diff options
author | Björn Gustavsson <[email protected]> | 2017-11-14 09:05:31 +0100 |
---|---|---|
committer | Björn Gustavsson <[email protected]> | 2017-11-30 12:46:51 +0100 |
commit | ef2e55c707a6c3d1f54f2acb9c8bfbe3fb7a8659 (patch) | |
tree | 7a9eae08bd786efbebddb53763888a66ed580abb /erts/emulator/beam/ops.tab | |
parent | 0c4574cc31414a5204eabfad7eb0ac5095c0ace1 (diff) | |
download | otp-ef2e55c707a6c3d1f54f2acb9c8bfbe3fb7a8659.tar.gz otp-ef2e55c707a6c3d1f54f2acb9c8bfbe3fb7a8659.tar.bz2 otp-ef2e55c707a6c3d1f54f2acb9c8bfbe3fb7a8659.zip |
Add syntax in try/catch to retrieve the stacktrace directly
This commit adds a new syntax for retrieving the stacktrace
without calling erlang:get_stacktrace/0. That allow us to
deprecate erlang:get_stacktrace/0 and ultimately remove it.
The problem with erlang:get_stacktrace/0 is that it can keep huge
terms in a process for an indefinite time after an exception. The
stacktrace can be huge after a 'function_clause' exception or a failed
call to a BIF or operator, because the arguments for the call will be
included in the stacktrace. For example:
1> catch abs(lists:seq(1, 1000)).
{'EXIT',{badarg,[{erlang,abs,
[[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20|...]],
[]},
{erl_eval,do_apply,6,[{file,"erl_eval.erl"},{line,674}]},
{erl_eval,expr,5,[{file,"erl_eval.erl"},{line,431}]},
{shell,exprs,7,[{file,"shell.erl"},{line,687}]},
{shell,eval_exprs,7,[{file,"shell.erl"},{line,642}]},
{shell,eval_loop,3,[{file,"shell.erl"},{line,627}]}]}}
2> erlang:get_stacktrace().
[{erlang,abs,
[[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,
23,24|...]],
[]},
{erl_eval,do_apply,6,[{file,"erl_eval.erl"},{line,674}]},
{erl_eval,expr,5,[{file,"erl_eval.erl"},{line,431}]},
{shell,exprs,7,[{file,"shell.erl"},{line,687}]},
{shell,eval_exprs,7,[{file,"shell.erl"},{line,642}]},
{shell,eval_loop,3,[{file,"shell.erl"},{line,627}]}]
3>
We can extend the syntax for clauses in try/catch to optionally bind
the stacktrace to a variable.
Here is an example using the current syntax:
try
Expr
catch C:E ->
Stk = erlang:get_stacktrace(),
.
.
.
In the new syntax, it would look like:
try
Expr
catch
C:E:Stk ->
.
.
.
Only a variable (not a pattern) is allowed in the stacktrace position,
to discourage matching of the stacktrace. (Matching would also be
expensive, because the raw format of the stacktrace would have to be
converted to the cooked form before matching.)
Note that:
try
Expr
catch E ->
.
.
.
is a shorthand for:
try
Expr
catch throw:E ->
.
.
.
If the stacktrace is to be retrieved for a throw, the 'throw:'
prefix must be explicitly included:
try
Expr
catch throw:E:Stk ->
.
.
.
Diffstat (limited to 'erts/emulator/beam/ops.tab')
-rw-r--r-- | erts/emulator/beam/ops.tab | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/erts/emulator/beam/ops.tab b/erts/emulator/beam/ops.tab index 3df91056cb..ba4ca67e5e 100644 --- a/erts/emulator/beam/ops.tab +++ b/erts/emulator/beam/ops.tab @@ -1575,3 +1575,9 @@ i_recv_mark recv_set Fail | label Lbl | loop_rec Lf Reg => \ i_recv_set | label Lbl | loop_rec Lf Reg i_recv_set + +# +# OTP 21. +# + +build_stacktrace |