Update of bug #56500 (group groff):
Status: None => Need Info
_______________________________________________________
Follow-up Comment #3:
[comment #2 comment #2:]
> The restriction does not seem to hold in other *roffs, either: The ascii and
> ps outputs of Ralph's example are the same in Heirloom troff as in groff.
I'm not sure I buy Ralph's example as demonstrative of the issue.
$ cat recurse.tr
.de t0
t0a
.sp 2i
t0b
.br
..
.de t1
t1
.br
..
.wh 1i t0
.wh 2i t1
foo
.sp 4i
bar
The reason I say that is that there may be a confounding factor: when the
formatter exits normally, vertical motion to the bottom of the page occurs,
and that can spring traps. That's a deliberate design feature, so that the
last page of a document flushes its footnotes, writes a page footer, and so
forth.
Consider this alternative example.
$ cat ATTIC/56500.roff
.de T1
This text is from trap macro T1.
. sp 4i
. ab
..
.de T2
This text is from trap macro T2.
..
.wh 1i T1
.wh 2i T2
This is the top of the document.
.sp 1i
$ groff -a ATTIC/56500.roff
<beginning of page>
This is the top of the document.
This text is from trap macro T1.
$ cd ~/heirloom
$ ./bin/troff -a 56500.roff
./bin/troff: User Abort; line 12, file 56500.roff; page 1
T1
$ ./bin/troff 56500.roff
./bin/troff: User Abort; line 12, file 56500.roff; page 1
T1
$ cd ~/dwb
$ DWBHOME=. ./bin/troff -a 56500.roff
./bin/troff: User Abort; line 12, file 56500.roff
stack: T1
$ DWBHOME=. ./bin/troff 56500.roff
./bin/troff: User Abort; line 12, file 56500.roff
stack: T1
I have to admit I don't understand why DWB/Heirloom _troff_ fails to produce
_any_ output before dying horribly. That rather confounds my hypothesis of
confounding.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?56500>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
