Prototypes: the good, the bad,and the ugly

This is a reply to chromatic's post The Problem with Prototypes, but his blog didn't allow me to post it there so I post it here.

Pretty much everyone agrees that there are good (such as blocks and sometimes reference) prototypes and bad ones (scalar most of all). Few discuss the third class: the ugly glob prototype.

perlsub describes them as such:

A * allows the subroutine to accept a bareword, constant, scalar expression, typeglob, or a reference to a typeglob in that slot. The value will be available to the subroutine either as a simple scalar, or (in the latter two cases) as a reference to the typeglob.

In other words, they are the same as scalar prototypes, except that they also accept globs and barewords. This is mainly used to pass filehandles, like this:

sub foo (*) {...}

foo STDIN;

but in fact it can be used to pass any bareword to function, as it leaves the interpretation of it to the function.

It's tempting to call this bad, but it offers some API possibilities that would otherwise not be possible, hence I would call it ugly rather than bad per se.

Casting magic against segfaults

The problem

For years, there has been the Sys::Mmap module, however, it has a few issues. For example, let's take this piece of code:

use Sys::Mmap; open my $fd, '+>', 'filename'; mmap my $var, -s $fd, PROT_READ|PROT_WRITE, MAP_SHARED, $fd, 0; $var = 'Foobar'; munmap $var;

First of all, it's simply not user-friendly. mmap takes 6 arguments in a weird order, and uses weird constants. Also munmap shouldn't be necessary: variables should dispose of themselves when they run out of scope.

But more importantly, this program does not do what you think it does, though the only hint of that is an Invalid argument exception when doing munmap. During the assignment, the link between the mapping and the variable is lost, so nothing is written to the file. Worse yet, this can even lead to a segfault in some circumstances.

Ouch!

Tying things up?

The documentation clearly says that you shouldn't do this (or anything else that changes the length of the variable), but IMHO this hole shouldn't be left open in the first place, if only because it is extremely counterintuitive (and thus a maintenance nightmare). Modules should fail more gracefully than this.

Sys::Mmap offers a tied interface as compensation, but this didn't work out. The tied interface indeed is safe, but it creates another problem.

Every time it is read, it copies the whole file into the variable. Every time the variable is modified, it writes the whole new value to the file, even if the change only affects a single byte.

Ouch!

Obviously, that doesn't scale at all. One user of the module reported a 10-20 times slowdown of his program after converting to ties. That's not a workable solution.

The solution

Perl has a powerful but rarely used feature called magic. (It's rare use by module authors is indicated by the fact that the prototypes of the magic virtual table as documented in perlguts aren't even complete: they lack pTHX_'s). They are used by the perl core to implement magic variables such as $! and ties (surprise, surprise). It offers 8 hooks into different stages of handling a variable, the three most important being fetching(svt_get), storing(svt_set) and destruction(svt_free).

In my case, I didn't need get magic, but I did use set and free magic. Freeing the variable is not that interesting (simply unmapping the variable), but setting it is. This function is called just after every write to the variable:

static int mmap_write(pTHX_ SV* var, MAGIC* magic) {     struct mmap_info* info = (struct mmap_info*) magic->mg_ptr;     if (SvPVX(var) != info->address) {         if (ckWARN(WARN_SUBSTR))             warn("Writing directly to a to a memory mapped file is not recommended");         Copy(SvPVX(var), info->address, MIN(SvLEN(var) - 1, info->length), char);         SvPV_free(var);         reset_var(var, info);     }     return 0; }

This function is called after every write to the variable to check if the variable is still linked to the map. If it isn't, it copies the new value into the map, frees the old value and restores the link. As copying is potentially expensive, it will issue a warning if warnings (or actually, 'substr' warnings) is in effect.

There is no perfect solution to this problem, but getting a friendly warning is undeniably better than getting a segmentation fault or data loss.

Anyway, you can find Sys::Mmap::Simple here. It offers more goodies, such as portability to Windows, a greatly simplified interface, and built-in thread synchronization.

