717407 - add null-arg checks and cleanup in mailnews/mime/src

Status: RESOLVED FIXED

Description

[:aceman]

I'll try to do some null-arg check in files of that directory. Also there are some possibilities to optimize 'if (!unlikely problem) return ERR' to 'NS_ENSURE_TRUE(unlikely problem, ERR)'.

Also, I'll add these from bug 413548:
--- a/mailnews/mime/src/nsMimeConverter.cpp	Tue Jan 22 16:02:08 2008 -0500
+++ b/mailnews/mime/src/nsMimeConverter.cpp	Tue Jan 22 16:31:14 2008 -0500
@@ -136,7 +136,8 @@ nsMimeConverter::EncodeMimePartIIStr(con
                                            PRInt32 encodedWordSize, 
                                            char          **encodedString)
 {
-
+  NS_ENSURE_ARG_POINTER(header);
+  NS_ENSURE_ARG_POINTER(mailCharset);
   // Encoder needs utf-8 string.
   nsAutoString tempUnicodeString;
   nsresult rv = ConvertToUnicode(mailCharset, header, tempUnicodeString);

Comment 1

[:aceman]

Created attachment 589647 [details] [diff] [review]
null checks, whitespace cleanup and NS_ENSUREs

Comment 2

[[e-mail]]

Comment on attachment 589647 [details] [diff] [review]
null checks, whitespace cleanup and NS_ENSUREs

>+  NS_ENSURE_ARG_POINTER(mailCharset);
Are you sure mailCharset can't be null? (I think it just defaults to ASCII.)

>-    char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
>+  char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
[This should actually use NS_Alloc. Sadly all_headers isn't null-terminated, so we can't simply clone it.]

>+  NS_ASSERTION (allHeaders, "nsMimeHeaders - out of memory");
>+  NS_ENSURE_TRUE(allHeaders, NS_ERROR_OUT_OF_MEMORY);
Not sure we need both checks here.

> static int
> Initialize(MimeObject *obj)
So, the rest of the file has a 4 space indent, which is fair enough, except that you changed this function from a 2 space indent to a weird 4+2 indent...

Comment 3

[:aceman]

(In reply to [e-mail] from comment #2)
> >+  NS_ENSURE_ARG_POINTER(mailCharset);
> Are you sure mailCharset can't be null? (I think it just defaults to ASCII.)
No idea, I took this from Joey Minta's patch in bug 413548.
I can remove it.

> >-    char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
> >+  char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
> [This should actually use NS_Alloc. Sadly all_headers isn't null-terminated,
> so we can't simply clone it.]
I assume I should not do anything about this.

> >+  NS_ASSERTION (allHeaders, "nsMimeHeaders - out of memory");
> >+  NS_ENSURE_TRUE(allHeaders, NS_ERROR_OUT_OF_MEMORY);
> Not sure we need both checks here.
Who will decide? :)

Comment 4

[[e-mail]]

(In reply to aceman from comment #3)
> (In reply to comment #2)
> > >+  NS_ENSURE_ARG_POINTER(mailCharset);
> > Are you sure mailCharset can't be null? (I think it just defaults to ASCII.)
> No idea, I took this from Joey Minta's patch in bug 413548.
Ah, well not all null in arguments are invalid. This was the only one that I spotted though; I guess if you mistakenly trap other null in arguments then exceptions will get thrown...

> > >-    char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
> > >+  char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
> > [This should actually use NS_Alloc. Sadly all_headers isn't null-terminated,
> > so we can't simply clone it.]
> I assume I should not do anything about this.
Well, I would appreciate it if you switched from PR_MALLOC to NS_Alloc.

> > >+  NS_ASSERTION (allHeaders, "nsMimeHeaders - out of memory");
> > >+  NS_ENSURE_TRUE(allHeaders, NS_ERROR_OUT_OF_MEMORY);
> > Not sure we need both checks here.
> Who will decide? :)
Well, unlike the previous code, NS_ENSURE_TRUE does log an error to stdout, which might make the assertion less useful.

Comment 5

[:aceman]

(In reply to [e-mail] from comment #4)
> (In reply to aceman from comment #3)
> > (In reply to comment #2)
> > > >+  NS_ENSURE_ARG_POINTER(mailCharset);
> > > Are you sure mailCharset can't be null? (I think it just defaults to ASCII.)
> > No idea, I took this from Joey Minta's patch in bug 413548.
> Ah, well not all null in arguments are invalid. This was the only one that I
> spotted though; I guess if you mistakenly trap other null in arguments then
> exceptions will get thrown...
Yes, I try to only check OUT parameters that are dereferenced and would cause a crash.
only this 'mailCharset' and 'header' IN parameters are checked because Joey had it that was. I can drop those 2.


> > > >-    char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
> > > >+  char *allHeaders = (char *) PR_MALLOC(mHeaders->all_headers_fp + 1);
> > > [This should actually use NS_Alloc. Sadly all_headers isn't null-terminated,
> > > so we can't simply clone it.]
> > I assume I should not do anything about this.
> Well, I would appreciate it if you switched from PR_MALLOC to NS_Alloc.
If it is just replacing the function name I can do that.

> > > >+  NS_ASSERTION (allHeaders, "nsMimeHeaders - out of memory");
> > > >+  NS_ENSURE_TRUE(allHeaders, NS_ERROR_OUT_OF_MEMORY);
> > > Not sure we need both checks here.
> > Who will decide? :)
> Well, unlike the previous code, NS_ENSURE_TRUE does log an error to stdout,
> which might make the assertion less useful.
Ok, I'll remove it.

Comment 6

[:aceman]

Created attachment 603387 [details] [diff] [review]
patch v2

Comment 7

[Ryan VanderMeulen [:RyanVM]]

http://hg.mozilla.org/comm-central/rev/1bb9b9a58eea

BTW Neil, the Modules wiki doesn't list you as a peer for MailNews Core::MIME. You should probably update that.
https://wiki.mozilla.org/Modules/MailNews_Core