Difference between revisions of "DIP83"

From D Wiki
Jump to: navigation, search
 
(122 intermediate revisions by 2 users not shown)
Line 25: Line 25:
  
 
== Abstract ==
 
== Abstract ==
Allow for assert's to do pretty printing of its failing expression. Printing is configurable via specific sets of (template) function overloads.
+
Allow for '''assert''' to do pretty printing of its failing expression when flagged for in call to compiler. Printing is configurable via specific sets of (template) function overloads.
  
 
== Rationale ==
 
== Rationale ==
  
Typically added/modified by the developer by changing the flags to calls to rdmd/dub/scons etc upon failure with `-unittest` for a specific failing module. This will be more convenient than explicitly adding the prints of `lhs` and `rhs` in the failing unittest.
+
A failing assert (in a unittest), currently, give no hint about why it failed. To aid the developer in debugging, the failing assert of a
 +
 
 +
* binary expression, such as '''assert(x == y)''', should print the values of '''x''' and '''y''' and
 +
* unary expression, such as '''assert(!x)''', should print the value of '''x'''.
 +
 
 +
This extra, so called, pretty printing can be enabled by changing the dmd flag '''-unittest''' to, say, '''-unittest=verbose''' for a specific failing module. This will be more convenient than explicitly adding the prints of the left-hand-side expression '''lhs''' and right-hand-side expression '''rhs''' directly before the call to the failing assert.
  
 
== Description ==
 
== Description ==
 +
 
This DIP proposes to add library-level-configurable diagnostics to failing calls to '''assert(expr)''' typically called from within '''unittest'''-blocks.
 
This DIP proposes to add library-level-configurable diagnostics to failing calls to '''assert(expr)''' typically called from within '''unittest'''-blocks.
  
This diagnostics is activated only when DMD is called with a specific command line flag, say `-unittest=verbose`.
+
This diagnostics is activated only when DMD is called with a specific command line flag, say '''-unittest=verbose''' or perhaps '''-diagnose=assert'''.
 +
 
 +
== Modifications needed in Compiler Frontend ==
 +
 
 +
=== Binary Operator Lowering ===
 +
 
 +
If DMD is called with this flag it will rewrite (lower) all assert expressions ('''AssertExpr''' in dmd source) such as
 +
 
 +
<syntaxhighlight lang="d">
 +
assert(a BINOP b)
 +
</syntaxhighlight>
 +
 
 +
into
 +
 
 +
<syntaxhighlight lang="d">
 +
(auto ref a, auto ref b) {
 +
    if (a BINOP b) return;
 +
    onAssertFailed!"BINOP"(a, b, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
 +
} (e1, e2)
 +
</syntaxhighlight>
 +
 
 +
where '''onAssertFailed''', in this binary operator case, is declared as
 +
 
 +
<syntaxhighlight lang="d">
 +
void onAssertFailed(string op, E1, E2)(E1 e1, E2 e2, string file, uint line, string function, string module);
 +
</syntaxhighlight>
 +
 
 +
=== Unary Operator Lowering ===
  
If DMD is called with this flag it will then rewrite the AST to replace calls to
+
Similarly, for unary expressions, rewrite (lower)
  
```D
+
<syntaxhighlight lang="d">
assert(lhs == rhs)
+
assert(UNOP e)
```
+
</syntaxhighlight>
  
 
with
 
with
  
```D
+
<syntaxhighlight lang="d">
assertBinOp(A_FILE, A_LINE, A_COLUMN, "==")(lhs == rhs, lhs, rhs)
+
(auto ref e) {
```
+
    if (UNOP e) return;
 +
    onAssertFailed!"UNOP"(e, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
 +
} (e)
 +
</syntaxhighlight>
 +
 
 +
where on onAssertFailed, in this unary operator case, is declared as
 +
 
 +
<syntaxhighlight lang="d">
 +
void onAssertFailed(string op, E)(E e, string file, uint line, string function, string module);
 +
</syntaxhighlight>
 +
 
 +
=== Non-Operator Lowering ===
  
where
+
For the case when no unary or binary operator is present in the top assert expression, unary overload of '''onAssertFailed''' is called with '''op''' being empty string. In other words the '''assert''' expression
  
```D
+
<syntaxhighlight lang="d">
assertBinOp(string file, uint line, uint column, string op)(E, L, R)(lazy E expression, lazy L lhs, lazy R rhs)
+
assert(e)
```
+
</syntaxhighlight>
and similarly for unary expressions.
 
  
The default implementation of `assertBinOp` (preferrably defined somewhere in `druntime`) would be some standard D code that mimics the current behaviour of `assert(expression)` by throwing an `AssertExpression` if `cast(bool)expression` is `false` (dont know about the behaviour of nothrow/@nogc discussed above though). The Phobos-developer can then do what he likes with the information he needs in the extra arguments in specific templated overloads of `assertBinOp`.
+
is rewritten (lowered) into
  
This specific behaviour could be extendable by adding (typically templated) overloads of `assertBinOp` for specific sets of types (concepts).
+
<syntaxhighlight lang="d">
 +
(auto ref e) {
 +
    if (e) return;
 +
    onAssertFailed!""(e, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
 +
} (e)
 +
</syntaxhighlight>
  
Then we could get the extendability we want in testing-frameworks such as this without adding a new `assert`-overload-set and without sacrifycing default memory usage in DMD/Phobos unittests.
+
If the assert lowering must also be sensitive to expressions like
  
Further with this solution we could add cool diagnostics behaviour in `assertBinOp` for failing array/range comparisons aswell in a format. This diagnostics could even have different pretty printing backends such as HTML.
+
=== Non-Equality Operator Lowering ===
 +
 
 +
Further note that operator '''!=''' may need special care because D only supports overloading via '''opUnary!"!"''' and '''opBinary!"=="'''. This because the D compiler rewrites
 +
 
 +
<syntaxhighlight lang="d">
 +
x != y
 +
</syntaxhighlight>
 +
 
 +
into
 +
 
 +
<syntaxhighlight lang="d">
 +
!(x == y)
 +
</syntaxhighlight>
 +
 +
To make assert diagnostics as configurable as possible it may become relevant to enable the possibility for the developer to to define the overload '''onAssertFailed!"!="''' aswell. If this is desirable the lowering logic in DMD must be sensitive to expressions
 +
 
 +
<syntaxhighlight lang="d">
 +
assert(!(a == b))
 +
</syntaxhighlight>
 +
 
 +
which, preferrably, will be rewritten to
 +
 
 +
<syntaxhighlight lang="d">
 +
(auto ref a, auto ref b) {
 +
    if (!(a == b)) return;
 +
    onAssertFailed!"!="(a, b, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
 +
} (e1, e2)
 +
</syntaxhighlight>
 +
 
 +
== Modifications needed in D Runtime ==
 +
 
 +
=== Configuring Diagnostics ===
 +
 
 +
The default implementations (in druntime) of all the '''onAssertFailed'''-overloads should, as a first step, by defined to exactly mimic the current behaviour of '''assert()'''. This because, memory requirements of the call to DMD will increase when the compiler must generate terminal printing-code of all the arguments for all the unittests found in the compilation unit of interest. And it is currently unclear if this, current increase in system requirements, will be compatible with system resources currently available in build servers connected to GitHub.
 +
 
 +
Specific printing behaviour of '''assert()''' diagnostics can then be extendable by adding (typically templated) overloads of '''onAssertFailed''' for specific sets of types (concepts).
 +
 
 +
For instance, diagnostics specifically when comparing arithmetic types, could be realized through the overload
 +
 
 +
<syntaxhighlight lang="d">
 +
import std.traits : isArithmetic;
 +
void onAssertFailed(string op, E1, E2)(E1 e1, E2 e2, string file, uint line, string function, string module)
 +
    if (isArithmetic!L &&
 +
        isArithmetic!R)
 +
{
 +
    version(assert)
 +
    {
 +
        import core.exception : AssertError;
 +
        throw AssertError("Failed arithmetic assert: " ~ e1 ~ " " ~ binOp ~ " " ~ e2);
 +
    }
 +
}
 +
</syntaxhighlight>
 +
 
 +
In this way D would get the extendability we want in testing-frameworks such as '''std.experimental.testing''' (https://github.com/D-Programming-Language/phobos/pull/3207) without adding a new '''assert'''-overload-set and without sacrifycing default memory usage in DMD/Phobos unittests.
 +
 
 +
Further, this solution enables the possibility to provide fancy diagnostics behaviour in '''onAssertFailed''' for failing '''array'''/'''range''' or aggregate ('''struct''' or '''class''') comparisons. This diagnostics could also have different pretty printing backends such as HTML.
 +
 
