano commands ‘Ctrl-k’ and ‘Ctrl-u’? As opposed to: select the current line, cut, then paste.
Thanks in advance.
Promote away. Thank you.
> On Sep 18, 2017, at 5:00 AM, textmate-request(a)lists.macromates.com wrote:
>
> Subject: [TxMt] Linting functionality for TextMate
> Message-ID: <D0C4F4EF-1C59-4999-A842-E8646270D387(a)me.com <mailto:D0C4F4EF-1C59-4999-A842-E8646270D387@me.com>>
> Content-Type: text/plain; charset=utf-8
>
> Apologies for the self-promotion but I thought this may be interesting to some of you. I?ve seen Atom and Sublime Text linting plugins and I was wanting something similar for TextMate so wrote my own plugin:
>
> https://github.com/MikeMcQuaid/Linter.tmbundle <https://github.com/MikeMcQuaid/Linter.tmbundle>
>
> Give me a shout if you like this or have feedback. If there?s enough interest I?d like to try and submit this to be an official TextMate plugin in future.
>
> Thanks!
>
> Mike McQuaid
> http://mikemcquaid.com <http://mikemcquaid.com/>
Works here, as they say.
--
Marc Wilson
msw(a)cox.net
On Sun, Sep 17, 2017, at 11:09 AM, George McGinley Smith wrote:
> Yes, it's supposed to pop out the changes in the commit. Do you have
> customisations in your Git bundle? I also had the same issue and
> reverting to the default bundle fixed the issue for me.>
> On Sun, 17 Sep 2017 at 19:56 Matt Neuburg <matt(a)tidbits.com> wrote:
>> When you do Command-Y and Log, and the commits are listed with plus
>> buttons at their left, are those plus buttons supposed to _do_
>> something, like show the diff? They don't.>>
>> m.
>>
>> --
>> matt neuburg, phd = http://www.apeth.net/matt/
>> pantes anthropoi tou eidenai oregontai phusei
>> Programming iOS 10! http://shop.oreilly.com/product/0636920055235.do>> iOS 10 Fundamentals! http://shop.oreilly.com/product/0636920055211.do>> RubyFrontier! http://www.apeth.com/RubyFrontierDocs/default.html
>>
>>
>> _______________________________________________
>> textmate mailing list
>> textmate(a)lists.macromates.com
>> http://lists.macromates.com/listinfo/textmate
>
> _________________________________________________
> textmate mailing list
> textmate(a)lists.macromates.com
> http://lists.macromates.com/listinfo/textmate
I'm getting the following error from the LaTeX documentation command:
Traceback (most recent call last):
File "/Users/tim/Library/Application Support/TextMate/Managed/Bundles/LaTeX.tmbundle/Support/bin/texdoc.py", line 223, in <module>
shell=True).strip()
File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/subprocess.py", line 573, in check_output
raise CalledProcessError(retcode, cmd, output=output)
subprocess.CalledProcessError: Command 'texdoc %' returned non-zero exit status 1
The command source code is:
#!/usr/bin/env bash
[[ -f "${TM_SUPPORT_PATH}/lib/bash_init.sh" ]] && \
. "${TM_SUPPORT_PATH}/lib/bash_init.sh"
. "${TM_SUPPORT_PATH}/lib/webpreview.sh"
html_header 'Package Documentation'
echo '<p><small>Hint: Cursor on package name shows doc for package.
Otherwise show package doc browser.</small></p>'
texdoc.py
My environment is
OS high Sierra beta
Has anyone else tried out the High Sierra beta and experienced performance
issues with the Quick Look plugin?
When I put a folder full of .h and .m files into Icon view in Finder, Quick
Look started taking up 100% CPU. Seems like there must be some infinite
loop, but I haven't investigated the root cause yet.
Jacob
when I type a quote or a left paren when using the fortran bundle, it automatically inserts a matching quote or paren and leaves the cursor between it and the one i typed myself.
i can see that some users will love this and find it very useful. if just find it annoying and a nuisance. i notice that this “helpful” behavior is not present in Plain Text, but I’d like to use the fotran bundle, just without the automatic addition of matching quotes and parens.
there was nothing obvious in the details of the fortran bundle, so I’m asking for help on how to manage this.
thanks
bill
I recently upgraded to State 15 and am using an older Stata bundle that worked with Stata 14. I am trying to understand how to make Textmate 2 run Stata 15 from within the Stata bundle, but am having trouble figuring it out.
Hello
Just wondering about the likelyhood of getting
https://github.com/textmate/git.tmbundle/pull/36 merged? I have been
running with local delta updates since creating that PR but have recently
run into issues which removing my customisations has fixed, rather than
having to go through and re-update the bundle, if we could get this merged
that would likely solve 80% of my required changes as it's one of my more
frequently used git commands.
Thanks,
George
I think TextMate’s bundle index should include a RuboCop bundle.
Disclaimer: I wrote a RuboCop bundle myself, and there are at least
three more that I know about:
1) https://github.com/mrdougal/textmate2-rubocop
2) https://github.com/goyox86/rubocop-tmbundle
3) https://github.com/fazibear/Rubocop.tmbundle
4) My bundle: https://github.com/noniq/RuboCop.tmbundle
1) and 2) seem stale (last commit in Nov resp. Dec 2014). 3) and 4) are
actively maintained.
One of the differences between 3) and 4) is that 3) supports different
gutter marks for different severities, while 4) only uses a generic
warning mark. On the other hand, 3) seems to only ever work on the
current file, while 4) uses the filebrowser selection (if any, similar
to the Git bundle for example).
In
https://github.com/textmate/ruby.tmbundle/pull/99#issuecomment-296400597
ff. there was discussion if the Ruby’s bundle already existing “Validate
Syntax” command should be changed / extended to use RuboCop (if present)
– this could be an alternative to adding an “official” RuboCop bundle.
Happy to hear your opinions!
Stefan
Looking at the code, TextMate appears to start with a list taken from
“Find in Folder Globs.default”, which doesn’t exist, so it falls
back to hard-coded defaults.
I find myself adding the same patterns to almost every project so I’d
like to change the defaults.
Is there a way to set that via `.tm_properties` or do I just have to add
a key named “default” to `com.macromates.TextMate.plist` manually?
Thanks.
--
Rob McBroom
http://www.skurfer.com/
Hi,
This week I installed Textmate 2.0 beta, and I am still learning a lot. One of the things I found myself wishing for was the ability to change some features of the themes I found most to my liking. Google got me to pages such as this (https://manual.macromates.com/en/themes <https://manual.macromates.com/en/themes>), referencing "Preferences → Fonts & Colors”, which I can’t find in Textmate; perhaps the manual does not apply to the version I installed?
Is there any (GUI-based much preferred) way of building a theme (modifying an existing one) in my installation? If not, should I perhaps uninstall Textmate and install an earlier version? If so, which?
Thank you.
Nicolae
I know it’s hard to keep old webpages up but it would be helpful if a banner could be put up on the TM 1 pages with a link to TextMate 2 info. The linked page looks “official,” so even more important. I’m far from a power user and am discouraged from exploring TM when I run into these pages without knowing which version they apply to.
Near top of page might be something like: “This page is about TextMate version 1. For TextMate 2 documents check <here> For the status TextMate 1 and 2 read< here>.” Probably should be stuck to top of page in case the link puts someone part way down the page.
> On Jul 22, 2017, at 6:00 AM, textmate-request(a)lists.macromates.com wrote:
>
>> (https://manual.macromates.com/en/themes <https://manual.macromates.com/en/themes> <https://manual.macromates.com/en/themes <https://manual.macromates.com/en/themes>>), referencing "Preferences ? Fonts & Colors?, which I can?t find in Textmate; perhaps the manual does not apply to the version I installed?
>
> That link applies to TextMate 1.
Hi there,
I upload generally the new versions of lftp with ‘brew'. But it is able to install this morning only the version 4.7.7
(Command in a the terminal of a Mac : 'brew install lftp’ - with link problems but it is easy to fix it)
Do you know of this new version is now included in ‘brew' and how I can find it?
Tkx
--
Bernd J.
I recently started to notice that the HTML output window often (not
always) no longer automatically scrolls down when new output is appended
(eg. from a build task thats running for several seconds).
Anyone else noticing this?
Stefan
Hello all,
I'm looking to fix up the old TextMate Python Jedi completion bundle.
The docstring viewer command returns this error.
