400195 - cert exception file has unix line endings on Windows

Status: RESOLVED FIXED

(Core :: Security: PSM, defect, P2)

Description

[Nelson Bolyard (seldom reads bugmail)]

The new file of bad cert exceptions is a text file.
It is written with Unix line endings on all platforms, including Windows.
This means it cannot be cleanly displayed or edited on Windows.
All text files should have platform-dependent line endings.

Comment 1

[Kai Engert [:KaiE:]]

Attachment: Patch v1

Comment 2

[Nelson Bolyard (seldom reads bugmail)]

Comment on attachment 285443 [details] [diff] [review]
Patch v1

This will write them correctly.  Will they be read correctly?

Comment 3

[Kai Engert [:KaiE:]]

(In reply to comment #2)
> (From update of attachment 285443 [details] [diff] [review])
> This will write them correctly.  Will they be read correctly?

Yes, tested.

Comment 4

[Daniel Veditz [:dveditz]]

If the file can be stored with different endings then the reading code needs to be able to handle any platform's variant at any time.

 - people copy profiles to new architectures
   (especially with the rising popularity of multi-OS machines)

 - people share the same profile directory on different OS's.
   Again, increasingly true with multi-boot and VM-based multi-OS
   machines, though from the earliest days Netscape had to worry 
   about network-mounted profiles shared amongst machines with
   differing endianness.

Comment 5

[Robert Sayre]

Comment on attachment 285443 [details] [diff] [review]
Patch v1

Please answer dveditz's questions before requesting approval.

Comment 6

[Kai Engert [:KaiE:]]

I think I answered in comment 3 that I tested that files will always be read correctly, independently of the line endings.

The backend code that reads lines in (somewhere in a global stream class) seemed to allow for any combinations of line endings.

I explicitly tested:
- produce file on Linux
- copy file to Windows 
- run on Windows
=> works

Comment 7

[Nelson Bolyard (seldom reads bugmail)]

Comment on attachment 285443 [details] [diff] [review]
Patch v1

Kai answered the questions

Comment 8

[Robert Sayre]

Comment on attachment 285443 [details] [diff] [review]
Patch v1

OK, I understand this bug better now. Thanks Kaie and Nelson.

Comment 9

[Kai Engert [:KaiE:]]

checked in, fixed.