* $* assignment regression in shwordsplit @ 2017-04-02 17:45 Martijn Dekker 2017-04-02 17:49 ` Martijn Dekker 2017-04-02 17:58 ` Daniel Shahaf 0 siblings, 2 replies; 4+ messages in thread From: Martijn Dekker @ 2017-04-02 17:45 UTC (permalink / raw) To: Zsh hackers list In the current zsh development version[*], there is a regression with assigning "$*" to a variable in shwordsplit mode when IFS is either unset or empty. The "$*" is only expanded the first time it occurs. $ Src/zsh -f -c 'set -- one two; unset IFS; var=$*/$*; echo $var $*/$*; setopt shwordsplit; var=$*/$*; echo "$var" "$*/$*"' one two/one two one two/one two one two/$* one two/one two $ Src/zsh -f -c 'set -- one two; IFS=; var=$*/$*; echo $var $*/$*; setopt shwordsplit; var=$*/$*; echo "$var" "$*/$*"' onetwo/onetwo onetwo/onetwo onetwo/$* onetwo/onetwo Thanks, - M. [*] as of commit eaeebeb8bb80cd7e2f5c703c7da31f44a1ab3182 ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: $* assignment regression in shwordsplit 2017-04-02 17:45 $* assignment regression in shwordsplit Martijn Dekker @ 2017-04-02 17:49 ` Martijn Dekker 2017-04-02 17:58 ` Daniel Shahaf 1 sibling, 0 replies; 4+ messages in thread From: Martijn Dekker @ 2017-04-02 17:49 UTC (permalink / raw) To: Zsh hackers list Op 02-04-17 om 19:45 schreef Martijn Dekker: > $ Src/zsh -f -c 'set -- one two; IFS=; var=$*/$*; echo $var $*/$*; > setopt shwordsplit; var=$*/$*; echo "$var" "$*/$*"' > onetwo/onetwo onetwo/onetwo > onetwo/$* onetwo/onetwo Correction, the output of this one is actually: onetwo/onetwo one two/one two onetwo/$* onetwo/onetwo Had accidentally left the output of a previous version of the test that quoted all 'echo' arguments even before setting shwordsplit. This correction is just for accuracy, it's not actually relevant to the bug report. - M. ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: $* assignment regression in shwordsplit 2017-04-02 17:45 $* assignment regression in shwordsplit Martijn Dekker 2017-04-02 17:49 ` Martijn Dekker @ 2017-04-02 17:58 ` Daniel Shahaf 2017-04-02 20:15 ` Bart Schaefer 1 sibling, 1 reply; 4+ messages in thread From: Daniel Shahaf @ 2017-04-02 17:58 UTC (permalink / raw) To: zsh-workers Martijn Dekker wrote on Sun, Apr 02, 2017 at 19:45:14 +0200: > In the current zsh development version[*], there is a regression with > assigning "$*" to a variable in shwordsplit mode when IFS is either > unset or empty. The "$*" is only expanded the first time it occurs. > > $ Src/zsh -f -c 'set -- one two; unset IFS; var=$*/$*; echo $var $*/$*; > setopt shwordsplit; var=$*/$*; echo "$var" "$*/$*"' > one two/one two one two/one two > one two/$* one two/one two > > $ Src/zsh -f -c 'set -- one two; IFS=; var=$*/$*; echo $var $*/$*; > setopt shwordsplit; var=$*/$*; echo "$var" "$*/$*"' > onetwo/onetwo onetwo/onetwo > onetwo/$* onetwo/onetwo Both of these bisect to: commit 74fe4d0950d5db0bba9d8ec182c4a827728cff60 Author: Barton E. Schaefer <schaefer@zsh.org> Date: Mon Feb 20 13:22:55 2017 -0800 40598: paramsubst() should always return scalar when PREFORK_SINGLE was passed :100644 100644 0d3352a8e807172c7adcde89a7a769c1c704320c 60a5164d638eba63dcef57a7e7bbf7951ada4426 M ChangeLog :040000 040000 843e4c1065cbcdb63fb554ab97de69741351f78c e3d9d89d533c0cd63a7aaa0d714e232e024302ca M Src ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: $* assignment regression in shwordsplit 2017-04-02 17:58 ` Daniel Shahaf @ 2017-04-02 20:15 ` Bart Schaefer 0 siblings, 0 replies; 4+ messages in thread From: Bart Schaefer @ 2017-04-02 20:15 UTC (permalink / raw) To: zsh-workers Hrm. Well, we can do it this way instead, but that changes the effects of ${(u)...} and and ${^...} in the expansion, requiring an extra level of nested braces. Obviously in the bit removed by this patch it's the wrong thing to join up the entire input linklist. I guess we'd need to join only the nodes added around the input node n (plus turn off LF_ARRAY), but there's no good way to keep track of that (?). Maybe it would work to join up everything to the left of n. Anybody with a clearer idea what's going on here care to interject? Meanwhile this fixes the bug without failing the other related tests. diff --git a/Src/subst.c b/Src/subst.c index e639c96..5b1bf89 100644 --- a/Src/subst.c +++ b/Src/subst.c @@ -446,7 +446,7 @@ singsub(char **s) * NULL to use IFS). The return value is true iff the expansion resulted * in an empty list. * - * *ms_flags is set to bits in the enum above as neeed. + * *ms_flags is set to bits in the enum above as needed. */ /**/ @@ -3779,6 +3779,13 @@ paramsubst(LinkList l, LinkNode n, char **str, int qt, int pf_flags, * as a scalar.) */ + if (isarr && ssub) { + /* prefork() wants a scalar, so join no matter what else */ + val = sepjoin(aval, NULL, 1); + isarr = 0; + l->list.flags &= ~LF_ARRAY; + } + /* * If a multsub result had whitespace at the start and we're * splitting and there's a previous string, now's the time to do so. @@ -4026,18 +4033,6 @@ paramsubst(LinkList l, LinkNode n, char **str, int qt, int pf_flags, y = dupstring(nulstring); setdata(n, (void *) y); } - if (isarr && ssub) { - /* prefork() wants a scalar, so join no matter what else */ - LinkNode tn; - - aval = hlinklist2array(l, 0); - val = sepjoin(aval, NULL, 1); - n = firstnode(l); - for (tn = lastnode(l); tn && tn != n; tn = lastnode(l)) - uremnode(l, tn); - setdata(n, (void *) val); - l->list.flags &= ~LF_ARRAY; - } if (eval) *str = (char *) getdata(n); ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2017-04-02 20:15 UTC | newest] Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-04-02 17:45 $* assignment regression in shwordsplit Martijn Dekker 2017-04-02 17:49 ` Martijn Dekker 2017-04-02 17:58 ` Daniel Shahaf 2017-04-02 20:15 ` Bart Schaefer
Code repositories for project(s) associated with this public inbox https://git.vuxu.org/mirror/zsh/ This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).