You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm just going to start cataloguing the bugs as I find then in test re-factoring.
There's a test in Manifest.t that ensures maniadd() does not attempt to write to the file if an entry already exists, and that adding an entry that doesnt exist causes a fatal exception to occur.
However, maniadd inside it calls this horrible function called _fix_manifest, which munges the state of the manifest on disk prior to maniadd actually doing anything.
This in turn can silently cause the file to be marked writeable, and the expected failure of maniadd vanishes.
repro steps:
path("MANIFEST")->spew_raw("foo #bar");
chmod(0400, "MANIFEST");
maniadd({ foo=>"bar" }); # no fatalsdie"Is now writeable"if-w"MANIFEST"
The text was updated successfully, but these errors were encountered:
kentfredric
added a commit
to kentfredric/ExtUtils-Manifest
that referenced
this issue
Apr 6, 2015
I'm just going to start cataloguing the bugs as I find then in test re-factoring.
There's a test in Manifest.t that ensures
maniadd()
does not attempt to write to the file if an entry already exists, and that adding an entry that doesnt exist causes a fatal exception to occur.However,
maniadd
inside it calls this horrible function called_fix_manifest
, which munges the state of the manifest on disk prior to maniadd actually doing anything.This in turn can silently cause the file to be marked writeable, and the expected failure of maniadd vanishes.
repro steps:
The text was updated successfully, but these errors were encountered: