Re^4: Escaping multiple escape chars

Perhaps this example helps better what I am trying to say that context matters ( ' vs " ):

use Devel::Peek;

my $txt = 'some\text';

print $txt;  #prints some\text
Dump($txt);
[download]

versus:

use Devel::Peek;

my $txt = "some\text";

print $txt; #prints some [tab] ext
Dump($txt);
[download]

If what you were saying was true, then the first line should print the same as the second.
The substitution:

use Devel::Peek;

my $txt = '\|/';
$txt =~s/\\/\\\\/g;

print $txt;
Dump($txt);
[download]

Produces the exact same results as:

use Devel::Peek;

my $txt = '\\|/';
$txt =~s/\\/\\\\/g;

print $txt;
Dump($txt);
[download]

Aren't they different literals? Change the ' to " and try the same test.
I think the logic is broken here. A literal should mean, I am to be taken literally, not, interpret part of me as a literal and part of me however you take a guess, because that is what is happening or else it would store 'some\text' the same as "some\text". The parser is just getting this case incorrect I think.

Comment on Re^4: Escaping multiple escape chars Select or Download Code

Replies are listed 'Best First'.
Re^5: Escaping multiple escape chars by ikegami (Patriarch) on Dec 10, 2005 at 14:22 UTC
' and " are different. All of this is documented. A literal should mean, I am to be taken literally Nonsense. Literals are always interpreted. The literal `1234` in the expression `$a = 1234;` is a group of characters, yet they are interpreted as a number. And somehome, the quotes in the literal `"abc"` are removed. Without interpretation, the compiler wouldn't know where the string ends and/or wouldn't allow certain characters to be part of the string. For example, what if there was a quote in your ASCII art? How would Perl know the string doesn't end at that quote, but rather the following one? (or the one after that?) For single quoted strings, you preceed the quote with a backslash. Of course, now we need a method of allowing backslashes followed by single quotes... Whenever something is embedded in something else, be it a string in a Perl source file, a object in a data file or text between HTML tags, some form of encoding or escaping is required. To be crystal clear: You can't have strings and Perl code in the same file without some form of escaping or encoding.	[reply] [d/l] [select]
Re^6: Escaping multiple escape chars by JamesNC (Chaplain) on Dec 10, 2005 at 16:29 UTC
I have attempted to explain my question with 2 examples using 'some\text' and "some\text" both of which get stored as I understood they should. However, 'some\\text' and "some\\text" don't behave as you would expect and are NOT stored with the same rules as the first case because in the first case, the compiler sees the single quotes and stores the characters correctly by pushing an additional '\' to the PV array as it should. What is going wrong, however, is the case when it sees a second '\' in the literal case, and doesn't push another '\' onto the PV. use Devel::Peek; # literal case - PV's are the same $txt = 'some\text'; print "Case 1 'some\\text': $txt\n-------\n"; # PV gets an extra '\' pushed on it. Dump($txt); $txt = 'some\\text'; print "\nCase 2 'some\\\\text: $txt\n-------\n"; #PV does NOT get an extra '\' pushed on to it Dump($txt); # quoted case - PV's handle '\' as expected $txt = "some\text"; print "\nCase 3 \"some\\text\": $txt\n-------\n"; Dump($txt); # $txt = "some\\text"; print "\nCase 4 \"some\\\\text\": $txt\n-------\n"; Dump($txt); [download] If you actually try these examples, then you can see for yourself how the scalars are getting stored. Nuff said. I will ask the perl porters from here.	[reply] [d/l]
Re^7: Escaping multiple escape chars by ikegami (Patriarch) on Dec 12, 2005 at 15:50 UTC
However, `'some\\text'` and `"some\\text"` don't behave as you would expect I got the following (abbreiviated) results: `Case 1 'some\text': some\text ------- PV = 0x1aa01f0 "some\\text"\0 Case 2 'some\\text: some\text ------- PV = 0x1aa01f0 "some\\text"\0 Case 3 "some\text": some ext ------- PV = 0x1aa01f0 "some\text"\0 Case 4 "some\\text": some\text ------- PV = 0x1aa01f0 "some\\text"\0` [download] If you obtained the above, `perl` behaved as expected (or at least as documented). According to to the relevant docs, double-quoted string literals are subject to backslash and variable substitution; single-quoted strings are not (except for `\'` and `\\`). In single quotes, `perl` interprets the backslash literally (instead of an escape) unless followed by another backslash or by a single quote. This DWIM most of the time, since few characters ever need to be escaped in single-quoted strings. For example, `'\n'` results in the string {backslash}+{lowercase n}. In double quotes, `perl` interprets the backslash and the following character as an escape (instead of literally). How `perl` interprets an invalid/unrecognized escape sequence is undocumented (or unclearly documented), but the behaviour has consistently been to replace the escape sequence with the character that followed the backslash. This DWIM most of the time, since the extra backslash was probably a case of overzealous escaping. For example, `"\<"` results in the string {right angle bracket}. If you don't see the necessity of the equivalency of `"some\\text"` and `'some\\text'` try answering the following questions: How would you code the string {backslash}+{backslash} using single quotes? How would you code the string {backslash}+{backslash} using double quotes? How would you code the string {backslash}+{backslash}+{single quote} using single quotes? How would you code the string {backslash}+{backslash}+{double quote} using double quotes? How would you code the string {backslash}+{single quote} using single quotes? How would you code the string {backslash}+{double quote} using double quotes?	[reply] [d/l] [select]
Re^8: Escaping multiple escape chars by JamesNC (Chaplain) on Dec 12, 2005 at 17:03 UTC


We don't bite newbies here... much
	PerlMonks