Activity for Martin Cooper
-
Martin Cooper posted a comment on merge request #15
You'll find the same functions in several other places already, scattered around the Linpac codebase. Eliminating duplicated headers is high on my personal "to do" list for Linpac internals, and I've actually started work on it in a local branch. Once the headers are centralised, getting common code in a single place will be the next logical step.
-
Integrate lptelnet package into Linpac proper
-
This was fixed 14 months ago (2019-02-11) on the develop branch. See: https://sourceforge.net/p/linpac/linpac/ci/2c78bb53df119e82511a957a1fe8f46e2aa1ba8c/
-
Integrate the lptelnet package
-
New functionality; bug fixes; documentation updates
-
Add minimal styling for readability
-
Fix a few misplaced tags
-
Significant revision and cleanup
-
Documentation for the shell API
-
Replace obsolete tt tag with code tag.
-
Regenerated text from substantially revised HTML
-
Fix initial scroll position in index view
-
Fix buffer overflow
-
Sign off before disconnecting
-
Gracefully handle missing messages
-
Fix confusing comment
-
Fix incorrect usage of vector::erase
-
Add a few minor debug niceties
-
Fix bug in modifying EventGate children
-
Add new command for integrated BBS mail retrieval
-
Fix message number lookup
-
Clean up mismatched allocations and frees
-
Fix line ends
-
Fix line ends
-
Fix stack corruption from blank lines in macros
-
Adjust ignores after moving files
-
Add support for PBBS message retrieval
-
Adjust build for moved mail utilities
-
Move mail tools into their own directory
-
Merge remote-tracking branch 'upstream/develop' into develop
-
Fix buffer overflow bugs
-
Fix possible buffer overflow in system info
-
Fix buffer overflow when reading long lines
-
Don't expect a response when none is required.
-
Don't send EasyComm I commands to EasyComm II rotators.
-
Fixes, corrections, and updates to the user manual.
-
Minor updates and corrections.
-
Minor changes to improve generated text.
-
Fix and sanitize some HTML rendering.
-
Fixes for language, typos, and the like.
-
Regenerated text manual from updated HTML.
-
Replace obsolete tt tag with code tag.
-
Fix incorrect character conversion indexing and add related docs.
-
Fix incorrect character conversion indexing.
-
Add a note about multi-byte encodings.
-
Fix bug in event gate when new process is created
-
Fix bug in event gate when new process is created
-
Fix several bugs in processing data from BBS
-
Fix several bugs in processing data from BBS
-
Fix badly broken config parsing in ulistd
-
Fix badly broken config parsing
-
Merge in cleanup of warnings and bug fixes.
-
Merge branch 'cleanup' into develop
-
Fix warning about ignoring system() result.
-
Fix warnings about signed array subscripts.
-
Fix warnings about unused variables and results.
-
Fix a bug in file name input handling.
-
Fix warnings about ambiguous 'else'; add braces.
-
Fix bugs related to mail header field lengths.
-
Fix warning about tmpnam; use mkstemp instead.
-
Address text pasting issues introduced by load reduction approach.
-
Correct loop control variable type.
-
Document the new EV_KEY_PRESS_MULTI event.
-
Add support for new EV_KEY_PRESS_MULTI event.
-
Fix a typo that appears in the mail UI.
-
Use the correct type (WINDOW) for ncurses windows.
-
Add support for new EV_KEY_PRESS_MULTI event.
-
Fix inconsistent args in wgetch() calls.
-
Use the correct type (WINDOW) for the key window.
-
Attached is a patch that combines the fix to the MonWindow constructor with the delete / free fixes to data.cc.
-
Hi David, To be clear, I haven't seen this segfault either. I was looking through the bug list here on SF, saw this one, looked at the associated code, and spotted an issue that could cause the reported segfault. I figured that was worth noting, along with a very minor change that would assist in nailing it down.
-
PMFJI. A few things: 1) The issue at windows.cc:877 is caused by a bug in the MonWindow constructor. The Window class defines a bunch of member variables, but has no constructor to initialise them, so it is up to derived classes to initialise them. MonWindow, derived from Window, fails to initialise conv (and convcnt), which belong to Window, so there's a random pointer here. Later, MonWindow::handle_event() tests conv and attempts to free it if it's non-null. Since it was not initialised, it is...
-
I see that MessageIndex::getMessage() does not protect against 'messages' being empty. If it is, index will be set to -1, and messages[index] will blow up. The behaviour of operator[] is undefined if the index is out of range, so a segfault is likely. Switching to messages.at(index) would cause an out_of_range exception to be thrown instead, since that function does bounds checking where operator[] does not. While that wouldn't solve the problem per se, it would allow us to determine whether in fact...
1