Re: commenting out blocks of code
Steve Holden wrote: Steven D'Aprano wrote: On Fri, 17 Feb 2006 19:12:01 -0500, Peter Hansen wrote: Scite, for example, lets me selected a block and hit Ctrl-Q to either comment or uncomment the block. (It does this by prefixing each line with #~ instead of just #, which allows it to detect when a line is already so commented and reverse the operation.) It is *easy* to detect when a line is already commented. It starts with a #. The ~ is superfluous. Commenting and uncommenting should be two different commands: the whole point of nested comments is that it allows you to comment a block of text which may already contain comments. Having one command do both commenting and uncommenting according to the presence or absence of semantic clues in the text is a recipe for failure (No you stupid computer, I want to COMMENT that block, not uncomment it!!!). Imagine if your text editor used cntl-C for both copy and paste, somehow guessing whether you wanted to copy selected text or paste over it according to some subtle clue in the text itself. Wouldn't that be fun? I agree that's a less-than-sensible feature implementation. I submit that Steve and Steven (neither of whom has apparently even *used* this feature of Scite before criticizing it so strongly), should consider whether in actual practice something like this might not be much more effective than they can apparently imagine. I'm strongly critical of many features of most editors, and thus use very few of them, and I also happen to have a solid background in GUI design and ergonomics. I'm also a heavy user of Scite, and have found it readily accepted amongst the groups of programmers with whom I've worked. I and my team have found the Ctrl+Q feature of Scite to be very sensible, highly effective and useful, and in fact *better* than alternatives that we've used before in other editors. Kudos to Neil for having the (apparently) unusual design sense to think of it and implement it! So there. Harumph. :-) -Peter -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
Atanas Banov: here is something you both seems to have not considered: imagine you make decision if ^Q has to comment or uncomment based on the 1st line and not on each line individually in the block. When first implementing Ctrl+Q, this looked to have similar advantages and disadvantages to the chosen implementation. However, it wouldn't allow a technique I've come to use which is to have two versions of some code or section of a configuration file and to toggle between the two by selecting the whole range and pressing Ctrl+Q. #code1 ##comment1 ##comment2 #code2 note how ## maintains where comments were. now, for the same selection, ^Q again? 1st character is #, so the editor is asked to uncomment, dwim: The problem with ## as opposed to #~ is that ## is common in much source code particularly in barrier comments and comment sections: Persistence section ## These functions are responsible for load ## and save for the core media database. Neil -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
Steven D'Aprano wrote: Scite, for example, lets me selected a block and hit Ctrl-Q to either comment or uncomment the block. I see the developers of Scite are either newbies to GUI programming, or they operate in a world of their own. Ctrl-Q is reserved for Quit (or Exit) in every GUI API I know of. Your experience obviously differs from mine. I just tried Ctrl-Q on both windows and ubuntu and in neither case did it exit applications consistently: For example Firefox uses Alt-F4 in both cases. My own favourite editor (Epsilon) uses Ctrl-Q for the 'quoted-insert' command (hardly suprising given its similarity to emacs). (It does this by prefixing each line with #~ instead of just #, which allows it to detect when a line is already so commented and reverse the operation.) It is *easy* to detect when a line is already commented. It starts with a #. The ~ is superfluous. Not so easy if the lines to be commented already contain some lines starting with comments. Messing around with the comments themselves sounds highly unsatisfactory though. I agree with you that the best solution is to use different commands (or as Epsilon does a single command which can be modified with a prefix Ctrl-U). That way I can select a large region and comment it all out, or uncomment disjoint parts of it as I choose. -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
On Sat, 18 Feb 2006 03:14:49 +, Neil Hodgson wrote: I am the developer responsible for SciTE and the Ctrl+Q decision. Ctrl+Q is not reserved for Quit on every platform. [snip] Thank you for your thoughtful response. Although I'm not convinced by your reasoning, I will concede that perhaps my post was a little hasty. -- Steven. -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
Steven D'Aprano wrote: On Fri, 17 Feb 2006 19:12:01 -0500, Peter Hansen wrote: john peter wrote: is there a similar mechanism in python, other than prefixing the '#' character to the start of each statement i do not want to execute (which gets old very quickly if one needs to comment and uncomment several statements a couple of times while playing around with code say during initial design)? In addition to the approach Paul has suggested, you should also consider using a better editor, as most decent editors should have some kind of block comment/uncomment feature that prevents that operation from getting old very quickly... Scite, for example, lets me selected a block and hit Ctrl-Q to either comment or uncomment the block. I see the developers of Scite are either newbies to GUI programming, or they operate in a world of their own. Ctrl-Q is reserved for Quit (or Exit) in every