Elegance in minimalism

Some time ago, I read this journal entry by Aristotle. I liked it and suspected it could be easily implemented in XS. It turned out to be the most elegant piece of XS I've ever written.

void induce(block, var)     SV* block;     SV* var;     PROTOTYPE: &$     PPCODE:         SAVESPTR(DEFSV);         DEFSV = sv_mortalcopy(var);         while (SvOK(DEFSV)) {             PUSHMARK(SP);             call_sv(block, G_ARRAY);             SPAGAIN;         }

I assume most readers don't know much C, let alone the perl API or XS, so I'll explain it piece by piece.

void induce(block, var)     SV* block;     SV* var;     PROTOTYPE: &$
This declares the xsub. It has two parameters, both scalar values. The function has the prototype &$. So far little surprises.

    PPCODE:
This declares that a piece of code follows. Unlike CODE blocks, PPCODE blocks pop the arguments off the stack at the start. This turns out to be important later on.

        SAVESPTR(DEFSV);         DEFSV = sv_mortalcopy(var);
These lines localizes $_ and initializes it to var.

        while (SvOK(DEFSV)) {
This line is equivalent to while (defined($_)).

Now comes the interesting part:
            PUSHMARK(SP);             call_sv(block, G_ARRAY);             SPAGAIN;
To understand what this block does, you have to know how perl works inside.

If you've read perlxs, you may notice this function does not push any values on the stack. Are careless reader might be mistaken and think this function doesn't return anything: they couldn't be more wrong!

If you've read perlcall, you would notice a lot more is missing. For starters, the function calls SPAGAIN (pretty much equivalent to saying I accept the values you return to me), but it doesn't do anything with them.
Also, you may notive that both ENTER/LEAVE(needed to delocalize $_) and SAVETMPS/FREETMPS (needed to get rid of temporary values) are missing. The function that calls the xsub automatically surrounds it by an ENTER/LEAVE pair, so that one isn't necessary. The lack of SAVETMPS/FREETMP however is not only deliberate but also essential.

The loop calls the block without arguments (PUSHMARK & call_sv). The xsub accept the return values on the stack and leaves them there. This sequence repeated. This way it assembles the induces values on the stack. PPCODE removing the arguments at the start prevents it from returning those as first two return values. It also adds a trailer that causes all elements that have been pushed on the stack to be recognized as return values of this function. That's why a SAVETMPS/FREETMPS pair would break this code: the values must live after the code returns.

That's the elegance of this function. It doesn't even touch it's return values, it delegates everyting to the block. All the things that are missing make that it does exactly what it should do.

Where java stopped

Yesterday I explained my problems with garbage collection. I don't think garbage collection is bad or any, I just think it isn't being used properly. GC was invented in the late 50's for LISP, the first of high level programming languages. Lambda calculus required that the memory is managed by the system. Having freed the programmer from memory management, Lisp and is brethren enabled the development of true high level features such as dynamic typing, higher-order functions, closures, macros and continuations. These are exactly the feature that give those languages their incredible power.

That is what doesn't make sense in Java and derivatives: those really powerful features are missing in Java. Java is mostly lacking the features that absolutely require a GC. Guy Steele once said "We were after the C++ programmers. We managed to drag a lot of them about halfway to Lisp". Considering the conservativeness of the industry it's understandable they stopped halfway (the step from C to C++ was even smaller). That kind of makes Java a middle level language; a watery compromise that fails to offer the best of both worlds.

Both high and low level languages have their niches, but what about the middle level languages? My intuition tells me their proper niche should be way smaller. It's hard to say what will be Java's successor, but I'm pretty certain of two things. It will not be a Java derivative and it will take the second half step, if not more.

Garbage collection revisited

Garbage collection must be one of the most misunderstood features of programming languages. GC has existed for 50 years, yet a lot of languages have not adopted it. One of the most important commonly cited advantaged of Java and similar languages over languages such as C++ is garbage collection. Having used a number of Java applications I'd dare to say most have memory problems though. They feel extremely bloated. I've seen a few programs improve drastically when some expert started to optimize the memory usage. Most Java programs can be lean enough to be usable, but it will take a lot of effort. I imagine it is a major disillusion for a lot of Java programmers to find out they haven't found a memory panacea after all.

One could see all variables and their resources as a directed graph. In the most elementary programs this graph will be a tree and in such cases manual memory management is trivial. Real programs aren't this simple. Having said that, most programs are not random or unpredictable. Usually substantial pieces of the graph are trees on their own. Acyclical graphs and even most cyclical ones can be solved using reference counting and similar techniques. If you understand the resource usage pattern of your program, you can usually solve it without resorting to garbage collection. However this often requires planning and thinking ahead of time. A GC on the other hand is able to manage any graph, mostly without help from the programmer, but does not guarantee to do so efficiently. So the question when do you need a GC? reduces to when can't I know the pattern? To date, I've only come across one concrete and common example that really can not be solved semi-manually: programming languages themselves. The reason for this is obvious: it is inherent to them that you don't know ahead how they will be used.