 +
== Examples of Possible Diagnostics Outputs ==
 +
 
 +
=== Short Array Assert Arguments ===
  
 
For example a failing
 
For example a failing
  
```D
+
<syntaxhighlight lang="d">
 
assert([1,2,3] == [1,2,4]);
 
assert([1,2,3] == [1,2,4]);
```
+
</syntaxhighlight>
  
could fancy-print
+
could pretty-print
  
```
+
<syntaxhighlight lang="d">
([1,2,3][2] being 3) != ([1,2,4][2] being 4)
+
([1,2,3][2] is 3) != ([1,2,4][2] is 4)
```
+
</syntaxhighlight>
  
or for aggregates a failing
+
=== Aggregate Assert Arguments ===
  
```D
+
or, for aggregates, a failing
 +
 
 +
<syntaxhighlight lang="d">
 
struct A { int x, y; }
 
struct A { int x, y; }
 
auto a = A(1,2);
 
auto a = A(1,2);
 
auto b = A(1,3);
 
auto b = A(1,3);
 
assert(a == b);
 
assert(a == b);
```
+
</syntaxhighlight>
  
could fancy-print
+
could pretty-print
  
```
+
<syntaxhighlight lang="d">
(a.y being 2) != (b.y being 3)
+
(a.y is 2) != (b.y is 3)
```
+
</syntaxhighlight>
  
Thereby giving D yet another feature, I've not seen in any other language :)
+
=== Large Array Assert Arguments ===
  
If needed I'd be more than happy to help out implementing this solution in dmd/druntime/phobos. I'm guessing that parts of the solution list at
+
or, for very large arrays, a failing
 +
 
 +
<syntaxhighlight lang="d">
 +
const n = 1_000_000;
 +
auto a = iota(0, n).array;
 +
auto b = a.dup;
 +
a[1_000 .. 1_002] = 0;
 +
assert(a == b);
 +
</syntaxhighlight>
 +
 
 +
could pretty-print
 +
 
 +
<syntaxhighlight lang="d">
 +
(a[1_000 .. 1_002] is [0, 0]) != (b[1_000 .. 1_002] is [1_000, 1_001])
 +
</syntaxhighlight>
 +
 
 +
An optional '''multiLine''' flag could in this case be motivated. When set this flag changes output to
 +
 
 +
<syntaxhighlight lang="d">
 +
(a[1_000 .. 1_002] is [0, 0]) !=
 +
(b[1_000 .. 1_002] is [1_000, 1_001])
 +
</syntaxhighlight>
 +
 
 +
This makes comparison more visually helpful to developers.
 +
 
 +
== References ==
 +
 
 +
Parts of the solution list at
  
 
https://issues.dlang.org/show_bug.cgi?id=5547#c3
 
https://issues.dlang.org/show_bug.cgi?id=5547#c3
  
can be reused, right?
+
including rewriting/expansion of '''AssertExpr''' could probably reused.
 +
 
 +
https://issues.dlang.org/show_bug.cgi?id=15889 - similar idea for bound checking
  
== Impact ==
+
=== Discussions on topic ===
 +
* http://forum.dlang.org/post/holdxspayjguauomrbcx@forum.dlang.org
  
 
== Copyright ==
 
== Copyright ==

Latest revision as of 21:46, 24 December 2016

Title: Configurable Assert Diagnostics
DIP: 83
Version: 1
Status: Draft
Created: 2015-10-01
Last Modified: 2016-12-24
Author: Per Nordlöw
Links:

Abstract

Allow for assert to do pretty printing of its failing expression when flagged for in call to compiler. Printing is configurable via specific sets of (template) function overloads.

Rationale

A failing assert (in a unittest), currently, give no hint about why it failed. To aid the developer in debugging, the failing assert of a

  • binary expression, such as assert(x == y), should print the values of x and y and
  • unary expression, such as assert(!x), should print the value of x.

This extra, so called, pretty printing can be enabled by changing the dmd flag -unittest to, say, -unittest=verbose for a specific failing module. This will be more convenient than explicitly adding the prints of the left-hand-side expression lhs and right-hand-side expression rhs directly before the call to the failing assert.

Description

This DIP proposes to add library-level-configurable diagnostics to failing calls to assert(expr) typically called from within unittest-blocks.

This diagnostics is activated only when DMD is called with a specific command line flag, say -unittest=verbose or perhaps -diagnose=assert.