Setting or variable 'fontName' not found
Traceback (most recent call last):
File "Docstrings for word", line 10, in <module>
completion.show_docstrings()
File "/Users/myaccount/Library/Application Support/TextMate/Pristine Copy/Bundles/python-jedi.tmbundle/Support/completion.py", line 132, in show_docstrings
font_name = subprocess.check_output([tm_query, '--setting', 'fontName']).rstrip() or "Menlo-Regular"
File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/subprocess.py", line 573, in check_output
raise CalledProcessError(retcode, cmd, output=output)
subprocess.CalledProcessError: Command '['/Applications/TextMate.app/Contents/Resources/tm_query', '--setting', 'fontName']' returned non-zero exit status 69
This is because the CLI command says no such key exists.
/Applications/TextMate.app/Contents/Resources/tm_query --setting fontName
Setting or variable 'fontName' not found
Is there a modern alternative to fontName?
Or, even better, a reference for that and other keys?
thanks!
JJ
Hi Josh,
(I apologize for my replies to be attached to the wrong places in the thread. I have just figured out how to turn off the digest format. Please bear with me until the changes take effect.)
> Thanks for pointing this out! I agree that in 2017 we should
> be taking security seriously. I'd love to learn though, what
> are some good ways to set up vetting and security policies?
Off the top of my head:
1. Create an organization on GitHub, and move the repo there. Or maybe get another GitHub organization (TextMate?) to adopt it. This ensures that in the long run, more than one single person has the power to grant or revoke _Owner_ and _Maintainer_ privileges.
2. You have already mentioned a very effective measure: make it a policy to not allow `sha256 :no_check` without good reason. For TextMate bundles, this means in practice: always pin your Cask to a release; in cases where upstream does not issue releases or tags, pin the Cask to the latest commit instead; do not point to nightlies, or to upstream `head`; in case users really want it, create a separate tap for nightlies within the same GitHub organization.
3. On GitHub, enable forced reviews for any commit on the master branch; also, tick the box that says this actually covers anyone’s commits, including the owner’s. Configure the GitHub repo so any single maintainer has veto rights against merging. Make regular audits to ensure these setting are still in place.
4. Have two maintainers at the very least, and grant them the _Maintainer_ privilege to make #3 technically possible.
5. In the `url` stanza, do not allow domains where the maintainer can’t reliably determine who has control over the domain. For example, if the author of `Foomatic.tmbundle` offers their releases only via a Dropbox link, maintainers will not be able to keep track of authenticity as updates get released.
6. Do not be afraid to refuse a Cask if it fails to meet those standards for whatever reason.
7. On the GitHub organization level, grant the _Owner_ privilege to at least one more maintainer. Not only does this help keep the organization alive in case the original owner fails to be responsive, or fails to fulfill their duties, it also allows for quick reaction in case another owner’s account be ever compromised.
8. Speaking of compromised: enforce that anyone with _Maintainer_ or _Owner_ status have two-factor authentication enabled for their GitHub accounts. Make regular audits to ensure everyone continues to comply.
9. Check the contributions of anyone before you grant them _Maintainer_ status on the GitHub organization. Ensure that a person is trustworthy before you grant them _Owner_ status on the GitHub organization.
10. When creating or reviewing a Cask, make sure it contains a line that says e. g.,
# downloads.example.com was verified to be a canonical download
# site at the time it was first introduced to this Cask
On each update, double-check whether the `url` stanza still points to the same domain name as the one in the comment.
11. On each update, download the payload from the given source, and verify the updated SHA256 matches the downloaded file. (It’s OK to automate this.) When generating the hash, never use plain HTTP to download the file if it’s also available via HTTPS from the same source.
12. Use common sense before you grant anyone commit access or other privileges.
Regards!
Claudia
Hi,
> Homebrew wants to own /usr/local as if nothing else is there.
That’s what Homebrew _used_ to do. These times are long gone.
(FULL DISCLOSURE for anyone who’s late to the party: Caskroom member here, with a few ties to the Homebrew project.)
I have also used MacPorts, and I see how you prefer the `/opt` approach. Homebrew has come a long way to play as nice with everyone as MacPorts does today. Homebrew now lives exclusively in `/usr/local/Homebrew`. It installs packages into `../Cellar` and `../Caskroom`, and occasionally into `/opt` if it makes more sense. Anything else in `/usr/local` is free for anyone. I believe that for a package manager, this behavior is adequately modest.
If you’re unsatisfied with the fact that Homebrew puts individual things into `/usr/local/bin` and under `/usr/local/lib`, then you might want to consider that the FHS has made them for exactly that purpose. Package managers on most Unix-ish OSes keep some stuff inside `/usr/local/bin` and `/usr/local/lib` because that’s what those directories were made for.
With `brew link`, Homebrew invites all and any software to make use of its infrastructure, such as creating and maintaining symlinks for you so your software won’t have to. That’s purely optional though; anyone is free to completely ignore Homebrew, and have their software do its thing anywhere at their own discretion, and that includes `/usr/local`.
> Many package managers do not really have strong vetting
> or security policies. Homebrew is one of those.
This is actually a quite important point you’ve made; I wish more people were aware of this!
You are mostly correct, The Homebrew project is one of those package managers for a reason. Homebrew doesn’t want to be a software discovery service. Instead, it wants Homebrew users to make their own informed decisions as to whether they want to install a particular package or not.
If `foobinator` is a known spyware, but at the same time notable and popular enough that users want a `foobinator` formula to be in Homebrew, then in Homebrew that formula be.
The bottom line for users is: Do your homework before you decide to trust a particular software, with or without Homebrew. Homebrew wants that decision to be up to you, and only you.
That said, Homebrew does have vetting processes in places. Not only does it support GnuPG signatures, it also enforces SHA256 hash verification for all formulas who include such a hash, which most formulas do. Whenever a formula needs to be updated, the tap maintainer double-checks that it still points to the same trusted upstream domain. The maintainer also downloads the payload from that trusted source, and verifies the updated SHA256 matches the downloaded file. This is not to make any guarantees that the software is safe to use. It is for you to be sure that you’re installing the exact same thing the tap maintainer has personally downloaded from an official source of the upstream package.
Speaking of trust, my impression so far is that contributors are vetted quite carefully before they are granted commit access. Even when such access is granted, the project’s GitHub repos enforce reviews before anything gets merged into master branches, which are also the branches that get released to the public. Recently, the Caskroom organization has introduced a policy that requires every maintainer to use 2FA for their GitHub account.
Of course, none of all those measures will guarantee 100 % security. Nor will any of it withstand any sufficiently sophisticated attack. But it’s enough to make me personally feel very safe using Homebrew and all its official taps. (CAUTION: as I have already stressed a number of times, the `claui/textmate` tap is NOT an official Homebrew tap, and will probably never be.)
> I am just not sure unmanaged homebrew github repos is the thing.
I can’t think of a reason for anyone to use an unmanaged tap. It absolutely needs to be properly maintained.
Until some individual or organization is going to step up and adopt the tap, or we as a community figure out that a Homebrew tap might not fit our bill at all, I’ll gladly continue to help maintain it in the meantime as my time allows. Needless to say, I’m committed to the same standards and degree of diligence as I’d expect from any maintainer of a source from which I download and install executable software.
What the repo needs most right now is more contributors and more maintainers. It needs people to add content so it becomes actually useful, and make people try it out. More maintainers means that we’d have forced reviews enabled for the entire master branch similar to the official taps, including veto rights for every maintainer. This is very important for security; as of right now, the repo is not safe to use because it requires users to trust me personally. Being a sole maintainer means nobody can guarantee that every commit, including my own, gets actually reviewed.
> I would like to hear more before people just jump on the convenience.
I fully agree, and I’m looking forward to discuss this further.
Regards,
Claudia
> ------------------------------
>
> Message: 2
> Date: Fri, 7 Jul 2017 16:19:42 -1000
> From: ???????
> To: TextMate users
> Subject: [TxMt] Re: Bundle Management improvements?
> Message-ID: <7CC8A118-F7C5-4318-A814-54636955B8B2(a)gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> I am actually trying to avoid using homebrew.
> There are lots of great things in homebrew but a few flaws are definitely the assumptions
> about the path and also the total lack of vetting.
> Homebrew wants to own /usr/local as if nothing else is there.
> Mac Ports on the other hand is a bit more careful to use an uncommon path in /opt and that
> is pretty valuable to me.
>
> Many package managers do not really have strong vetting or security policies. Homebrew
> is one of those.
> Mac Ports is also not great at that.
>
> It would be great if there were more guarantees than just whatever is on github and visible
> to homebrew.
>
> A directory of sorts to work with would be great but I feel like this is clearly a modern
> and missing feature, I am just not sure unmanaged homebrew github repos is the thing.
> I would like to hear more before people just jump on the convenience.
>
>
> > On Jul 7, 2017, at 12:02, Lewis Overton >
> wrote:
> >
> > I use Textmate and Brew. Linking them sounds good.
> >
> > On Tue, Jul 4, 2017 at 11:03 AM, Ryan Fitzer >
> wrote:
> > > To install a tmbundle, you?d run:
> > >
> > > $ brew tm install arduino
> > > If more than three people find this useful, I?ll be happy to piece together a small PoC.
> >
> > This would extremely useful. I use Homebrew for so much already. Having it for bundles
> would a huge benefit.
> >
> > Ryan
> >
> > On Jul 4, 2017, at 1:51 AM, Claudia Pellegrino >
> wrote:
> >
> > Hi,
> >
> > The features that stand out to me immediately are: [?] CLI searching and installing.
> > Caskroom member here. With our recent integration into Homebrew last year, I think
> Homebrew now brings to the table a big part of what would be needed to manage TM bundles
> via CLI today.
> >
> > In fact, anyone is free to create an (unofficial) Homebrew tap for TextMate bundles.
> A tap is just a GitHub repo, e. g. github.com/textmate/homebrew-bundles .
> It would contain a number of package files, which essentially point to the individual
> download URLs of the tmbundles.
> >
> > This is what the user would do once to tap into the repo:
> >
> > $ brew tap textmate/bundles
> > To search for plugins, you?d run:
> >
> > $ brew tm search arduino
> > ==> Exact Match
> > tm-arduino
> > To install a tmbundle, you?d run:
> >
> > $ brew tm install arduino
> > If more than three people find this useful, I?ll be happy to piece together a small PoC.
> >
> > Regards,
> > Claudia
> >
> > _______________________________________________
> > textmate mailing list
> > textmate(a)lists.macromates.com
> > http://lists.macromates.com/listinfo/textmate
> >
> > _______________________________________________
> > textmate mailing list
> > textmate(a)lists.macromates.com
> > http://lists.macromates.com/listinfo/textmate
> >
> > _______________________________________________
> > textmate mailing list
> > textmate(a)lists.macromates.com
> > http://lists.macromates.com/listinfo/textmate
>
Hi,
The features that stand out to me immediately are: […] CLI searching and installing.
Caskroom member here. With our recent integration into Homebrew last year, I think Homebrew now brings to the table a big part of what would be needed to manage TM bundles via CLI today.
In fact, anyone is free to create an (unofficial) Homebrew tap for TextMate bundles. A tap is just a GitHub repo, e. g. github.com/textmate/homebrew-bundles. It would contain a number of package files, which essentially point to the individual download URLs of the tmbundles.
This is what the user would do once to tap into the repo:
$ brew tap textmate/bundles
To search for plugins, you’d run:
$ brew tm search arduino
==> Exact Match
tm-arduino
To install a tmbundle, you’d run:
$ brew tm install arduino
If more than three people find this useful, I’ll be happy to piece together a small PoC.
Regards,
Claudia
Hello じょいすじょん,
I have pieced together a small proof of concept and put it online for anyone to try: https://github.com/claui/homebrew-textmate
Feedback and pull requests welcome.
Regards!
Claudia
HI!
I’ve installed JSLintMate and am trying to use it on the JS for my Rails project. I’ve tried to turn off complaints about “$” and “document” being undefined, but can’t seem to do so. I’ve tried with a “.jshintrc” resource file, and also with inline configuration. My inline configuration currently looks like:
/* jshint undef: false, unused: false */
// global $, document
…
I read in the docs that single-line comments work as well as open-close comments, so I tried both formats.
Not only did it not turn off all messages regarding undefined and unused variables, but it also did not turn them off for “$” and “document”. Does anyone have an idea why?
In case it helps, the messages look like:
JSHint found problems:
‘$’ is not defined. line 3
$(document).ready(function() {
‘document’ is not defined. line 3
$(document).ready(function() {
…
I have several desktops ("Spaces" in Apple marketing parlance), typically with a few Terminal and TextMate windows on each. When I open a file (whether using File > Open… or the `mate` CLI), it often opens a tab in a TextMate window on another desktop, so I have to hunt through my desktops, detach the tab, and drag it back to the desktop I was originally using.
Is it possible to change this behavior, so TextMate always opens the file in a window on the current desktop?
- Kurt
Hi,
I tried uninstalling and reinstalling, but the OCaml bundle does not work:
- I can't select the OCaml language
- I doesn't appear in the Edit Bundles menu.
The /Users/louisabraham/Library/Application Support/TextMate/Managed/Bundles/OCaml.tmbundle folder exists and looks fine.
Has anybody the same problem?
Thanks,
Louis Abraham
Is there any option or defaults setting which allows me to drag a TM tab
out of the TM window (onto desktop for example) and have it work as if
you're trying to 'Move Tab to New Window'
I'm probably imagining it, but I feel like you could do this at one point.
Not a huge issue, as I can see I can assign a keyboard shortcut for it, it
just seems a little odd that it always treats dragging as the file contents
Thanks
When I am using TextMate to edit a LaTeX file, I find that command+Z (i.e., undo) no longer works. Rather than undoing something, it moves the cursor to about the middle of the document I am working on.
Anyone else having this behavior?
I’m using TexMate 2.0-rc.7 on 10.12.5.
Kyle
How can I send code from a source .R file to a Rdaemon window?
Or is this not how Rdaemon is intended to be used? Is it intended to be used ad hoc within a TextMate window for tasks such as simple calculations?
Many thanks
Ross
Hi all,
Like many of us, I suspect, I’ve got core bundles (like markdown, regex etc.) that I have customised with new snippets and commands, or just just tweaks to language.
I’d like extricate my code from these, so that
1. I can let TM2 easily update the mainstream bundle
2. I can control my mods
3. From that base of clarity, I can submit generally-useful changes back to the widely used central repo.
Are there any solutions to
1. Diffing the pristine bundle with my dirty copy?
2. Exporting modified snippets and commands to my personal repo
Ideally, the UUID exported commands would be entered into the plist of my personal repo, perhaps under “exported from <bundlename> as a menu name.
Hope others have had this problem and solved it :-) Otherwise bit of a manual trek in store for summer :-)
Best, tim
Hi,
I’ve noticed for a while that if you load a file that takes a while to syntax highlight and prompty press "⌘⇧T” (the function browser), textmate spins the beachball, becomes unresponsive and never seems to come back (hard crash out).
best, tim
Would be unfortunate if MiniTest isn’t supported in Rails since it appears that many Rails developers use TextMate, and Minitest is the default testing framework.
(Not to take away from RSpec, it is popular. I started with MiniTest and since I’m a Newbie I’m reluctant to add more confusion by changing testing methods.
Am I right in understanding that the Ruby on Rails bundle has the testing baked in and doesn’t use the version installed on the users computer?
More info on the change and background:
Rake commands have been changed to Rails commands. This was done to avoid confusion (no way to tell whether to use rake or rails). https://github.com/rails/rails/issues/18878
Overview of Rails 5 changes: https://www.youtube.com/watch?v=OaDhY_y8WTo
A Guide to Testing Rails Applications — Ruby on Rails Guides <http://guides.rubyonrails.org/testing.html>
>
> Rails 5 uses `rails` instead of `rake` for everything, including running
> tests.
>
> I may look into this if I find some time, but unfortunately all Rails
> projects I?m working on use RSpec instead of Test::Unit, so this is not
> really a problem for me personally and thus rather low-priority ?
>
> (On a side note: the RSpec bundle works fine with Rails 5 :-)
>
>
> Stefan
>
> Am 17.03.17 um 09:05 schrieb Koen Punt:
>> What's the problem with rake?
>>
>> Koen
>> On Mar 16 2017, at 11:47 pm, Greg <web(a)web.knobby.ws> wrote:
>>
>> I usually use the command line for testing (I?m more or less a
>> newbie to testing), but thought I?d try staying in TextMate, but see
>> it's still using Rake.
>>
>> I wish I could help, but I?m already in over my head.
>>
>> Greg
I usually use the command line for testing (I’m more or less a newbie to testing), but thought I’d try staying in TextMate, but see it’s still using Rake.
I wish I could help, but I’m already in over my head.
Greg
I'm hoping a TM dev could point me in the right direction of where TM's
inode caching takes place please?
I'm struggling to find it, but I think it might be one of the sources of my
issues using TM on a mounted remote filesystem
Quite often, if I change git branch on the remote filesystem (done via ssh
cmdline) it can end up with TM 'quick open' opening a completely different
file. This issue persists until the volume is unmounted and remounted.
I'd like to compile a custom TM which disables inode caching if possible
Thanks
Hello Ed, All,
None of my previous theories worked out, but I think I’ve had a break
through on this.
I can trivially reproduce it now by switching between iTerm2 and TextMate,
very quickly. Finder might work too, but I’ve been using iTerm. The key
seems to have less to do with which app, and more to do with how fast. My
machine is slow which may make the problem worse.
What I do know is if I reset my configs to their defaults (by removing
~/Library/Application Support/TextMate), this problem is eliminated. Adding
folders back one at a time TextMate still works until I add the
Session/Info.plist back.
I’ve attached my info.plst and made a video of what I see while I’m working
on this; https://youtu.be/C0nyIr5_dTQ
Ed, does removing this file do anything for you?
Everyone; Does this make any sense? Does anyone know where to begin solving
this bug? I sure don’t.
Thanks,
Graham Heath
On February 15, 2017 at 7:48:39 AM, Ed Wong (scampy(a)me.com) wrote:
Graham, I don’t have Diff Gutter enabled, just the standard line numbers,
folding arrows, bookmarks in the gutter.
Ed
On 14 Feb2017, at 3:56 PM, Graham Heath <graham.p.heath(a)gmail.com> wrote:
Thanks for the answers. What all is happening here is still a mystery but
I’ve got an update;
I’ve now had 24 hours without incident. My current theory is this relates
to SCM Diff Gutter, as after disabling it I haven’t seen the issue, but
thats just a theory. I’m re-enabling Emmet today, and I’ll add my linters
tomorrow.
Ed, do you have SCM Diff Gutter, or anything adding icons to your gutter?
Thanks,
Graham Heath
On February 10, 2017 at 9:08:03 PM, Ed Wong (scampy(a)me.com) wrote:
1: have more than a few TextMate windows or tabs open?
Usually many windows with many tabs. Some windows are minimized (those that
I’m not actively working on at the moment).
2: have a customized window bar title?
Yes - full document path.
3: have expensive to render documents?
A few files with > 5000 lines but generally < 1500 lines of PHP, ObjC,
HTML, JS or CSS. Some customizations done to PHP bundle and Mac Classic
theme.
4: SCM integration?
Git on some projects.
Hope this helps.
Ed
On Feb 10, 2017, at 5:33 PM, Graham Heath <graham.p.heath(a)gmail.com> wrote:
This continues to happen to me a few times a day, exclusively in TextMate.
Some days are worse, maybe better after a reboot, but thats not for sure.
Some more theories for you Ed, and anyone else :)
Do you …
1: have more than a few TextMate windows or tabs open?
2: have a customized window bar title?
3: have expensive to render documents?
4: SCM integration?
Here’s what I’m thinking _might_ be happening for each of these. Hopefully
one of these sparkles for someone.
1) TM is confused about what tab / window should have focus, and fails to
pick one.
2) I have a pretty complex stack of replacements to build my window title,
perhaps one of these replacements has a edge case that causes an error
3) TM may be selecting the right tab, but the selection doesn’t happen in a
reasonable time and aborts
4) The timing between focus and TM failing to select a tab seems related to
when the title bar updates when SCM is working
I’ve been experimenting with disabling plugins and add ons. Today I’m going
to try disabling the SCM plugin :(
If that doesn’t work I guess I’ll build TM and see what I can find. Short
of building it, there’s not a way better debug this issue, is there?
Thanks,
Graham Heath
On February 1, 2017 at 7:05:23 AM, Ed Wong (scampy(a)me.com) wrote:
Hi Graham,
My TM setup is pretty basic. No plug-ins, Shell Support installed, Git
support sometimes depending on the project.
I think Allan may be right though as I’ve seen this behaviour with Safari
and Finder as well so it could be an general OS issue.
Ed
On Jan 31, 2017, at 10:23 PM, Graham Heath <graham.p.heath(a)gmail.com> wrote:
Allan, I’m sorry to say, that this is behavior that I associate with
TextMate, not the os at large. I will keep an eye out for it and let you
know if I change my mind. But, as you’ll see below, I don’t really blame
your code either, I’m just wondering whats causing it. Its so erratic that
its probably not something simple.
Jacob, Ed, and others; Do you have any plugins? I haven’t done the full
reset, but I have disabled things that I feel might be offending. Its too
early to say, but I wonder if a plugin might be triggering when the app
gets focus, but something isn’t going quite right.
Tossing this on the pile; I’m using a early 2013 mbp that does seem to have
issues with apps hanging or freezing fairly often, so maybe whatever’s
causing that, is causing this.
Thanks,
Graham Heath
On January 14, 2017 at 11:44:42 PM, Allan Odgaard (mailinglist(a)textmate.org)
wrote:
On 14 Jan 2017, at 5:54, Graham Heath wrote:
> I’m regularly experiencing a bug where TextMate seems to receive
> partial
> focus. The title bar will appear to be selected, but the tabs never
> come
> into focus […]
This issue has existed with macOS for as long as I can remember, and I
have seen it for other apps as well, though it is visually much more
obvious with TextMate because of the tabs bleeding into the title bar.
I do not think the issue is with any of my code, since window management
is not the responsibility of the individual app and I am not aware of
code in TextMate that could affect this.
_______________________________________________
textmate mailing list
textmate(a)lists.macromates.com
http://lists.macromates.com/listinfo/textmate
_______________________________________________
textmate mailing list
textmate(a)lists.macromates.com
http://lists.macromates.com/listinfo/textmate
_______________________________________________
textmate mailing list
textmate(a)lists.macromates.com
http://lists.macromates.com/listinfo/textmate
Hi all,
since a few days, my textmate2 commands using ruby won’t start and just yield an error :
env: ruby18: Permission denied
I can’t recall any major change in my system recently.
Has anyone an idea?
David Bourguignon
All of a sudden I have *.pyc files showing up in my project drawer. I have
not changed any textmate config files or application versions.
I am using 2.0_b10, but have also tested 2.0_rc4.
I have also restarted my computer. I am using OSX 10.12.3.
I am using the default config values for inclusion and exclusion:
*/Applications/TextMate2.0_b10.app/Contents/Resources/Default.tmProperties*exclude
= "{*.{o,pyc},Icon\r,CVS,_darcs,_MTN,\{arch\},blib,*\~.nib}"
include = "{*,.tm_properties,.htaccess}"
binary = "{*.{ai,bz2,flv,gif,gz,icns,ico,jpg,jpeg,m4v,nib,o,pdf,
png,psd,pyc,rtf,scssc,swf,tar,tbz,tgz,tif,tiff,zip},Icon\r}"
I’ve hit a bug when using “Comment Line” in Terraform documents,
although it only occurs on certain lines.
# Steps to reproduce
1. Install the Terraform bundle. TextMate > Preferences > Bundles >
Terraform.
2. Create a file called "~/Desktop/example.tf" and change the document
syntax to "Terraform".
3. Enter the following contents in the file (between the '===' bars):
===
variable "aws_region" {
default = "us-east-1"
}
===
Note that line 4 contains whitespace, although the exact whitespace
chars seem to be arbitrary.
4. Place the cursor as follows:
a) At the start of line 1
b) Select all of line 1
c) Select all of line 4 (whitespace)
d) At the start of line 4.
After each, press ⌘+/ to Toggle Comment.
# Expected results
a/b) Line 1 will be commented out.
c/d) Line 4 will be commented out.
Note: in Terraform syntax, a comment is denoted by /* ... */
# Actual results
d) Works correctly.
All of a)-c) result in 'Failure running “Comment Line”.', with different
errors:
a) Comment Line:140:in `<main>': undefined method `to_a' for
"\u0000variable \"aws_region\" {":String (NoMethodError)
b) Comment Selection:140:in `<main>': undefined method `to_a' for
"variable \"aws_region\" {\n":String (NoMethodError)
${0:}
c) Comment Selection:140:in `<main>': undefined method `to_a' for "
\n":String (NoMethodError)
${0:}
Broadly similar – it just seemed a little odd that it doesn’t also
affect the case where your cursor is on a line containing whitespace.
# Environment
* macOS 10.12.1 (16B2657)
* TextMate version 2.0-rc.7
* ruby 2.0.0p648 (2015-12-16 revision 53162) [universal.x86_64-darwin16]
I'm using Textmate on Yosemite 10.10.15
The AS3 bundle has been updated for TM2 and I have that one installed
Each line I type adds an extra indent to the one before, where they should line up
If I delete the extra indent and start typing it just adds the indent again
I have two installs of TM2 both on Yosemite 10.10.5 and each one displays the same behaviour.
I tried another bundle (I forget which – maybe python) and the indent behaviour was normal.
How can I work out where the error is and how to trouble-shoot it, please?
Hi Allan and others,
Lately there hasn't been any (public) development of TextMate.
So I'm wondering if there's an update to be expected anytime soon?
I'm currently expecting performance issues with tab switching, closing, and
opening (CMD+T) of files. Is this a known issue, or how can I investigate what
is causing this?
– Koen
Sent from [Nylas Pro](https://link.nylas.com/link/33wd9n5m5l3cdcz4d1oikbwcz
/local-294f3511-5e38/0?redirect=https%3A%2F%2Fnylas.com%2Fnylas-
pro%3Fref%3Dn1&r=dGV4dG1hdGVAbGlzdHMubWFjcm9tYXRlcy5jb20=), the most powerful
email app for work
I’ll try using another email address for TextMate. Although I don’t have trouble with any other emails.
⌃⇧⌘F
⌃⌥⌘G
⇧⌥⌘F
⇧⌥⌘G
We’ll see how this comes across in the two emails. And a copy bcc’d to other addresses on the same server.
I’m guessing not many have this problem. BTW I’m on 1and1..com
Greg
> On Feb 24, 2017, at 8:01 AM, textmate-request(a)lists.macromates.com wrote:
>
> Message: 2
> Date: Fri, 24 Feb 2017 11:01:21 -0500
> From: Per Olofsson <magervalp(a)fastmail.fm <mailto:magervalp@fastmail.fm>>
> To: textmate(a)lists.macromates.com <mailto:textmate@lists.macromates.com>
> Subject: [TxMt] Re: textmate Digest, Vol 105, Issue 22
> Message-ID:
> <1487952081.2030038.891782960.07379631(a)webmail.messagingengine.com <mailto:1487952081.2030038.891782960.07379631@webmail.messagingengine.com>>
> Content-Type: text/plain; charset="utf-8"
>
> On Fri, Feb 24, 2017, at 10:52 AM, Greg wrote:
>
>> Why the question marks?
>
>
>
> Something between the mailing list and you appears to be having
> Unicode issues:
>
>
> http://lists.macromates.com/textmate/2017-February/040275.html <http://lists.macromates.com/textmate/2017-February/040275.html>
>
>
>
> --
>
> Per Olofsson
>
> magervalp(a)fastmail.fm <mailto:magervalp@fastmail.fm>
>
>
>
>
>
I'm on my third attempt on switching from TM1, but this time it's
actually going pretty smooth. One thing has me stumped though, there's
no keyboard shortcut for Replace & Find? I checked the commit history
and it was reassigned to Replace in f39ac28, leaving Replace & Find
shortcutless. There doesn't seem to be many options left for F and G (is
there a way to bring up a list of all active bindings?), but adding one
for Replace & Find would make my life a lot easier.
Thanks,
--
Per Olofsson
magervalp(a)fastmail.fm
Why the question marks?
Greg
> On Feb 24, 2017, at 4:00 AM, textmate-request(a)lists.macromates.com wrote:
>
> Message: 3
> Date: Fri, 24 Feb 2017 03:40:10 -0500
> From: Per Olofsson <magervalp(a)fastmail.fm <mailto:magervalp@fastmail.fm>>
> To: textmate(a)lists.macromates.com <mailto:textmate@lists.macromates.com>
> Subject: [TxMt] Re: Missing keyboard shortcut for Replace & Find
> Message-ID:
> <1487925610.1055419.891365096.02F0D25F(a)webmail.messagingengine.com <mailto:1487925610.1055419.891365096.02F0D25F@webmail.messagingengine.com>>
> Content-Type: text/plain; charset="utf-8"
>
> On Thu, Feb 23, 2017, at 10:36 PM, Allan Odgaard wrote:
>> On 23 Feb 2017, at 14:59, George McGinley Smith wrote:
>>
>>> It means please add it to TextMate ?
>>
>> But using what key equivalent?
>
> Is there a way of getting a definitive list of all active keyboard
> shortcuts? As far as I can tell the only free ones are:
>
> ???F
> ???G
> ???F
> ???G
>
> With my short fingers they're all two handed, but I find ???G is by far
> the easiest to type.
>
> Personally I use Replace & Find a lot more than I see myself using Find
> All, and I'd argue for giving back Replace & Find ??F and relegating
> Find All to something like ???F, but that's largely colored by my recent
> switch from TM1.
>
> --
> Per Olofsson
> magervalp(a)fastmail.fm <mailto:magervalp@fastmail.fm>
>
Hi, all, if you are not interested in the Modelica system definition and
simulation language, you can stop here ;-)
I have been looking for a TextMate bundle that supported Modelica —see,
for instance, http://openmodelica.org—, and could not find one, but I found
a SublimeText version, and instructions on how to create a TextMate one.
If you are interested, you might want to look for
https://github.com/juandesant/Modelica.tmbundle
Cheers,
Juande.
--
Juande Santander-Vela
System Engineer (Science Data Processor/Telescope Manager)
Square Kilometre Array/SKA Organisation
Jodrell Bank Observatory, Lower Withington
Macclesfield SK11 9DL, United Kingdom
Hi
I have to use a new MacBook Pro at work (a mixed blessing). It has a virtual real ESC key. I find this really awkward when typing Actionscript and using autocomplete. I need a real key (for tactile feedback).
I would like to use the § key at the top left as it is nearest to where the ESC key should be.
I have found and looked at KeyBindings.dict but cannot work out which is the correct key mapping to edit to get what I need.
I can see "$\033" = "previousCompletion:" , but not anything for 'nextCompletion'.
Is it best to attempt this here or at Bundle level??
Thank-you for any ideas, please
Simon
Hey there,
I’m regularly experiencing a bug where TextMate seems to receive partial
focus. The title bar will appear to be selected, but the tabs never come
into focus.
Please see https://imgur.com/a/ePa4j for screenshots.
In this "half focus" mode I can select, enter and delete text, but not
select, copy, paste or do many other things. Cmd-c and Cmd-v both cause the
system bell to ding. I can switch focused TextMate windows, but the "half
focus" window title bar remains focused, and the tabs never get selected.
I still haven’t found a reliable way to reproduce this.
I have 2 theories about what is happening;
1) Its related to some ongoing process, such as a linter still running
in the background or the code highlighting for complex code. This seems
relevant because I work some days without encountering the issue, others I
run into it many times. Its possible that the offensive days I have more
windows, or a more difficult to parse file open.
2) Its related to multiple displays. I’m on a MBP15in from early 2013
with a display hooked up over thunderbolt or hdmi. It does seem that its
much easier to reproduce this bug with a second display.
Perhaps there’s a combination of 1 and 2?
For a while I was switching between apps to get "full focus"; Switch from
Chrome to TextMate (get "half focus"), switch to Finder, switch to Chrome,
switch to TextMate (get "full focus").
Seems that if no one else is having the issue, its probably #1 and related
to my custom bundles.
Whatever the cause, hiding TextMate with Cmd+H and restoring it seems to
cause "full focus" reliably.
Is this making sense? Is there anything I can do to help debug?
This weekend I’ll try reverting to defaults and see if that helps. In the
mean time, what do you all think?
Thanks a million,
Graham Heath
Hello *,
not sure if this is a TM problem/feature at all; but I found it there, so …
Using TM 2.0-rc.4 and Skim 1.4.26 (100) on El Capitán 10.11.6, I have an odd phenomenon concerning search strings: they persist over several programs.
Example: In TM, I type something into the “Find:” field of the search dialogue, say “test123”. Then I change to Skim, and typing Cmd-F there I see “test123”. Same effect in the other direction (search words from Skim appear in TM-search), same if I use one of the search dialogues in Skims side panes. Also same with the mail search in Apple Mail.
I find this “feature” highly annoying. Is this some new Finder thing (I skipped from 10.6.8 directly to 10.11.x) or is it some TM thing? How can I disable it?
Thanks for all hints!
-Moss-
--
Debugging? Klingons do not debug. Our software does not coddle the weak.
-- not by Peter Dyballa (as he vows)
Hello.
Does anybody experience random selections while clicking on tabs or other
textmate windows (like search dialog) on Sierra?
I can't find reproducible sequence of events, but it is quite annoying.
Sometimes when I click on other tab text in the current tab gets selected
as if shift-clicked somewhere above the current view area.
I’ve been working with some old html that is moving to a markdown-based site.
Textmate has a neat feature where if you drop an html file onto your markdown window, its contents are pasted… as markdown! (neat!).
Question/feature request:
1. Can we have a “convert from html" command in markdown?
2. Is there an option on the html processor to [give](normal) links instead of [bibliography][1]
[1] Style links?
Currently in TextMate, it’s possible to implement completion using a bundle command. This can be invoked using a keyboard shortcut. Some IDE’s have implemented something like dot completion, that is, when a dot is typed the dot will be inserted in the text and a completion list will be shown. Some editors will even show this completion list for most characters that are typed.
