This is not because of fovalues or foreach. See help comments:
"However, if the // indicator is at the end of a line, it must be
preceded by one or more blanks."
This is quite a unique requirement of Stata. I can't come up with any
other system which would demand anything like that. Adding a space in
the beginning of every line should not change the program, but in this
case in does (and helps to fix the error).
The fact that it works when the delimiter is CR is hence inconsistent
with the documented behavior.
Best regards, Sergiy Radyakin
On Thu, Oct 2, 2008 at 3:27 PM, Steven Samuels <[email protected]> wrote:
> Here's another example where * works and // doesn't. The circumstances: the
> delimiter is ";", and the code block contains a -forvalues- or -foreach-
> section. /* and */ around the -foreach- section also works.
>
> -Steve
>
> *****************CODE BEGINS****************
> sysuse auto, clear
> forvalues i=1/5 {
> di "rep78 = " `i'
> }
>
> #delim ;
>
> forvalues i=1/5 { ;
> di "rep78 = " `i';
> };
> ;
>
> ***************** (*) COMMENT WORKS ****************
> * forvalues i=1/5 { ;
> * di "rep78 = " `i';
> * };
> * ;
>
> ***************** // COMMENT FAILS****************
> // forvalues i=1/5 { ;
> // di "rep78 = " `i';
> // };
> // ;
> *****************CODE ENDS****************
>
> On Oct 2, 2008, at 12:44 PM, Nick Cox wrote:
>
>> Very good catch, Sergiy. Thanks for the correction! I should have
>> checked about the //.
>> Conversely, this is another argument for the deprecated *.
>>
>> Nick
>> [email protected]
>>
>> Sergiy Radyakin
>>
>> Nick,
>>
>> local c //
>> will not work in the program because "//" is interpreted as the start
>> of the comment
>> it works in the command line though, because "//"-style comments are
>> prohibited in the command line input. To fix this:
>> 1) open quotes before // like so: local c "//" , or
>> 2) write it down as an expression, e.g. local c="/"+"/"
>> In any case, Stata fails to recognize "//" as a comment when it comes
>> as a result of macro expansion. But it works for a star ("*").
>>
>> // --- begin of file testc.do ---
>> local c //
>> display `"`c'"'
>> local c ="/"+"/"
>> display `"`c'"'
>> local c "//"
>> display `"`c'"'
>> local c "*"
>> display `"`c'"'
>> `c' just nothing
>> local c "//"
>> `c' just nothing
>> // --- end of file testc.do ---
>>
>> If writing out debugging messages, it may be more comfortable to use
>> di_g. See: help di_g.
>>
>> Best regards, Sergiy Radyakin
>>
>> On Tue, Sep 30, 2008 at 11:28 AM, Nick Cox <[email protected]> wrote:
>>>
>>> Another is to prefix lines by e.g.
>>>
>>> `c'
>>>
>>> Setting
>>>
>>> local c *
>>>
>>> or
>>>
>>> local c //
>>>
>>> comments out those lines.
>>>
>>> Setting
>>>
>>> local c
>>>
>>> zaps the macro.
>>>
>>> Naturally any definition of the local as "*" or "//" must precede all
>>
>> uses
>>>
>>> that require that interpretation.
>>>
>>> The converse does not apply: if the local is not defined, Stata sees
>>
>> the
>>>
>>> line.
>>>
>>> More generally, you can define or undefine the macro at will.
>>>
>>> I don't especially recommend this method, but it is an alternative
>>
>> that is
>>>
>>> sometimes useful, and it matches practices in many programming
>>
>> languages.
>>>
>>> Nick
>>> [email protected]
>>>
>>>
>>> Maarten buis wrote:
>>>>
>>>> --- Paul O'Brien <[email protected]> wrote:
>>>>
>>>>> at times i need to comment out a number of lines, usually with * and
>>>>> remove the * later to activate the commands.
>>>>> is there a simple way to remove and add the * to a number of line at
>>
>> one
>>>>>
>>>>> go?
>>>>
>>>> One way to comment out a block of code is to surround them with /*
>>
>> and
>>>>
>>>> */ this way you have to remove only these two.
>>
>> *
>> * For searches and help try:
>> * http://www.stata.com/help.cgi?search
>> * http://www.stata.com/support/statalist/faq
>> * http://www.ats.ucla.edu/stat/stata/
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
