MantisBT - NetSurf |
| View Issue Details |
|
| ID | Project | Category | View Status | Date Submitted | Last Update |
| 0002143 | NetSurf | RISC OS-specific | public | 2014-06-10 21:37 | 2015-03-10 11:06 |
|
| Reporter | Harriet Bazley | |
|---|
| Assigned To | Vincent Sanders | |
|---|
| Priority | normal | Severity | minor | Reproducibility | always |
|---|
| Status | closed | Resolution | fixed | |
|---|
| Platform | ARM | OS | RISC OS | OS Version | 5.19 |
|---|
| Product Version | 3.2 | |
|---|
| Target Version | | Fixed in Version | 3.2 | |
|---|
| Fixed in CI build # | |
|---|
| Reported in CI build # | 1966 |
|---|
| URL of problem page | https://www.fanfiction.net/r/7302546/ |
|---|
|
| Summary | 0002143: Lower part of page hidden/crash on fanfiction.net |
|---|
| Description | Loading the page https://www.fanfiction.net/r/7302546/ displays only the first few paragraphs of content, below which the window is blank.
Moreover subsequently attempting to edit the URL (in order to see if the mobile version of the site would render any better) causes a segmentation fault in the browser. |
|---|
| Steps To Reproduce | Double-click on URL in text file (or via web link).
Place input focus in URL bar and delete a couple of 'w's from "http://www...."
Browser crashes without return key being pressed (could this be associated with auto-complete?) |
|---|
| Additional Information | URL editing was fine as of Netsurf v.1932. Tried Netsurf v1963 but this crashes with the same symptoms as v1966. |
|---|
| Tags | No tags attached. |
|---|
| Relationships | |
| Attached Files |  Log.zip (6,761) 2014-06-10 21:37
https://bugs.netsurf-browser.org/mantis/file_download.php?file_id=157&type=bug |
|---|
|
Notes |
|
|
|
|
The URL-editing issue appears to be a generic problem with the current release which will cause any site to crash - hence rather more serious than I had assumed! |
|
|
|
|
this is caused by idna_encode() in utils/idna.c:594
Breakpoint 1, idna_encode (host=0x7fffdc02eff0 ".fanfiction.net", len=15,
ace_host=0x7fffffffce88, ace_len=0x7fffffffce80) at utils/idna.c:598
598 size_t label_len, output_len, ucs4_len, fqdn_len = 0;
when entered thusly the while loop immediately terminates due to idna__host_label_length() returning 0 and then decrements the pointer back off the front of the stack based char array and tries to duplicate the "string" followed by setting the output length to -1
this all ends very badly. |
|
|
|
|
|
fixed idna en/de coder to not explode if the first part of host is missing |
|
|
|
|
|
Confirmed fixed in 3.2 release |
|