digitalmars.D - std.regex horribly broken in 2.080?
- WebFreak (69/69) May 04 2018 I am currently working on workspace-d/serve-d but I am
- Seb (4/12) May 04 2018 FWIW ctRegex might be deprecated soon:
- Dmitry Olshansky (7/15) May 04 2018 I could take a look. Any chance to have the exact reproduction
- Dmitry Olshansky (3/19) May 04 2018 I think I see the code where the trace comes from. Might take a
I am currently working on workspace-d/serve-d but I am
continously running into segfaults with std.regex ctRegex with
the captures in dub and dscanner. I can't provide any other
information really because my internet is dead right now. This
issue is happening all the time there, yet I can't reproduce it.
Can someone with internet maybe check where that issue might come
from?
[Switching to Thread 0x7f3db0ba7700 (LWP 14593)]
0x000055f92a385dd3 in
_D3std5regex__T10matchFirstTAyaTSQBfQBe__T11ctRegexImplVQBca7_5e5c647b352c7dVQBxA0Z7WrapperZQCzFNfQCsQCrZSQEaQD
__T8CapturesTQDtZQo (__HID720=0x7f3d00000001, re=...,
input=...) at /usr/include/dlang/dmd/std/regex/package.d:980
980 import std.algorithm.iteration : map;
(gdb) bt
_D3std5regex__T10matchFirstTAyaTSQBfQBe__T11ctRegexImplVQBca7_5e5c647b352c7dVQBxA0Z7WrapperZQCzFNfQCsQCrZSQEaQD
__T8CapturesTQDtZQo (__HID720=0x7f3d00000001,
re=..., input=...) at
/usr/include/dlang/dmd/std/regex/package.d:980
_D8analysis7numbers16NumberStyleCheck5visitMFxS3std12experimental5lexer__T14TokenStructureThVAyaa305_0a20202020737472696e6720636f6d6d656e743b0a20202020737472696e6720747261696c696e67436f6d6d656e743b0a0a20202020696e74206f70436d702873697a655f7420692920636f6e73742070757265206e6f7468726f77204073616665207b0a202020202020202069662028696e646578203c2069292072657475726e202d313b0a202020202020202069662028696e646578203e2069292072657475726e20313b0a202020202020202072657475726e20303b0a202020207d0a0a20202020696e74206f70436d702872656620636f6e737420747970656f66287468697329206f746865722920636f6e73742070757265206e6f7468726f77204073616665207b0a202020202020202072657475726e206f70436d70286f746865722e696e646578293b
a202020207d0aZQYoZv (
this=0x7f3db3002100, t=...) at
../../.dub/packages/dscanner-0.4.2/dscanner/src/analysis/numbers.d:38
_D6dparse3ast17PrimaryExpression6acceptMxFCQBpQBl10ASTVisitorZv
(this=0x7f3da805bde8, visitor=0x7f3db3002100)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d-mixin-2503:2506
_D6dparse3ast10ASTVisitor5visitMFxCQBhQBd17PrimaryExpressionZv
(this=0x7f3db3002100, primaryExpression=0x7f3da805bde8)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d:279
_D6dparse3ast15UnaryExpression6acceptMxFCQBnQBj10ASTVisitorZv
(this=0x7f3da805bcd8, visitor=0x7f3db3002100)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d-mixin-3252:3252
_D6dparse3ast10ASTVisitor5visitMFxCQBhQBd15UnaryExpressionZv
(this=0x7f3db3002100, unaryExpression=0x7f3da805bcd8)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d:332
_D6dparse3ast10ASTVisitor5visitMFxCQBhQBd14ExpressionNodeZv
(this=0x7f3db3002100, n=0x7f3da805bcd8)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d:137
_D6dparse3ast16AssignExpression6acceptMxFCQBoQBk10ASTVisitorZv
(this=0x7f3da805bc98, visitor=0x7f3db3002100)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d-mixin-867:868
_D6dparse3ast10ASTVisitor5visitMFxCQBhQBd16AssignExpressionZv
(this=0x7f3db3002100, assignExpression=0x7f3da805bc98)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d:175
_D6dparse3ast10ASTVisitor5visitMFxCQBhQBd14ExpressionNodeZv
(this=0x7f3db3002100, n=0x7f3da805bc98)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d:108
_D6dparse3ast10Expression6acceptMxFCQBiQBe10ASTVisitorZv
(this=0x7f3da805b798, visitor=0x7f3db3002100)
at
../../.dub/packages/libdparse-0.7.2-alpha.6/libdparse/src/dparse/ast.d-mixin-1575:1575
_D6dparse3ast10ASTVisitor5visitMFxCQBhQBd10ExpressionZv
(this=0x7f3db3002100, expression=0x7f3da805b798)
May 04 2018
Seb <seb wilzba.ch> On Friday, 4 May 2018 at 11:39:18 UTC, WebFreak wrote:I am currently working on workspace-d/serve-d but I am continously running into segfaults with std.regex ctRegex with the captures in dub and dscanner. I can't provide any other information really because my internet is dead right now. This issue is happening all the time there, yet I can't reproduce it. Can someone with internet maybe check where that issue might come from? [...]FWIW ctRegex might be deprecated soon: https://github.com/dlang/phobos/pull/6164 Does the normal regex work?
May 04 2018
On Friday, 4 May 2018 at 11:39:18 UTC, WebFreak wrote:I am currently working on workspace-d/serve-d but I am continously running into segfaults with std.regex ctRegex with the captures in dub and dscanner. I can't provide any other information really because my internet is dead right now. This issue is happening all the time there, yet I can't reproduce it. Can someone with internet maybe check where that issue might come from?I could take a look. Any chance to have the exact reproduction steps? By the look of it something is wrong with Captures struct being destroyed twice or something. Could also be compiler issue. Another thing is to disable inlining and I could at least see detailed stack trace.[...]
May 04 2018
On Friday, 4 May 2018 at 14:33:09 UTC, Dmitry Olshansky wrote:On Friday, 4 May 2018 at 11:39:18 UTC, WebFreak wrote:I think I see the code where the trace comes from. Might take a shot at hackathon.I am currently working on workspace-d/serve-d but I am continously running into segfaults with std.regex ctRegex with the captures in dub and dscanner. I can't provide any other information really because my internet is dead right now. This issue is happening all the time there, yet I can't reproduce it. Can someone with internet maybe check where that issue might come from?I could take a look. Any chance to have the exact reproduction steps? By the look of it something is wrong with Captures struct being destroyed twice or something. Could also be compiler issue. Another thing is to disable inlining and I could at least see detailed stack trace.[...]
May 04 2018