playing with backreferences in list-colors

[Alexandre Duret-Lutz <duret_g@xxxxxxxx>]

I have been playing with list-colors to colorize 
process listing and so on.  This is quite fun.
The snipsets below show two problems I had
1) patterns containing letters don't seems to match;
2) backreferences when (x)# style patterns don't match make zsh segfault.


% zsh -f
phobos% unset {,Z}LS_COLO{,U}RS          
phobos% autoload -U compinit; compinit; zmodload zsh/complist
phobos% zstyle ':completion:*' group-name ''
phobos% zstyle ':completion:*:files' list-colors '=(#b)(a)*=0=1'
phobos% setopt extendedglob 
phobos% touch a{a,b,c}foo
phobos% ls a<TAB>
aafoo   abfoo   acfoo                 # nothing is highlighted
phobos% zstyle ':completion:*:files' list-colors '=(#b)(?)*=0=1'
phobos% ls a<TAB>          
aafoo   abfoo   acfoo                 # first letters are highlighted
phobos% zstyle ':completion:*:files' list-colors '=(#b)a(*)=1=0'
phobos% ls a<TAB>
aafoo   abfoo   acfoo                 # nothing is highlighted

Can't I put letters in those patterns?

And now :

phobos% zstyle ':completion:*:files' list-colors '=(#b)(a)#*=1=0'  
phobos% ls a<TAB>
zsh: segmentation fault (core dumped)  zsh -f

(I know, `(a)#' is weird, but actualy I would like to be able
to write something like `(*/)#([^ ]*)*' at the end of the pattern
for my processes listings, to colorize only basename of processes)

(gdb) bt
#0  0x80a7a37 in ztrsub (t=0x40014c50 "", 
    s=0x40018001 <Address 0x40018001 out of bounds>)
    at ../../last/Src/utils.c:2594
#1  0x8094d5c in pattryrefs (prog=0x40014a98, string=0x40014c58 "./aafoo", 
    nump=0x401e2150, begp=0x401e2160, endp=0x401e21a0)
    at ../../last/Src/pattern.c:1382
#2  0x401de36e in putfilecol (c=0x401e2100, group=0x814d470 "files", 
    n=0x40014c58 "./aafoo", m=33204) at ../../../last/Src/Zle/complist.c:585
#3  0x401de92c in clprintm (g=0x814d670, mp=0x814d6e0, mc=0, ml=0, lastc=0, 
    width=8, path=0x40014c58 "./aafoo", buf=0xbffff308)
    at ../../../last/Src/Zle/complist.c:703

[...]

Does string="./aafoo" in frame #1 means that `(#b)(a)#*' is matched
over "./aafoo" instead of only "aafoo"?

BTW zsh seems to go into troubles when such a pattern don't match :

~ % zsh -f
phobos% setopt extendedglob
phobos% [[ ab = (#b)(a)#* ]] 
phobos% echo $match
a
phobos% [[ .ab = (#b)(a)#* ]] 
phobos% echo $match          
zsh: bad pattern: (
phobos% [[ ./ab = (#b)(a)#* ]] 
phobos% echo $match           
zsh: bad pattern: (
phobos% [[ ab = (#b)(a)#* ]] 
phobos% echo $match         
a
phobos% [[ aafoo = (#b)(a)#* ]] 
phobos% echo $match            
a
phobos% [[ .aafoo = (#b)(a)#* ]] 
phobos% echo $match             
O
phobos% [[ ./aafoo = (#b)(a)#* ]] 
phobos% echo $match              
O
phobos% [[ ./aaffffffffffffffffffffoo = (#b)(a)#* ]] 
phobos% echo $match                                 
f
phobos% [[ .ab = (#b)(a)#* ]] 
zsh: segmentation fault (core dumped)  zsh -f

(sometime it's faster to coredump)

(gdb) bt
#0  0x80a7a37 in ztrsub (t=0x40014fc1 "O\001@", 
    s=0x40018001 <Address 0x40018001 out of bounds>)
    at ../../last/Src/utils.c:2594
#1  0x8094ed7 in pattryrefs (prog=0x810b650, string=0x400153f8 ".ab", 
    nump=0x0, begp=0x0, endp=0x0) at ../../last/Src/pattern.c:1417
#2  0x80949f0 in pattry (prog=0x810b650, string=0x400153f8 ".ab")
    at ../../last/Src/pattern.c:1286
#3  0x805c025 in evalcond (state=0xbffff6c0) at ../../last/Src/cond.c:206
#4  0x8064407 in execcond (state=0xbffff6c0, do_exec=0)
    at ../../last/Src/exec.c:2856

[...]

-- 
Alexandre Duret-Lutz