Transfer issue tickets? #2
Comments
|
Yes, we closed svn access and some programs seem to issue cryptic error messages. I replied already to IngwiePhoenix/FLTK#4 (Wayland) and suggest to close this issue (don't transfer it). I also replied to IngwiePhoenix/FLTK#3 (FLTKConfig.cmake). I believe this issue could also be closed (I answered the OP's question). I believe you can close the PR IngwiePhoenix/FLTK#5 because I committed updates in both the master branch (FLTK 1.4) and branch-1.3 (FLTK 1.3). The fix will be in FLTK 1.3.5. Hence I believe there's nothing left to transfer and we can close this issue. Do you agree? |
|
Fair enough. I will go ahead and actually terminate this mirror alltogether - after all, it's purpose is pretty much gone now. Thank you for replying to the issues! Though I'd like to know when exactly you stopped SVN access? |
|
We closed svn access on Dec 01, 2018 at about 21:00 CET. |
|
Note: all messages like "user mentioned this issue" above and (in case this happens again) are very likely not intended (false '#'nnn interpretation). |
Error was reported by Address Sanitizer (ASAN) when picking a file.
Platform: Unix/Linux.
==1734703==ERROR: AddressSanitizer: alloc-dealloc-mismatch (malloc vs operator delete []) on 0x607000108420
#0 0x7f3357d846ef in operator delete[](void*) ../../../../src/libsanitizer/asan/asan_new_delete.cc:168
#1 0x4e195f in Fl_Native_File_Chooser_Driver::strfree(char*) ../../src/Fl_Native_File_Chooser.cxx:262
...
0x607000108420 is located 0 bytes inside of 66-byte region [0x607000108420,0x607000108462)
allocated by thread T0 here:
#0 0x7f3357d0a3ed in __interceptor_strdup ../../../../src/libsanitizer/asan/asan_interceptors.cc:445
#1 0x57951a in Fl_Kdialog_Native_File_Chooser_Driver::filter(char const*) ../../src/Fl_Native_File_Chooser_Kdialog.cxx:245
#2 0x4e14a0 in Fl_Native_File_Chooser::filter(char const*) ../../src/Fl_Native_File_Chooser.cxx:176
[WIP: squashed and rebased by Albrecht-S, leaving all commit messages...] Commit message #1: Removes some window positioning bodges, which mess with coordinates. Window Managers should be responsible for placing windows correctly if they're adding title bars and such to them. Note that negative X and Y will only end up truly accurate if the screen scale is (or has been set to) 1.0. This is because information about the actual screen dimensions has been lost scaling integers. Updated patch fixes use of uninitialised variable, with help from Albrecht Schlosser. Commit message fltk#2: Offset position in X11 for east and south gravity Adds new virtual functions x_from_right() and y_from_bottom() to Fl_Screen_Driver. These are overridden in Fl_X11_Screen_Driver to return the offset between the actual screen edge (right or bottom) and the scaled up screen edge. The calculation is: scaled delta = scaled limit - scaled coordinate real coordinate = real limit - (scaled delta * screen scale) These offsets then applied when creating a window with east or south gravity to adjust for the loss of precision in the screen dimensions when they were scaled. Commit message fltk#3: X11 geometry: use parsed width and height in negative calculations If the user has specified a new dimension, that's how much we need to offset the coordinates. Using gw, gh is safe here: Fl_Screen_Driver::XParseGeometry() will only update their values if there was a user-specified value.
Hello!
For quite a while, I had tried to maintain an svn-to-git mirror of FLTK: https://github.com/IngwiePhoenix/FLTK - BUT... git-svn stopped doing what it is supposed to, and keeps throwing "Checksum missmatch" messages at me each time I try to fetch. Let alone rebase...
But, there happen to be issue and PR tickets on this repo, that I'd like to point towards:
And this PR:
I would like to close my mirror, seeing as there is an official repository now. Thank you for making FLTK - and finally an official FLTK repo... :)
The text was updated successfully, but these errors were encountered: