lagrange: Segmentation fault

[haary]

System

• xuname:
  Void 5.13.19_1 x86_64 GenuineIntel uptodate rFF
• package:
  affected package(s) including the version: lagrange-1.6.5_1

Expected behavior

Does not crash

Actual behavior

Crashes with "Segmentaion fault"

Steps to reproduce the behavior

• Open Lagrange
• Open some Gemini sites
• Scroll across those sites. Mark some words with the cursor
• Lagrange crashes with "Segmentation fault" after a few seconds.

This does not happen with the AppImage version of Lagrange from the author's Gitub site https://github.com/skyjake/lagrange/releases

[paper42]

I can not reproduce this, do you have an example page where this happens? Are you able to post a backtrace?

[haary]

It does happen at gemini://senioradmin.de/GreyMode.gmi. It involves scrolling and clicking

Backtrace:

Thread 1 "lagrange" received signal SIGSEGV, Segmentation fault.
0x0000555555606ec1 in ?? ()
(gdb) backtrace
#0  0x0000555555606ec1 in ?? ()
#1  0x0000555555577460 in ?? ()
#2  0x000055555559cc94 in ?? ()
#3  0x00005555555f074b in ?? ()
#4  0x00005555555ed4a0 in ?? ()
#5  0x0000555555567c05 in ?? ()
#6  0x0000555555568c57 in ?? ()
#7  0x0000555555561a30 in main ()

[paper42]

Could you install lagrange-dbg from the debug repository? This backtrace doesn't tell us much.

[haary]

Thread 1 "lagrange" received signal SIGSEGV, Segmentation fault.
decodeBytes_MultibyteChar (bytes=0x498000003aa <error: Cannot access memory at address 0x498000003aa>, end=0x0, 
    ch_out=ch_out@entry=0x7fffffffdc8c) at ../lib/the_Foundation/src/string.c:927
927     ../lib/the_Foundation/src/string.c: Datei oder Verzeichnis nicht gefunden.
(gdb) backtrace
#0  decodeBytes_MultibyteChar (bytes=0x498000003aa <error: Cannot access memory at address 0x498000003aa>, end=0x0, 
    ch_out=ch_out@entry=0x7fffffffdc8c) at ../lib/the_Foundation/src/string.c:927
#1  0x0000555555577460 in findLoc_GmRun (d=0x5555590381d0, pos=...) at ../src/gmdocument.c:2063
#2  0x00005555555774d4 in findLoc_GmDocument (d=<optimized out>, pos=...) at ../src/gmdocument.c:1910
#3  0x00005555555947b9 in sourceLoc_DocumentWidget_ (d=d@entry=0x5555556becc0, pos=...) at ../src/ui/documentwidget.c:2006
#4  0x0000555555594e1e in beginMarkingSelection_DocumentWidget_ (d=d@entry=0x5555556becc0, pos=...) at ../src/ui/documentwidget.c:3355
#5  0x000055555559df15 in processEvent_DocumentWidget_ (d=0x5555556becc0, ev=<optimized out>) at ../src/ui/documentwidget.c:3751
#6  0x00005555555f074b in dispatchEvent_Widget (d=d@entry=0x5555556becc0, ev=ev@entry=0x7fffffffe080) at ../src/ui/widget.c:1002
#7  0x00005555555ed4a0 in processEvent_Window (d=0x5555556ea3f0, ev=0x7fffffffe080, ev@entry=0x7fffffffe170) at ../src/ui/window.c:885
#8  0x0000555555567c05 in processEvents_App (eventMode=eventMode@entry=waitForNewEvents_AppEventMode) at ../src/app.c:1261
#9  0x0000555555568c57 in run_App_ (d=<optimized out>) at ../src/app.c:1367
#10 run_App (argc=argc@entry=1, argv=argv@entry=0x7fffffffe3d8) at ../src/app.c:1456
#11 0x0000555555561a30 in main (argc=1, argv=0x7fffffffe3d8) at ../src/main.c:78

[paper42]

I can still not reproduce this, could you report this (with the backtrace) upstream?

[ericonr]

Can you capture your screen while doing this? Knowing the exact movements could help in reproducing.

What does xbps-pkgdb -a say?

[sgn]

Recipe for crash in my side:

• Nagivate to gemini://senioradmin.de/GreyMode.gmi
• Fetch an image
• Click into the image

[sgn]

#0  0x000055f420f6ef11 in u8_mbtouc (n=18446726344084553426, s=0x10200000012e <error: Cannot access memory at address 0x10200000012e>, puc=0x7ffcf1bc484c, 
    puc@entry=0x7ffcf1bc480c) at /usr/include/unistr.h:233
#1  decodeBytes_MultibyteChar (bytes=0x10200000012e <error: Cannot access memory at ad
dress 0x10200000012e>, end=0x0, ch_out=ch_out@entry=0x7ffcf1bc484c)                   
    at ../lib/the_Foundation/src/string.c:927                                         
#2  0x000055f420edf540 in findLoc_GmRun (d=0x55f4239d4540, pos=...) at ../src/gmdocument.c:2063
#3  0x000055f420edf5b4 in findLoc_GmDocument (d=<optimized out>, pos=...) at ../src/gmdocument.c:1910
#4  0x000055f420efc899 in sourceLoc_DocumentWidget_ (d=d@entry=0x55f423860800, pos=...) at ../src/ui/documentwidget.c:2006
#5  0x000055f420efccde in beginMarkingSelection_DocumentWidget_ (d=d@entry=0x55f423860800, pos=...) at ../src/ui/documentwidget.c:3355
#6  0x000055f420f05ff5 in processEvent_DocumentWidget_ (d=0x55f423860800, ev=<optimized out>) at ../src/ui/documentwidget.c:3751
#7  0x000055f420f587eb in dispatchEvent_Widget (d=d@entry=0x55f423860800, ev=ev@entry=0x7ffcf1bc4c40) at ../src/ui/widget.c:1002
#8  0x000055f420f55530 in processEvent_Window (d=0x55f423122dc0, ev=0x7ffcf1bc4c40, ev@entry=0x7ffcf1bc4d30) at ../src/ui/window.c:885
#9  0x000055f420ecfca5 in processEvents_App (eventMode=eventMode@entry=waitForNewEvents_AppEventMode) at ../src/app.c:1261
#10 0x000055f420ed0cf7 in run_App_ (d=<optimized out>) at ../src/app.c:1367           
#11 run_App (argc=argc@entry=1, argv=argv@entry=0x7ffcf1bc4ee8) at ../src/app.c:1456  
#12 0x000055f420ec9a30 in main (argc=1, argv=0x7ffcf1bc4ee8) at ../src/main.c:78      

[paper42]

I can not get lagrange to crash anywhere, is it still happening with 1.7 (#33276)? cc @skyjake

[skyjake]

I am able to reproduce this crash. Looking into a fix...

This seems to correct the issue: https://git.skyjake.fi/gemini/lagrange/commit/1fe2b10bc1a2def2af86390272907206f819b7ae

[sgn]

@skyjake The patch indeed fixes the issue.