about summary refs log tree commit diff
path: root/lib/debug.nix
blob: 925e0d405a79aaa1eb2203b948dfa995bbb6986b (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
let lib = import ./default.nix;

inherit (builtins) trace attrNamesToStr isAttrs isFunction isList isInt
        isString isBool head substring attrNames;

inherit (lib) all id mapAttrsFlatten elem;

in

rec {

  inherit (builtins) addErrorContext;

  addErrorContextToAttrs = lib.mapAttrs (a: v: lib.addErrorContext "while evaluating ${a}" v);

  traceIf = p: msg: x: if p then trace msg x else x;

  traceVal = x: trace x x;
  traceXMLVal = x: trace (builtins.toXML x) x;
  traceXMLValMarked = str: x: trace (str + builtins.toXML x) x;

  # strict trace functions (traced structure is fully evaluated and printed)

  /* `builtins.trace`, but the value is `builtins.deepSeq`ed first. */
  traceSeq = x: y: trace (builtins.deepSeq x x) y;

  /* Like `traceSeq`, but only down to depth n.
   * This is very useful because lots of `traceSeq` usages
   * lead to an infinite recursion.
   */
  traceSeqN = depth: x: y: with lib;
    let snip = v: if      isList  v then noQuotes "[…]" v
                  else if isAttrs v then noQuotes "{…}" v
                  else v;
        noQuotes = str: v: { __pretty = const str; val = v; };
        modify = n: fn: v: if (n == 0) then fn v
                      else if isList  v then map (modify (n - 1) fn) v
                      else if isAttrs v then mapAttrs
                        (const (modify (n - 1) fn)) v
                      else v;
    in trace (generators.toPretty { allowPrettyValues = true; }
               (modify depth snip x)) y;

  /* `traceSeq`, but the same value is traced and returned */
  traceValSeq = v: traceVal (builtins.deepSeq v v);
  /* `traceValSeq` but with fixed depth */
  traceValSeqN = depth: v: traceSeqN depth v v;


  # this can help debug your code as well - designed to not produce thousands of lines
  traceShowVal = x: trace (showVal x) x;
  traceShowValMarked = str: x: trace (str + showVal x) x;
  attrNamesToStr = a: lib.concatStringsSep "; " (map (x: "${x}=") (attrNames a));
  showVal = x:
      if isAttrs x then
          if x ? outPath then "x is a derivation, name ${if x ? name then x.name else "<no name>"}, { ${attrNamesToStr x} }"
          else "x is attr set { ${attrNamesToStr x} }"
      else if isFunction x then "x is a function"
      else if x == [] then "x is an empty list"
      else if isList x then "x is a list, first element is: ${showVal (head x)}"
      else if x == true then "x is boolean true"
      else if x == false then "x is boolean false"
      else if x == null then "x is null"
      else if isInt x then "x is an integer `${toString x}'"
      else if isString x then "x is a string `${substring 0 50 x}...'"
      else "x is probably a path `${substring 0 50 (toString x)}...'";

  # trace the arguments passed to function and its result
  # maybe rewrite these functions in a traceCallXml like style. Then one function is enough
  traceCall  = n: f: a: let t = n2: x: traceShowValMarked "${n} ${n2}:" x; in t "result" (f (t "arg 1" a));
  traceCall2 = n: f: a: b: let t = n2: x: traceShowValMarked "${n} ${n2}:" x; in t "result" (f (t "arg 1" a) (t "arg 2" b));
  traceCall3 = n: f: a: b: c: let t = n2: x: traceShowValMarked "${n} ${n2}:" x; in t "result" (f (t "arg 1" a) (t "arg 2" b) (t "arg 3" c));

  # FIXME: rename this?
  traceValIfNot = c: x:
    if c x then true else trace (showVal x) false;

  /* Evaluate a set of tests.  A test is an attribute set {expr,
     expected}, denoting an expression and its expected result.  The
     result is a list of failed tests, each represented as {name,
     expected, actual}, denoting the attribute name of the failing
     test and its expected and actual results.  Used for regression
     testing of the functions in lib; see tests.nix for an example.
     Only tests having names starting with "test" are run.
     Add attr { tests = ["testName"]; } to run these test only
  */
  runTests = tests: lib.concatLists (lib.attrValues (lib.mapAttrs (name: test:
    let testsToRun = if tests ? tests then tests.tests else [];
    in if (substring 0 4 name == "test" ||  elem name testsToRun)
       && ((testsToRun == []) || elem name tests.tests)
       && (test.expr != test.expected)

      then [ { inherit name; expected = test.expected; result = test.expr; } ]
      else [] ) tests));

  # create a test assuming that list elements are true
  # usage: { testX = allTrue [ true ]; }
  testAllTrue = expr: { inherit expr; expected = map (x: true) expr; };

  strict = v:
    trace "Warning: strict is deprecated and will be removed in the next release"
      (builtins.seq v v);

  # example: (traceCallXml "myfun" id 3) will output something like
  # calling myfun arg 1: 3 result: 3
  # this forces deep evaluation of all arguments and the result!
  # note: if result doesn't evaluate you'll get no trace at all (FIXME)
  #       args should be printed in any case
  traceCallXml = a:
    if !isInt a then
      traceCallXml 1 "calling ${a}\n"
    else
      let nr = a;
      in (str: expr:
          if isFunction expr then
            (arg:
              traceCallXml (builtins.add 1 nr) "${str}\n arg ${builtins.toString nr} is \n ${builtins.toXML (strict arg)}" (expr arg)
            )
          else
            let r = strict expr;
            in trace "${str}\n result:\n${builtins.toXML r}" r
      );
}