Modifications needed in Compiler Frontend

Binary Operator Lowering

If DMD is called with this flag it will rewrite (lower) all assert expressions (AssertExpr in dmd source) such as

assert(a BINOP b)

into

(auto ref a, auto ref b) {
    if (a BINOP b) return;
    onAssertFailed!"BINOP"(a, b, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
} (e1, e2)

where onAssertFailed, in this binary operator case, is declared as

void onAssertFailed(string op, E1, E2)(E1 e1, E2 e2, string file, uint line, string function, string module);

Unary Operator Lowering

Similarly, for unary expressions, rewrite (lower)

assert(UNOP e)

with

(auto ref e) {
    if (UNOP e) return;
    onAssertFailed!"UNOP"(e, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
} (e)

where on onAssertFailed, in this unary operator case, is declared as

void onAssertFailed(string op, E)(E e, string file, uint line, string function, string module);

Non-Operator Lowering

For the case when no unary or binary operator is present in the top assert expression, unary overload of onAssertFailed is called with op being empty string. In other words the assert expression

assert(e)

is rewritten (lowered) into

(auto ref e) {
    if (e) return;
    onAssertFailed!""(e, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
} (e)

If the assert lowering must also be sensitive to expressions like

Non-Equality Operator Lowering

Further note that operator != may need special care because D only supports overloading via opUnary!"!" and opBinary!"==". This because the D compiler rewrites

x != y

into

!(x == y)

To make assert diagnostics as configurable as possible it may become relevant to enable the possibility for the developer to to define the overload onAssertFailed!"!=" aswell. If this is desirable the lowering logic in DMD must be sensitive to expressions

assert(!(a == b))

which, preferrably, will be rewritten to

(auto ref a, auto ref b) {
    if (!(a == b)) return;
    onAssertFailed!"!="(a, b, __FILE__, __LINE__, __FUNCTION__, __MODULE__);
} (e1, e2)

Modifications needed in D Runtime

Configuring Diagnostics

The default implementations (in druntime) of all the onAssertFailed-overloads should, as a first step, by defined to exactly mimic the current behaviour of assert(). This because, memory requirements of the call to DMD will increase when the compiler must generate terminal printing-code of all the arguments for all the unittests found in the compilation unit of interest. And it is currently unclear if this, current increase in system requirements, will be compatible with system resources currently available in build servers connected to GitHub.

Specific printing behaviour of assert() diagnostics can then be extendable by adding (typically templated) overloads of onAssertFailed for specific sets of types (concepts).

For instance, diagnostics specifically when comparing arithmetic types, could be realized through the overload

import std.traits : isArithmetic;
void onAssertFailed(string op, E1, E2)(E1 e1, E2 e2, string file, uint line, string function, string module)
    if (isArithmetic!L && 
        isArithmetic!R)
{
    version(assert)
    {
        import core.exception : AssertError;
        throw AssertError("Failed arithmetic assert: " ~ e1 ~ " " ~ binOp ~ " " ~ e2);
    }
}

In this way D would get the extendability we want in testing-frameworks such as std.experimental.testing (https://github.com/D-Programming-Language/phobos/pull/3207) without adding a new assert-overload-set and without sacrifycing default memory usage in DMD/Phobos unittests.

Further, this solution enables the possibility to provide fancy diagnostics behaviour in onAssertFailed for failing array/range or aggregate (struct or class) comparisons. This diagnostics could also have different pretty printing backends such as HTML.

Examples of Possible Diagnostics Outputs

Short Array Assert Arguments

For example a failing

assert([1,2,3] == [1,2,4]);

could pretty-print

([1,2,3][2] is 3) != ([1,2,4][2] is 4)

Aggregate Assert Arguments

or, for aggregates, a failing

struct A { int x, y; }
auto a = A(1,2);
auto b = A(1,3);
assert(a == b);

could pretty-print

(a.y is 2) != (b.y is 3)

Large Array Assert Arguments

or, for very large arrays, a failing

const n = 1_000_000;
auto a = iota(0, n).array;
auto b = a.dup;
a[1_000 .. 1_002] = 0;
assert(a == b);

could pretty-print

(a[1_000 .. 1_002] is [0, 0]) != (b[1_000 .. 1_002] is [1_000, 1_001])

An optional multiLine flag could in this case be motivated. When set this flag changes output to

(a[1_000 .. 1_002] is [0, 0]) !=
(b[1_000 .. 1_002] is [1_000, 1_001])

This makes comparison more visually helpful to developers.

References

Parts of the solution list at

https://issues.dlang.org/show_bug.cgi?id=5547#c3

including rewriting/expansion of AssertExpr could probably reused.

https://issues.dlang.org/show_bug.cgi?id=15889 - similar idea for bound checking

Discussions on topic

Copyright

This document has been placed in the Public Domain.