buffer = SvPVutf8(bufsv, blen);
}
else {
- if (DO_UTF8(bufsv))
- sv_utf8_downgrade(bufsv, FALSE);
- buffer = SvPV(bufsv, blen);
+ if (DO_UTF8(bufsv)) {
+ /* Not modifying source SV, so making a temporary copy. */
+ bufsv = sv_2mortal(newSVsv(bufsv));
+ sv_utf8_downgrade(bufsv, FALSE);
+ }
+ buffer = SvPV(bufsv, blen);
}
if (PL_op->op_type == OP_SYSWRITE) {
* AIX getpwnam() is clever enough to return the encrypted password
* only if the caller (euid?) is root.
*
- * There are at least two other shadow password APIs. Many platforms
+ * There are at least three other shadow password APIs. Many platforms
* seem to contain more than one interface for accessing the shadow
* password databases, possibly for compatibility reasons.
* The getsp*() is by far he simplest one, the other two interfaces
* char *(getespw*(...).ufld.fd_encrypt)
* Mention HAS_GETESPWNAM here so that Configure probes for it.
*
+ * <userpw.h> (AIX)
+ * struct userpw *getuserpw();
+ * The password is in
+ * char *(getuserpw(...)).spw_upw_passwd
+ * (but the de facto standard getpwnam() should work okay)
+ *
* Mention I_PROT here so that Configure probes for it.
*
* In HP-UX for getprpw*() the manual page claims that one should include
* Divert the urge to writing an extension instead.
*
* --jhi */
-# ifdef HAS_GETSPNAM
+ /* Some AIX setups falsely(?) detect some getspnam(), which
+ * has a different API than the Solaris/IRIX one. */
+# if defined(HAS_GETSPNAM) && !defined(_AIX)
{
struct spwd *spwent;
int saverrno; /* Save and restore errno so that