In TextMate it’s possible to set the key equivalent of a bundle command to a dot, the problem is that the dot will then not be inserted in the text. It’s also possible to manually write out a dot in the bundle command. The problem with that is that the dot is not actually in the text buffer when accessing it in the bundle command. This will most likely not work because whatever is parsing the source code to generate the completion expects the dot to actually be there.
Is this something that can be implemented in TextMate, perhaps as a semantic class? Ideally it should be configurable to support any character, ideally more than one. Example, for many languages a dot is what’s needed, but for the C, C++ and Objective-C an arrow (->) would be necessary as well. Just to be clear, I’m only talking about a new way to trigger a bundle command, not implementing the actually completion.
--
/Jacob Carlborg
Hi,
The Hyperlink helper: "Lookup Word / Selection on Wikipedia and link" appears to be broken?
/Users/tim/Library/Application Support/TextMate/Ruby/1.8.7/lib/ruby/1.8/open-uri.rb:174:in `open_loop': redirection forbidden: http://en.wikipedia.org/w/api.php?format=xml&action=opensearch&search=Consc… -> https://en.wikipedia.org/w/api.php?format=xml&action=opensearch&search=Cons… (RuntimeError)
from /Users/tim/Library/Application Support/TextMate/Ruby/1.8.7/lib/ruby/1.8/open-uri.rb:132:in `open_uri'
from /Users/tim/Library/Application Support/TextMate/Ruby/1.8.7/lib/ruby/1.8/open-uri.rb:518:in `open'
from /Users/tim/Library/Application Support/TextMate/Ruby/1.8.7/lib/ruby/1.8/open-uri.rb:30:in `open'
from Lookup Selection on Wikipedia and link:19:in `getWikiEntries'
from Lookup Selection on Wikipedia and link:31:in `getWikiDef'
from Lookup Selection on Wikipedia and link:37
I'm not sure if this is a known issue, or if it is only Python, but the Jump to Symbol window has trouble properly displaying Python functions & methods that have multi-line parameter lists.
A function heading like this
def makeWeek(geo=AMERICAS,
post_by=STAFF_TYPE,
staff_types=[],
start=S_0630_AM,
end=E_0730_PM
):
Shows up in the Jump to Symbol window's table as
def makeWeek(geo=AMERICAS, post_by=STAFF_TYPE...)
Aligning the parameter names' leading edges with the opening parenthesis tends to work better with Python linters in this situation, but kind poorly for the Jump to Symbol window.
My way to quickly ge to the bottom of a document is to select it all and then right-arrow to place the cursor at the last character in the document.
For some time now, this doesn’t always work: Often one ends up some-way down, but not at the bottom. Repeating 2-4 times needed to get to the bottom.
It seems to occur when line wrap is on, and the document contains long lines (i.e., text paragraphs)
Perhaps something is a bit off in the algorithm for scrolling the page when lines are wrapping?
Best, tim
Hi all -
I keep a lot of functions folded in my normal workflow. This usually works fine, but when a non-TM program changes a file, (usually switching branches in Git from a client or command line) the folding settings aren't updated and I end up with strange folding in the middle of a line that may throw off everything in the file.
This isn't a huge problem (and I bet if I always used TM's source control plugin, this would be handled automatically, right?) but I wonder if there's any way to just tell TM "hey, forget all the folding in this file". One might think that cmd-opt-0 (Un/fold All Levels) would do this, but it seems to toggle all the folds without scanning the syntax to see if the fold locations are valid. Anybody know how to just delete all folding for a given file? As it is, I have to scan through big files, looking for hidden '...' glyphs.
I see this most often in Python, which has weird folding behavior, but I've also seen it in Objective C & Javascript, which should have much simpler folding grammars since they're {}-delimited.
Cheers,
Evan
Hi,
For creating web templates I use Xdebug. In Xdebug you can set
`txmt://open/?` in order to open the file with the error in TM. On previous
versions of Safari (< 10.0.2) TM opens without any problems. But now (>
10.0.2) Safari every time ask if it may open TM: see screenshot:
http://feek.d.pr/V1pH
Is there a way to disable this pop-up? I cannot find any settings in Safari.
Or is it maybe related to a "unsigned developer"?
how to reproduce: just paste `txmt://open/` into Safari.
I use TM 2.0-rc.4 on OSX 10.11.6
Regards,
Feek
--
View this message in context: http://textmate.1073791.n5.nabble.com/Safari-10-popup-when-opening-a-txmt-o…
Sent from the textmate users mailing list archive at Nabble.com.
Hi all,
I have found that I had to finally learn to create Commands in a TM Bundle to move beyond what snippets alone can do (even though they have such a brilliantly simple syntax...)
Is there any canonical listing of values to use Semantic Class hooks?
Also I couldn't find it in the docs, but realized by reading other examples, the way to get the whole document text reliably is to read stdin in a script, then write to stdout.
It might be really great stuff to get documented for everyone.
Finally, what is the recommended flow for managing & sharing bundles?
Certainly, putting changes into Github would be ideal where possible.
But here are two use cases I have.
In my workplace we have stuff we cannot share outside, but we would like to share and maintain a group repo internally to grow up some TM Bundles.
I also sometimes need to work on multiple systems.
I'd like to understand the best way to sync or centralize and update bundles.
What are the best practices with this?
Hi there, I’m currently working on updating an old plugin (EditorConfig — https://github.com/mr0grog/editorconfig-textmate) and it needs to be able to trim the trailing whitespace from lines. I’m currently doing this by getting the OakDocument’s `content` property, editing it, and setting it back, but that can cause selections to move around (not really surprising).
To address selections, the most straightforward method I could find was to use `OakTextView accessibilityAttributeValue:
NSAccessibilitySelectedTextRangesAttribute` (and the associated setter), like so:
NSString *content = [document performSelector:@selector(content)];
NSMutableArray<NSValue *> *selections =
[textView accessibilityAttributeValue:NSAccessibilitySelectedTextRangesAttribute];
//
// do some stuff to manipulate `content` and `selections` here...
//
[document performSelector:@selector(setContent:) withObject:content];
[textView accessibilitySetValue:selections forAttribute:NSAccessibilitySelectedTextRangesAttribute];
This works great with normal selections, but I had expected column selections to show up as multiple selections here. However, they show up as a single contiguous selection from the starting column on the first row of the selection to the ending column on the last row of the selection. That means that, by the end of this operation, any column selections become normal contiguous text selections.
Is there any straightforward way to get and set column selections from a plugin, where I can’t easily mess with the C++ objects? Is there a better way I should be approaching this whole operation in the first place? (I recognize that my approach above might not be great, but couldn’t find anything else workable; I tried poking at OakTextView’s `filterDocumentThroughCommand` method, but had trouble with it since it requires arguments that are C++ types).
Thanks for any advice,
-Rob
Hi all,
I wished to enable spell checking for comments in source languages.
I edited the settings items in the source bundle, enabling spelling for my scope.
This worked (yay), but I realised it would be better to stash this in my private bundle, not fork the main source.tmbundle.
So I reverted the scope as a first step to making a new settings item. The bundle is now marked as delta, and that isn’t right.
SO I deleted the is delta key. That was a mistake: Now I have an item labeled “untitled” that is empty, and no spellcheck enable setting item.
Try reinstating the delta key, and setting to false. No worky.
Try reverting the bundle: you can’t unselect it.
Deleted the bundle and installed from in TM2: still has my untitled item and no. Must be buried in an invisible cache..? Where…
Curse :-)
Email list: How do you really revert a core bundle to base?
tim
PS: It would be GREAT to have a revert to default right-click action for TM bundles, which cleared invisible caches etc… Or make the grayed out source bundle tick function as a togglle; click it, and it throws a dialog asking “really revert to base source bundle (bundle cannot be deactivated, only reverted)?” then does that.
PS: Feel free to improve this stack overflow answer so others can share this knowledge.
http://stackoverflow.com/questions/41565216/how-do-i-get-textmate-2-to-enab…
PPS: bit of a hot mess of 10-year old TM1 info and partial TM2 info on the interweb at present: we should try and clean up where possible now everyone is moving to TM2 beta
t
>> Prerequisites listed include Boost.
>>
>> My question is:
>> What is the recommended way to build/install Boost for TextMate
>> building?
>
> Using homebrew is the easiest, this should take care of all
> prerequisites:
>
> brew install ragel boost multimarkdown hg ninja capnp
> google-sparsehash libressl
>
>> Using Homebrew, boost has additional options and they seem possibly
>> valuable but also potentially problematic.
>
> TextMate only use a few header-only libraries from boost, so I don?t
> think any of the options have any effect on TextMate.
Ah. I didn't even think to survey the header includes.
Thanks it's exactly what I wanted to know!
I might run it through Doxygen to get a list easily going forward.
Thanks!
