On Mar 20, 2006, at 6:26 PM, Pawel Szymczykowski wrote:

On 3/20/06, Trevor Harmon trevor@vocaro.com wrote:

...

Has anyone written a "zap gremlins" command for TextMate? BBEdit has this, but an equivalent doesn't seem to be available in TextMate.

Please forgive me if I react in total and absolute confusion but.. what????

From the BBEdit User Manual:

This command displays a sheet which allows you to remove or replace various non- printing characters, often known as “gremlins”. Use this command when you have a file that may contain extraneous control characters, or any non- ASCII characters, which you wish to identify or remove.

It allows you to selectively remove non-ASCII characters, control characters, and/or null (ASCII 0) characters, and either delete them, replace them with their character code, or replace them with a specified character (defaulting to the standard bullet (•) character).

I can't answer Trevor's question, but hopefully that clears up what he's asking for those of you who haven't used BBEdit before.

Thanks, Jeff

Text > Remove Unprintable Characters in Document

the name "zap gremlins" is way cooler ^_^ let's rename this command to "stab lemmings" or something like that! (you know, to avoid trademark disputes)

ciao,

D.

On 21/3/2006, at 2:57, Trevor Harmon wrote:

[...] I think they should be shorter. For instance, is having "in Document" as part of the name really necessary?

It is, yes. Because if you have a selection, it is “Remove Unprintable Characters in Selection”.

Another example: "Convert Spaces to Tabs (Estimating Indent)" That's quite a mouthful. Why not just "Entab"?

Because personally I had no idea what Entab or Zap Gremlens meant, before BBEdit users told me (and I would think Entab would be to add one tab, e.g. increase indent).

The goal with TextMate’s menu labeling is not verbosity, but it is non-ambiguity. There is a lot of commands, so it is a priority that the names spell out what the commands do. And if you try to rename all the commands to shorter names, I think you will find this may work for only very few of them, causing asymmetry in the naming system.

On 21/3/2006, at 4:19, Dan Ridley wrote:

On 20 Mar 2006, at 18:40, Allan Odgaard wrote:

...

if you try to rename all the commands to shorter names, I think you will find this may work for only very few of them, causing asymmetry in the naming system.

Oh, I don't know about that. Given "Zap Gremlins" and "Entab" as examples, you could go on for quite a while in that vein.

I suggest: Poke Edges Chop Super Chop Transmogrify Halo Effect Decaffeinate

ROTFL! :)

Dan Ridley wrote:

I suggest: Poke Edges Chop Super Chop Transmogrify Halo Effect Decaffeinate

Ooooh, that could be a great idea for a masochistic plugin to TextMate: Convert words in commands to one chosen at random from a list like that ;)

Jeroen.

On Mar 20, 2006, at 4:51 PM, Chris Thomas wrote:

...

Has anyone written a "zap gremlins" command for TextMate? BBEdit has this, but an equivalent doesn't seem to be available in TextMate. Thanks,

Text > Remove Unprintable Characters in Document

Ah, thanks for the pointer. I completely missed that.

Unfortunately, it doesn't work. That is, it doesn't work like BBEdit's Zap Gremlins. The problem is that TextMate's version is too conservative. With Zap Gremlins, I have the option of removing any non-ASCII characters, while TextMate only removes "unprintable" characters. A case in point:

http://vocaro.com/trevor/files/t.cc

Compiling the above in g++ gives: t.cc:8: error: stray '\194' in program t.cc:8: error: stray '\160' in program

The problem characters are 0xC2 and 0xA0, which apparently are "printable" (according to TextMate) but are not compilable.

Perhaps we simply need an extra command: "Remove non-ASCII Characters in Document".

Trevor

On 21/3/2006, at 3:02, Corey O'Connor wrote:

Does the "Remove Unprintable Characters in Document" command take into consideration the character encoding of the file?

Commands always see the document as UTF-8 and work with that. The encoding of the file (on disk) has no influence on that, as TextMate convert from/to the encoding used on disk when loading and saving the file.

In a brief experiment with the supplied file I found that in some encodings the byte sequences was printed, but in others (UTF-8) the byte sequence was not printed.

A file has one encoding -- if you load it with the wrong encoding, the result is effectively a broken file.

On Mar 20, 2006, at 6:34 PM, Allan Odgaard wrote:

On 21/3/2006, at 2:51, Trevor Harmon wrote:

...

[...] Perhaps we simply need an extra command: "Remove non-ASCII Characters in Document".

You might, and you can just create it yourself.

I understand, but what I meant by that was, "I think this should be built-in to TextMate."

Go to the Bundle Editor (Window menu). Locate the current command in the Text bundle. Duplicate it and change it to:

perl -pe 's/[^\t\n\x20-\x7F]//g'

Thanks for the tip! I didn't expect it to be that simple, but that's because I'm always forgetting the power of regex...

Trevor