It is undisputed that GCs make it easier to write programs, but this issue makes me wonder if it also is easier to write good programs. Java programmers are happy to think they don't have to think about memory, but it turns out they will have to think ahead (though less than C or C++) if they want decent performance. On the server side that doesn't really matter that much, but in client-side programs it does.

In the end, it is a matter of trade-offs. Both approaches can solve most problems. I don't believe a GC by itself gives us better programs than RAII management. Having said that, garbage collection makes business sense in a LOT of situations. Companies don't earn money by building the best program they could, they earn money by selling a finished program, good or not.

Operators and types

Ruby and Python overload the + operator for a large number of things, the most common ones being addition of numbers, concatenation of strings, and concatenation of tuples. Very different things are represented by the same syntax. In Perl these three roles are occupied by three different operators (+,. and ,). For that matter almost all operators for numbers are separated from operators on strings in Perl. This causes is one of the most common misconception among non-natives about Perl's typing system: it's not weakly typed. This piece of code:

my $foo = "1";
return $foo + 0;

does not cause an implicit type conversion, nor is the last statement in any way ambiguous. The addition operator causes an explicit conversion of its argument. It's not an implicit one for a simple reason: it's the Perl idiom for converting any variable to a number.

I think this is an excellent example of the waterbed theory of complexity. To reduce the number of operators in the language Python and Ruby use runtime polymorphism on data whose behavior is already known to the programmer (not to the compiler) at compile time. I cannot think of real-world code where you don't know if your variable is a number, a string or a tuple but want to do addition/concatenation nonetheless. It is trading semantic clarity for syntactic clarity. It's a valid choice, just as Perl's choice of separating them. I think a lot of rubyists and pythonistas fail to see that their choice has its disadvantages too.

Control structures in Perl

As I said in my previous entry, there is a need for education in social coding skills in Perl. Therefor I'll put my money where my mouth is ;-)

Basically there are four patterns for branching in Perl.

1. Conditional statements: if(condition) { true_action }
else { false_action }
2. Statement modifiers action if condition;
3. Logical operators: condition and action
4. Ternary operator: condition ? true_action : false_action

How to decide which one to use? That depends on the situation of course. Only conditional statements and the ternary operator can provide an else clause. If you need one, your choices are already limited. All four put their emphasis differently. For example print $line unless $line =~ /^#/; may be better than
$line =~ /^#/ or print $line; Because the principles of prominence. This is a linguistic notion that tells us that people tend to prefer important things to be in front and details to be in the end so they can skip over the details when scanning the code. When skipping the second part the code still makes sense (even though it may not be correct anymore). In general statement modifiers are best used when the action is much more important than the condition.

Logical operators are useful in two situations. The first one is exactly the opposite of the statement modifier: When the condition is vastly more important than the action. This usage typically uses the low precedence version. For example: open my $filehandle, $filename or die "Can't open $filename: $!"; is better than die "Can't open $filename: $!" unless open my $filehandle, $filename; because the first one communicates the intend of the programmer (opening a file) better. Error handling is not important for understanding the big picture of the code.

It has another trait that is important for deciding when to use this pattern. Unlike the previous two patterns, logical operators are expressions not statements. As such they can be used in places where the former can not. parse($filename || "default.filename") is significantly easier to read than if(!$filename) { $filename = "default.filename"; } parse($filename);

Similarly my $id; if($input >= 0) { $id = $input; } else { $id = 1; } can be simplified using the ternary operator to my $id = $input >= 0 ? $input : 1

You may wonder now, when should I use old fashioned conditional statements then? First of all, if the action contains multiple statements and isn't suitable for putting in a function. It puts an equal emphasis on the condition and the action. It makes sense to use this pattern if you don't have a reason to do otherwise.

Summary
	Conditional statements	Statement modifiers	Logical operators	Ternary operator
Emphasis	None	Action	Condition	Condition
Expression	No	No	Yes	Yes
Else clause	Yes	No	No	Yes
Nestable/Chainable	Yes, very well	No	Yes	Yes
Multiple statements	Yes	No	Yes	No