JJ
Hi,
Since yesterday evening, pasting to pastie.org <http://pastie.org/> does not work anymore.
It is not the usual error 503 service unavailable, but there are 2 possible errors:
There was a problem pasting your text, pastie.org responded with 522 Origin Connection Time-out
There was a problem pasting your text, pastie.org responded with 400 Bad Request
Am I the only one? Or did pastie change their endpoint?
Thanks,
Louis
Hi,
There are quite some commands that when invoked result in an error. The error
itself is correct in a way, but most of the time not very helpful, and I think
they can be avoided by adding some preconditions to these commands.
Take for example "Browse Annotated File (Blame)"; when running without having
a file open or selected in the filebrowser, it displays the following error:

Or the scm gutter; when deleting or renaming a folder outside of TextMate, but
which is currently open in TextMate, then after deleting closing one of the
tabs displays the following error:

I think all these kinds of errors have to be solved at bundle level, but like
to hear some opinions before I start opening pull requests.
– Koen
Sent from [Nylas N1](https://link.nylas.com/link/33wd9n5m5l3cdcz4d1oikbwcz
/local-aa207fc5-447b/0?redirect=https%3A%2F%2Fnylas.com%2Fn1%3Fref%3Dn1&r=dGV4
dG1hdGVAbGlzdHMubWFjcm9tYXRlcy5jb20=), the extensible, open source mail
client.
Prerequisites listed include Boost.
My question is:
What is the recommended way to build/install Boost for TextMate building?
Using Homebrew, boost has additional options and they seem possibly valuable but also potentially problematic.
Sadly, Boost doesn't ship in parts, even though it can be used linked to in independent parts (mostly).
brew says this about Boost:
$ brew info boost
boost: stable 1.63.0 (bottled), HEAD
Collection of portable C++ source libraries
https://www.boost.org/
Not installed
From: https://github.com/Homebrew/homebrew-core/blob/master/Formula/boost.rb
==> Dependencies
Optional: icu4c
==> Requirements
Optional: mpi
==> Options
--c++11
Build using C++11 mode
--universal
Build a universal binary
--with-icu4c
Build regexp engine with icu support
--with-mpi
Build with MPI support
--without-single
Disable building single-threading variant
--without-static
Disable building static library variant
--HEAD
Install HEAD version
==> Caveats
Building of Boost.Log is disabled because it requires newer GCC or Clang.
I tried:
brew install --c++11 --with-icu4c --with-mpi boost
I know macOS has its own ICU, but I also know Apple prefers people build against their own ICU if they need direct ICU support.
MPI wanted to build its own gcc and then I got a warning about gcc's dependency on a different C++ being a potential issue using boost with clang.
(seems insane to me)
"Warning: boost dependency gcc was built with a different C++ standard library (libstdc++ from clang). This may cause problems at runtime."
MPI adds these dependencies:
gmp, mpfr, libmpc, isl, gcc
Bundle commands support autoRefresh, but the option isn't exposed in the
bundle editor.
Is this a known issue or an intentional limitation of the bundle editor?
– Koen
Sent from [Nylas N1](https://link.nylas.com/link/33wd9n5m5l3cdcz4d1oikbwcz
/local-1313e4da-5a74/0?redirect=https%3A%2F%2Fnylas.com%2Fn1%3Fref%3Dn1&r=dGV4
dG1hdGVAbGlzdHMubWFjcm9tYXRlcy5jb20=), the extensible, open source mail
client.
Hi guys,
I guess this is primarily of interest to Rene Schwaiger. A long time ago we discussed LaTeX Tidy, \labels and code folding. I love everything about LaTex Tidy except for the fact that it puts \labels on a separate line to \begin or \end, so when I use code folding all I see is a list of \begins, \ends, \sections and the like, without a \label in sight. Rene pointed me to latextidy.pl, which was the help I needed, and I hacked the file so the code now ends
$string =~ s/({[^}]*}[ \t]*)(\n[ \t]*)((\[[^\]]*\][ \t]*)*(\\label[ \t]*(\[[^\]]*\][ \t]*)*{[^}]*}[ \t]*)*)\n?[ \t]*/$1$3$2/mg;
$string =~ s/((\\label[ \t]*(\[[^\]]*\][ \t]*)*{[^}]*}[ \t]*)+)(\n[ \t]*)(\\end{[^}]*})/$4$1$5/mg;
$string =~ s/(\\section.*(\\label[ \t]*(\[[^\]]*\][ \t]*)*{[^}]*}[ \t]*))(\S)/$1\n$4/mg;
$string =~ s/(?<=\S)[ \t]*(?:\n[ \t]*){0,2}(\\section)/\n\n\n$1/mg;
$string =~ s/(?<=\S)[ \t]*(?:\n[ \t]*){0,1}(\\subsection)/\n\n$1/mg;
print $string;
I’ve been using this for a while now and it has proved fairly reliable. I’ve allowed for optional arguments on \labels, labels pinned to \end (for captioned environments), and insert blank lines before \sections and \subsections. Now when I code fold I see not just the environments but their \labels, which mean something to me.
This message is intended in the spirit of giving back. Please feel free to ignore this hack, or try it out, change it, whatever.
Regards
Robert Milton
Hey
I'm occasionally experiencing smaller performance hickups in TM. Writing a
line of code, there's a short delay before the input appears; delay is maybe
in the range of half a second. It's infrequent enough not to be really
bothersome, but ideally, I'd like to know the cause.
It could in fact be some (too expensive) regexes I defined in my customized
language grammars, but I can't say for sure right now.
That's why I was wondering: is there a relatively simple way I can profile
TM/TM processes, in order to find out if it is in fact my own grammar rules
causing these hickups, or if it's something else?
--
View this message in context: http://textmate.1073791.n5.nabble.com/Occasional-performance-hickups-when-e…
Sent from the textmate users mailing list archive at Nabble.com.
I often see in my "Open Recent" menu the full path to a file and the same
file is repeated several times in the menu. It makes the menu quite large
and a bit hard to read. See attached screenshot.
In this case I had just created a new text file and saved to it about a
dozen times.[image: Inline image 1]
Dear All:
I am unable to typeset a latex file using the latest TextMate version with Skim. I can typeset the text file first time, but cannot typeset it with subsequent changes to the text file. Can you help me resolve this obvious bug?
Thanks,
Tao
Hi all
I’ve updated TextMate to version “2.0-rc.4”. Since updating to this version, several Keyboard Maestro macros that are triggered by keyboard shortcuts have stopped working. The problem appears to be because the new version of TextMate has introduced some new keyboard shortcuts. For example, I have a Keyboard Maestro macro that is triggered by the keyboard shortcut Shift + CMD + O. But in the new version of TextMate, this keyboard shortcut is assigned to “Open Recent Project”. Is there are way I can stop Shift + CMD + O triggering “Open Recent Project” in TextMate?
Ross
I haven’t been using this command. But tried it with RC3 and I stopped looking at 30k files. Is it looking at my entire drive. The files are beyond scope that I have open, they might be in scope of folders I’ve opened. Ah, I now guess I see what’s going on. I Cmd-T on an untitled, unsaved file. Guess this is an unintended consequence.
OT: I’m surprised to see in this modern era, question marks in place of apostrophes, and things like "seeing the ?T + ? issue with.” In case this gets munged on the way, in Mail on my up to date Mac this has a question mark in front of the T and another question mark following the plus sign.
Greg
> On Nov 19, 2016, at 4:00 AM, textmate-request(a)lists.macromates.com wrote:
>
> In rc.3 the main thread is stalled until the list has updated with the
> currently open files,
Hi,
Congratulations on reaching the significant milestone of RC1 on Textmate2. Unfortunately a bug has crept in in this release.
In the attached video, I press Cmd+T followed by Enter to switch between the current and previous files. The first time, I wait for the list to appear before pressing Enter, and it works as expected. The 2nd, 3rd and 4th times I quickly press Cmd+T followed by Enter before the list appears – the dialog closes but the previous file is not loaded. On the 5th time, I wait before pressing Enter again, and it works.
Clearly the problem is probably caused by the number of files. Prior to RC1, the list appeared instantaneously, but additional files filled in the list later. Is it possible to go back to that behaviour which has worked well for many years?
Regards,
Jonathan Monahan
Hi,
It happens often that after I move a file to another directory, textmate
doesn't handle the updated path to the file.
See this screen recording: http://recordit.co/3sdYVDOcF8
When I open the file, it still opens the file from the old path, where it no
longer exists, and thus results in an empty document.
– Koen
Sent from [Nylas N1](https://link.nylas.com/link/33wd9n5m5l3cdcz4d1oikbwcz/loc
al-28028436-6279/0?redirect=https%3A%2F%2Fnylas.com%2Fn1%3Fref%3Dn1&r=dGV4dG1h
dGVAbGlzdHMubWFjcm9tYXRlcy5jb20=), the extensible, open source mail client.

