Feedback on 0.3.8 - bugs

Tue Aug 9 18:29:45 CEST 2005

OK, I have been having another big bout of therion use. This has generated a
few bugs, some observations for improvements and some more queries.

I'll split these up into a few messages to avoid one huge list.

The result so far is a 30km system in one plan which may be one of the
largest therion maps drawn so far?

A new expedition is going in 3 weeks so there is a lot of things I need to
sort out before they go. Your help will be appreciated as ever - I hope
everyone isn't on holiday....

OK. lets start with a few bugs, which are not major but should be fixed (I
understand why some of these things are happening and that they may be hard
to fix but I am just being a 'user' here - these things need to be improved
somehow to make therion useable by non-geeks):

1) If you split a line the view moves to be centered on the previous/next
line in the list (not sure which), whilst the selection moves to one half of
the line. Sometimes this means the view moves so you can't see the selected
line. The view should move to the line that is now selected (one half or the
other of the split line).

2) If you try to open a new image it asks you to create a .th2 file first
presenting "(new file)" as the file name. If you exit without choosing one
(cancel) then when you close the file in the map window it asks if you want
to save "blah/blah/(new file)". A blank white square also appears in place
of the image you didn't load. It should remain in the state it was before
you tried to load an image.

3) If you run out of disk space while Therion is processing then 
therion-in-xtherion just blanks the log
window and says 'error' - no way to find out what error without runninng
therion in raw terminal mode. There it gives:
This is pdfeTeX, Version 3.14159-1.10b-2.1 (Web2C 7.4.5)
entering extended mode
(./data.tex{/usr/share/texmf/pdftex/config/pdftex.cfg} (./th_enc.tex)
(./th_texts.tex) (./th_resources.tex) (./th_fontdef.tex)
(./th_formdef.texpdfetex: data: No space left on device
therion: error -- pdftex exit code -- 256

The 'No space left on device' should appear in the processing window (or
some other way of working out what went wrong). 

4) 'Sort of bug' (english language bug rather than computer language):

There is a 'vtreshold' option. This makes no sense and presumably should be 
'vthreshold' as it refers to a threshold value. The thbook explanation
should have the spelling changed too.

5) tcl/tk problem: If you maximise a window then swap to 'F2' window the
menus disappear (I think they may be 'under' the window title bar). They can
be brought back by un-maximising the window. This is using icewm window
manager under Debian GNU/Linux and has been present for many versions.
I suppose I should try it with another window manager and see what happens.

6) Wish/xtherion still hangs regularly so I have to kill it. I save after
every line (and _during_ long lines(!)) or boulder drawn to avoid losing a
lot of work. Very tiresome. No pattern to it at all - sometime minutes,
sometimes hours, sometimes days, doing various things. I don't know if it
happens outside xtherion as I don't use anything else tcl/tk for hours on
end. I tried running xtherion under strace a couple of times and caught the
point where it hangs. I don't know if this is a clue. It seemed to crash
faster when running under strace, and the output suggests that it is some
kind of threading problem, perhaps in the underlying tcl/tk implementation?

This time it crashed whilst I was using another window (the only time I have
seen that - normally it only crashes when I am using xtherion - presumably
the presence of strace caused this). This is the last 30 lines or so of
output:
futex(0x804a8c0, FUTEX_WAKE, 1)         = 1
clock_gettime(CLOCK_REALTIME, {1123365892, 37854000}) = 0
futex(0x88b3590, FUTEX_WAIT, 12826, {0, 17829000}) = 0
futex(0x804a8c0, FUTEX_WAKE, 1)         = 0
gettimeofday({1123365892, 44121}, {4294967236, 0}) = 0
ioctl(5, FIONREAD, [32])                = 0
read(5, "\6\0\216\2306@!\26@\0\0\0\30\0\340\0\0\0\0\0\21\2\223\1"..., 32) = 32
write(4, "\0", 1)                       = 1
futex(0x804a8c0, FUTEX_WAKE, 1)         = 1
futex(0x88b3580, FUTEX_WAKE, 1)         = 1
futex(0x88b3590, FUTEX_WAIT, 12828, NULL) = -1 EAGAIN (Resource temporarily
unavailable)
futex(0x804a8c0, FUTEX_WAKE, 1)         = 0
gettimeofday({1123365892, 53678}, {4294967236, 0}) = 0
write(5, "5\30\4\0.\2\340\0b\1\340\0o\0\24\0F\0\5\0.\2\340\0\2\0"..., 200) =
200
gettimeofday({1123365892, 54293}, {4294967236, 0}) = 0
write(4, "\0", 1)                       = 1
gettimeofday({1123365892, 54442}, {4294967236, 0}) = 0
futex(0x804a8c0, FUTEX_WAKE, 1)         = 1
futex(0x88b3580, FUTEX_WAKE, 1)         = 1
clock_gettime(CLOCK_REALTIME, {1123365892, 55967000}) = 0
futex(0x88b3580, FUTEX_WAKE, 1)         = 0
futex(0x804a8c0, FUTEX_WAKE, 1)         = 0
gettimeofday({1123365892, 56132}, {4294967236, 0}) = 0
ioctl(5, FIONREAD, [32])                = 0
read(5, "\16\0\227\230b\1\340\0\0\0>\0\0\0\0\0\0\0\0\0.\2\340\0"..., 32) = 32
gettimeofday({1123365892, 56372}, {4294967236, 0}) = 0
gettimeofday({1123365892, 58108}, {4294967236, 0}) = 0
write(4, "\0", 1)                       = 1
futex(0x804a8c0, FUTEX_WAKE, 1)         = 1
futex(0x88b3580, FUTEX_WAKE, 1)         = 1
futex(0x88b3590, FUTEX_WAIT, 12831, NULL

Clearly it is getting stuck on a futex that never exits/returns/is released.
Any suggestions as to what I should ask to try and get to the bottom of this
are welcome. perhaps some other tcl/tk test I could run?

Wookey
-- 
Aleph One Ltd, Bottisham, CAMBRIDGE, CB5 9BA, UK  Tel +44 (0) 1223 811679
work: http://www.aleph1.co.uk/     play: http://www.chaos.org.uk/~wookey/