GUI API I know of. I don't think this is common enough to regard as any kind of de facto standard. Just to give two further counterexamples, neither TextPad nor Notepad action a Ctrl/Q in that way. Perhaps I'm just using the wrong operating system? (It does this by prefixing each line with #~ instead of just #, which allows it to detect when a line is already so commented and reverse the operation.) It is *easy* to detect when a line is already commented. It starts with a #. The ~ is superfluous. Commenting and uncommenting should be two different commands: the whole point of nested comments is that it allows you to comment a block of text which may already contain comments. Having one command do both commenting and uncommenting according to the presence or absence of semantic clues in the text is a recipe for failure (No you stupid computer, I want to COMMENT that block, not uncomment it!!!). Imagine if your text editor used cntl-C for both copy and paste, somehow guessing whether you wanted to copy selected text or paste over it according to some subtle clue in the text itself. Wouldn't that be fun? I agree that's a less-than-sensible feature implementation. regards Steve -- Steve Holden +44 150 684 7255 +1 800 494 3119 Holden Web LLC www.holdenweb.com PyCon TX 2006 www.python.org/pycon/ -- http://mail.python.org/mailman/listinfo/python-list
commenting out blocks of code
in java, i can prevent a block of code from executing by bracketing the block with comment indicators, as shown below: /* statement1 will not execute; statement2 will not execute; */ statement3 will executeis there a similar mechanism in python, other than prefixing the '#' character to the start of each statement i do not want to execute (which gets old very quickly if one needs to comment and uncomment several statements a couple of times while "playing around with code" say during initial design)? Yahoo! Autos. Looking for a sweet ride? Get pricing, reviews, & more on new and used cars.-- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
john peter wrote: in java, i can prevent a block of code from executing by bracketing the block with comment indicators, as shown below: /* statement1 will not execute; statement2 will not execute; */ statement3 will execute is there a similar mechanism in python, other than prefixing the '#' character to the start of each statement i do not want to execute (which gets old very quickly if one needs to comment and uncomment several statements a couple of times while playing around with code say during initial design)? IMO this is a missing feature in Python. However, if the block of code you are wanting to comment out doesn't happen to contain any triple-quotes, you can surround the code with those. For example: def myFunc(arg): return arg + 2 def myFunc(arg): return arg + 1 print myFunc(2) 4 -- Paul -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
john peter wrote: is there a similar mechanism in python, other than prefixing the '#' character to the start of each statement i do not want to execute (which gets old very quickly if one needs to comment and uncomment several statements a couple of times while playing around with code say during initial design)? In addition to the approach Paul has suggested, you should also consider using a better editor, as most decent editors should have some kind of block comment/uncomment feature that prevents that operation from getting old very quickly... Scite, for example, lets me selected a block and hit Ctrl-Q to either comment or uncomment the block. (It does this by prefixing each line with #~ instead of just #, which allows it to detect when a line is already so commented and reverse the operation.) Other decent editors have their own approach. -Peter -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
On Fri, 17 Feb 2006 19:12:01 -0500, Peter Hansen wrote: john peter wrote: is there a similar mechanism in python, other than prefixing the '#' character to the start of each statement i do not want to execute (which gets old very quickly if one needs to comment and uncomment several statements a couple of times while playing around with code say during initial design)? In addition to the approach Paul has suggested, you should also consider using a better editor, as most decent editors should have some kind of block comment/uncomment feature that prevents that operation from getting old very quickly... Scite, for example, lets me selected a block and hit Ctrl-Q to either comment or uncomment the block. I see the developers of Scite are either newbies to GUI programming, or they operate in a world of their own. Ctrl-Q is reserved for Quit (or Exit) in every GUI API I know of. (It does this by prefixing each line with #~ instead of just #, which allows it to detect when a line is already so commented and reverse the operation.) It is *easy* to detect when a line is already commented. It starts with a #. The ~ is superfluous. Commenting and uncommenting should be two different commands: the whole point of nested comments is that it allows you to comment a block of text which may already contain comments. Having one command do both commenting and uncommenting according to the presence or absence of semantic clues in the text is a recipe for failure (No you stupid computer, I want to COMMENT that block, not uncomment it!!!). Imagine if your text editor used cntl-C for both copy and paste, somehow guessing whether you wanted to copy selected text or paste over it according to some subtle clue in the text itself. Wouldn't that be fun? -- Steven. -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
