Mirror of forum.nim-lang.org

5878 :: Is "danger" define supposed to also define "release"?

• Yardanico (orginal) [2020-02-02T17:26:11+01:00] view original

  Nim code (even stdlib) contains some code which has some debugging options enabled with check like when defined(release) or when not defined(release). I think that "danger" define should also imply "release", is that correct?

  Because for example https://github.com/nim-lang/Nim/blob/devel/lib/pure/asyncfutures.nim#L26 checks for release, so with "danger" that code will still be there (I actually checked and it is indeed compiled with -d:danger since it doesn't define release now)

---

• Yardanico (orginal) [2020-02-02T17:31:03+01:00] view original

  Or maybe we're just supposed to do both -d:release -d:danger when we want a release build with all checks disabled?

---

• Yardanico (orginal) [2020-02-02T17:34:07+01:00] view original

  Also, docs in https://nim-lang.org/docs/nimc.html#compiler-usage-configuration-files say that you should just use "-d:danger" for release builds without "-d:release"

---

• ktamp (orginal) [2020-02-02T19:06:08+01:00] view original

  The only thing is "-d:danger" does not define "-d:release"...

---

• Yardanico (orginal) [2020-02-02T22:50:43+01:00] view original

  That's the point of this thread - figure out if this is expected or a bug and -d:danger should also define -d:release

---

• juancarlospaco (orginal) [2020-02-03T01:02:03+01:00] view original

  Expected. (?)

---

• Araq (orginal) [2020-02-03T01:53:53+01:00] view original

  This is a serious bug, -d:danger should imply -d:release to the point that defined(release) is true. We will fix it and backport the fix.

---

• Vindaar (orginal) [2020-02-03T12:17:39+01:00] view original

  Oh wow, I somehow thought it was intended behavior, since @mratsim advocated to compile with -d:release -d:danger almost immediately after -d:danger was introduced. So I thought you were aware of this.

---

• dom96 (orginal) [2020-02-03T15:48:43+01:00] view original

  Even though this happened before 1.0 (AFAIK) it would have broken a lot of code, so it makes perfect sense that we should consider this to be a bug and a serious one at that.

---