Hi,
Congrats with RC1. Unfortunately, my bundles are messed up. I lost a
number of my shortcut commands that I have always worked with. I lost my
RSpec bundle functionality, which I managed to restore by manually
reinstalling the bundle from Github.
Is there something I must do/should have done to keep my bundles working
as they have done before? I did accept the request to move the bundles
to the new structure when I restarted Textmate, but that hasn't helped.
Some guidance on this topic would be very helpful.
Thank you!
Pascal
Hi,
As a follow up to the problems with my bundles: the shortcut for
duplicate line is not working anymore: it clears the whole file now
instead. I have checked if there are any conflicting shortcut keys in
the "Select bundle item" menu, but there aren't any. Any thoughts,
because this is very annoying.
Thank you!
Pascal
I seem to be ending up with two TM dock icons quite a bit lately, not 100%
sure but it seems to coincide with when I accept a LiveUpdate.
No causing me a problem but I just wonder if this is normal? Feels like it
might be a bug that nobody has yet reported....
Getting the following when system submitting a crash report:
Internal Server Error
PDOStatement::execute(): SQLSTATE[HY093]: Invalid parameter number: no
parameters were bound
Hi,
I remember something about customizable tab titles, but couldn't find what I
was looking for in the archives.
So here's the question; there's a windowTitle property for .tm_properties, but
is something like that also possible for tab titles? Because I'm currenty
working on a project with a lot of identical named files:
My tabs currently look like this:

But if it could be possible to include the name of parent directory, that
would make it usable.
Koen
Sent from [Nylas N1](https://link.nylas.com/link/33wd9n5m5l3cdcz4d1oikbwcz
/local-eb0dadc9-7075/0?redirect=https%3A%2F%2Fnylas.com%2Fn1%3Fref%3Dn1&r=dGV4
dG1hdGVAbGlzdHMubWFjcm9tYXRlcy5jb20=), the extensible, open source mail
client.
Hey all,
I've been having this issue for a while now where the global find window
doesn't search .hbs (Handlebars) files. It's now occurring too with .elm
files. There's nothing in my .tm_properties that would suggest anything
like this:
windowTitle = "$TM_DISPLAYNAME – $TM_DIRECTORY (git: $TM_SCM_BRANCH)"
excludeInFolderSearch =
"{$excludeInFolderSearch,$extraExcludes,dist,node_modules,deps}"
excludeInFileChooser =
"{$excludeInFolderSearch,$extraExcludes,dist,node_modules}"
And the global find is matching all file extensions. Not sure where else to
look here, anyone know what I'm doing wrong?
Thanks!
Hello,
I cloned https://github.com/textmate/textmate <https://github.com/textmate/textmate> to my Mac, then successfully built and ran it.
But when I copy the binary to another Mac, it cannot be launched because of missed dylibs.
I compared linked dependencies with public Beta version and detected the following difference:
/Applications/TextMate.app/Contents/Frameworks/plist.dylib:
@rpath/plist.dylib (compatibility version 1.0.0, current version 1.0.1)
@rpath/cf.dylib (compatibility version 1.0.0, current version 1.0.1)
@rpath/io.dylib (compatibility version 1.0.0, current version 1.0.1)
/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation (compatibility version 150.0.0, current version 1258.1.0)
/usr/lib/libc++.1.dylib (compatibility version 1.0.0, current version 120.1.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1226.10.1)
vs
/Users/vadim/build/TextMate/Applications/TextMate/TextMate.app/Contents/Frameworks/plist.dylib:
@rpath/plist.dylib (compatibility version 1.0.0, current version 1.0.1)
@rpath/cf.dylib (compatibility version 1.0.0, current version 1.0.1)
@rpath/io.dylib (compatibility version 1.0.0, current version 1.0.1)
/System/Library/Frameworks/CoreFoundation.framework/Versions/A/CoreFoundation (compatibility version 150.0.0, current version 1348.15.0)
/usr/lib/libc++.1.dylib (compatibility version 1.0.0, current version 307.4.0)
/usr/local/lib/libcapnp--no-undefined.dylib (compatibility version 0.0.0, current version 0.0.0)
/usr/local/lib/libkj--no-undefined.dylib (compatibility version 0.0.0, current version 0.0.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1238.0.0)
Apparently that another Mac does not have capnp and kj in /usr/local/lib/.
Maybe you can hint why these dylibs are linked and why the paths are absolute? 🙏
Thanks in advance,
Vadim
P. S. This is my first request, and I submitted it to the IRC channel because not sure where to look for help, sorry for duplication.
I use the pdflatex option and view in Skim. Ever since I installed macOS Sierra, when I typeset my latex file, the page in the created pdf file no longer corresponds to the location in the latex file. In fact, the opened page is where the last opened page is. I would appreciate your help with any solution to this problem.
Tao
Hi,
I’m running TextMate 2.0-beta.12.26 on OSX El Capitan (10.11.6)
I ran into the following issue today:
While going through some code refactoring, I changed a couple of files, but I forgot to save one of these files (so it showed up half transparant in the file list sidebar).
I then modified the name of the folder which contained that open file (I did that in TextMates file list sidebar). The project is in a git repo by the way.
After that I moved focus to another program (I’ve got TextMate set to automatically save on focus lost) but TextMate then switched to the open file, presenting me a modal sheet “No parent folder for <filename>. Do you wish to create a folder at <previous_name_of_renamed_folder>” with buttons ‘Cancel’ and ‘Create Folder’.
None of these buttons are reacting to mouse clicks, space bar or return keys. Due to the sheet being modal, my only recourse is to force quite TextMate.
After relaunching TextMate, I found the file in the renamed folder (without the unsaved changes I made to it).
I made the changes again, saved this time, changed the file name and then changed the folder name.
On focus lost showed me the same modal sheet again! (after which I had to force quite TextMate again).
Now when I open this project in TextMate, it opens in the tab containing the above file, with the new contents, but with the *old* name. The file list sidebar shows the new name.
As soon as I move focus elsewhere, it shows me the unresponsive modal sheet (with the old file name in the error message!), so I’m actually kind of stuck now…
Best regards,
Bas
Hi,
I recently created a [TextMate bundle][] for [ABNF][]. The [RFC][] that describes the current version of ABNF explicitly states that the standard uses CRLF (Carriage Return + Line Feed) for line endings.
[TextMate bundle]: http://github.com/sanssecours/ABNF.tmbundle
[ABNF]: http://en.wikipedia.org/wiki/Augmented_Backus–Naur_Form
[RFC]: https://tools.ietf.org/html/rfc5234
TextMate already supports different line endings in the “Save Dialog”. This works quite well, as far as I can tell. We can also explicitly tell TextMate to always use specific line endings via “Preferences” → “Files” → “Line Endings”. If we set this option to “CRLF” then TextMate adds the entry
lineEndings = '\r\n'
to the global properties file `Global.tmProperties`. However, in both the current development version of TextMate (2.0-beta.12.26), and the lastest normal release (TextMate 2.0-beta.12.4), this feature does not seem to work as expected:
---
# Steps to Reproduce
1. Set line Endings to “CRLF” in “Preferences…”
2. Create a new file via `mate test.txt`
3. Enter the following text in the new document
line 1
line 2
4. Check the line endings via the command `file test.txt`
# Expected Result
The `file` utility displays the following output:
test.txt: ASCII text, with CRLF line terminators
# Actual Result
The `file` utility displays the following output:
test.txt: ASCII text
---
Maybe I am doing something wrong here?
I would also expect that the default “Save Dialog” reflects the line ending option currently set via “Preferences…”. This seems not to be the case. If I save a new document via `⇧` + `⌘` + `S`, TextMate always shows “LF” as default option for line endings. Again, maybe this might be a problem with my expectations on how the system works.
After this long introduction :o) here is my feature request. It would be very nice if it would be possible to set scope- and file-specific line endings. For example, I want to create a new “Settings” item set to the scope `source.abnf` that contains the following code:
{ lineEndings = '\r\n'; }
The “TextMate Settings” grammar currently marks `lineEndings` as invalid key. I suspect this feature is currently not available/implemented? If this is correct, then please consider this a (low priority) request to add this feature.
Kind regards,
René
I'm on OS 10.12.1, TM2 (12.26) and Skim 1.4.22 as viewer. Syncing pdf to source works, but not source to pdf. I looked at Skim's user list but couldn't find a solution. I wonder if something has to be done to the LaTeX bundle to restore this ability? Or is it an Apple 10.12 problem and I should wait for an Apple fix....
--Gildas