On Sat, 18 Feb 2006 12:36:06 +1100, Steven D'Aprano [EMAIL PROTECTED] wrote: [snip] Commenting and uncommenting should be two different commands: the whole point of nested comments is that it allows you to comment a block of text which may already contain comments. Having one command do both commenting and uncommenting according to the presence or absence of semantic clues in the text is a recipe for failure (No you stupid computer, I want to COMMENT that block, not uncomment it!!!). Imagine if your text editor used cntl-C for both copy and paste, somehow guessing whether you wanted to copy selected text or paste over it according to some subtle clue in the text itself. Wouldn't that be fun? emacs' comment-dwim (M-;) has never guessed incorrectly, in my experience. Of course, if it ever does, I can always invoke comment-region or uncomment-region, as appropriate. -- Steven. -- http://mail.python.org/mailman/listinfo/python-list -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
In article [EMAIL PROTECTED], Paul McNett [EMAIL PROTECTED] wrote: . . . IMO this is a missing feature in Python. However, if the block of code you are wanting to comment out doesn't happen to contain any triple-quotes, you can surround the code with those. For example: def myFunc(arg): return arg + 2 def myFunc(arg): return arg + 1 print myFunc(2) 4 . . . ... and note that, even if the block *does* contain triple-quotes of one flavor, you can use the other: '''This is syntactic data: . ''' -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
Hi Steven, I see the developers of Scite are either newbies to GUI programming, or they operate in a world of their own. Ctrl-Q is reserved for Quit (or Exit) in every GUI API I know of. I am the developer responsible for SciTE and the Ctrl+Q decision. Ctrl+Q is not reserved for Quit on every platform. Perhaps you have never used Windows where Exit does not normally have a distinct command key of its own but can be performed with Alt+F4 (Close) for single window applications. Since exiting an application is a far less common action than modifying the data within that application, it doesn't deserve one of the more easily typed key bindings. Some applications use Ctrl+Q but, of the applications I currently have open, Firefox, Internet Explorer, Explorer, iTunes, Visual Studio .NET, WordPad, QuickTime, PythonWin, and WinMerge don't respond to Ctrl+Q by quitting and Thunderbird does. Windows guidelines: http://msdn.microsoft.com/library/en-us/dnwue/html/appxB.asp It is *easy* to detect when a line is already commented. It starts with a #. The ~ is superfluous. It is not usual to change a line from being code to being a comment as most lines of code make no sense as English text. If you do sometimes want to do this, the # key can be pressed at the beginning of a line. Ctrl+Q exists to allow commenting out sections of code that you do not want to be active. The ~ acts to differentiate these different uses of one language feature and makes it easy to apply the operation over ranges that include comments and then invert the operation. Commenting and uncommenting should be two different commands: the whole point of nested comments is that it allows you to comment a block of text which may already contain comments. Having one command do both commenting and uncommenting according to the presence or absence of semantic clues in the text is a recipe for failure (No you stupid computer, I want to COMMENT that block, not uncomment it!!!). Comments and commented out code are distinct. Using the same marker for both is ambiguous. Few languages support true nested comments. SciTE's implementation of Ctrl+Q is very popular although you can change the text inserted/removed from #~ to # if you prefer and SciTE also includes a command to add stream style comments for languages like /* C++ */. Imagine if your text editor used cntl-C for both copy and paste, somehow guessing whether you wanted to copy selected text or paste over it according to some subtle clue in the text itself. Wouldn't that be fun? Your imaginary command makes no sense as it is described. What are its full semantics? Why was it implemented this way? BTW, I have been implementing GUIs for over 20 years so maybe me and the other SciTE users are in our own world. It does feel like a friendly, enjoyable, comfortable world, though. Neil -- http://mail.python.org/mailman/listinfo/python-list
Re: commenting out blocks of code
Neil Hodgson wrote: Hi Steven, It is *easy* to detect when a line is already commented. It starts with a #. The ~ is superfluous. It is not usual to change a line from being code to being a comment as most lines of code make no sense as English text. If you do sometimes want to do this, the # key can be pressed at the beginning of a line. Ctrl+Q exists to allow commenting out sections of code that you do not want to be active. The ~ acts to differentiate these different uses of one language feature and makes it easy to apply the operation over ranges that include comments and then invert the operation. hi. here is something you both seems to have not considered: imagine you make decision if ^Q has to comment or uncomment based on the 1st line and not on each line individually in the block. so let's say i mark and ^Q this: -- code1 #comment1 #comment2 code2 1st line is not commented, so i want to comment-out the whole blcok, the result is -- #code1 ##comment1 ##comment2 #code2 note how ## maintains where comments were. now, for the same selection, ^Q again? 1st character is #, so the editor is asked to uncomment, dwim: -- code1 #comment1 #comment2 code2 so there is a way to do it without #~ noise... now if we can only think of how to do DWIM copypaste :-)! - nas -- http://mail.python.org/mailman/listinfo/python-list
