Clarify that the pass key new function can fail.

[iphydf]

This change is 

[GrayHatter]

---

Reviewed 2 of 2 files at r1.
Review status: all files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[robinlinden]

Reviewed 2 of 2 files at r1.
Review status: all files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[sudden6]

Review status: all files reviewed at latest revision, 2 unresolved discussions, some commit checks failed.

---

toxencryptsave/toxencryptsave.api.h, line 168 at r1 (raw file):

 * @param ciphertext_len The length of the cipher text array. At least $PASS_ENCRYPTION_EXTRA_LENGTH.
 * @param passphrase The user-provided password. Can be empty or NULL.
 * @param passphrase_len The length of the password. Must be 0 if passphrase is NULL.

Does it really have to be 0? IMO it would be better to ignore it if passphrase is NULL

---

toxencryptsave/toxencryptsave.api.h, line 204 at r1 (raw file):

   * initialise it, use one of the derive_* functions below.
   *
   * In case of allocation failure, this function returns NULL.

Only on allocation failure? Else it would be better to write In case of failure.

---

Comments from Reviewable

[iphydf]

Review status: all files reviewed at latest revision, 2 unresolved discussions, some commit checks failed.

---

toxencryptsave/toxencryptsave.api.h, line 168 at r1 (raw file):

Previously, sudden6 wrote…

Does it really have to be 0? IMO it would be better to ignore it if passphrase is NULL

Why do you think it would be better? If you pass NULL, but a non-zero length, you indicate that you want the function to read more than 0 bytes from the password, which is invalid. Reading exactly 0 bytes from a NULL pointer is valid.

---

toxencryptsave/toxencryptsave.api.h, line 204 at r1 (raw file):

Previously, sudden6 wrote…

Only on allocation failure? Else it would be better to write In case of failure.

Allocation failure is the only possible failure here. The only thing this function will ever do (in this version of the API) is call malloc.

---

Comments from Reviewable

[sudden6]

Review status: all files reviewed at latest revision, 2 unresolved discussions, some commit checks failed.

---

toxencryptsave/toxencryptsave.api.h, line 168 at r1 (raw file):

Previously, iphydf wrote…

Why do you think it would be better? If you pass NULL, but a non-zero length, you indicate that you want the function to read more than 0 bytes from the password, which is invalid. Reading exactly 0 bytes from a NULL pointer is valid.

If the function throws an error when passphrase == NULL AND passphrase != 0 it's ok for me.

---

toxencryptsave/toxencryptsave.api.h, line 204 at r1 (raw file):

Previously, iphydf wrote…

Allocation failure is the only possible failure here. The only thing this function will ever do (in this version of the API) is call malloc.

Removing allocation from the sentence also removes the thought "are there any other failure modes I have to care for?" from the API users mind, which is IMO a good thing. But if you don't see this as a problem leave it.

---

Comments from Reviewable

[iphydf]

Review status: 0 of 2 files reviewed at latest revision, 2 unresolved discussions.

---

toxencryptsave/toxencryptsave.api.h, line 168 at r1 (raw file):

Previously, sudden6 wrote…

If the function throws an error when passphrase == NULL AND passphrase != 0 it's ok for me.

Actually, this is incorrect. I just checked. The API is inconsistent: it returns an error when passphrase is NULL here. tox_pass_key_derive_with_salt on the other hand accepts it. Amended.

---

toxencryptsave/toxencryptsave.api.h, line 204 at r1 (raw file):

Previously, sudden6 wrote…

Removing allocation from the sentence also removes the thought "are there any other failure modes I have to care for?" from the API users mind, which is IMO a good thing. But if you don't see this as a problem leave it.

Clarified.

---

Comments from Reviewable

[robinlinden]

Reviewed 2 of 2 files at r2.
Review status: all files reviewed at latest revision, 2 unresolved discussions.

---

Comments from Reviewable

[sudden6]

---

Review status: all files reviewed at latest revision, all discussions resolved.

---

Comments from Reviewable

[sudden6]

Reviewed 2 of 2 files at r2.
Review status: all files reviewed at latest revision, all discussions resolved, some commit checks failed.

---

Comments from Reviewable

[robinlinden]

---

Review status: all files reviewed at latest revision, all discussions resolved, some commit checks failed.

---

Comments from Reviewable

[nurupo]

---

Comments from Reviewable

[nurupo]

@iphydf you forgot to reword the commit message, it still says

Also clarify that passwords can be empty or NULL.