> I've refactored the appropriate code, in src/roff/troff/input.cpp,
> with a view to accommodating this; see patch attached. While I've
> not yet progressed any implementation for PDF handling, I have
> indicated the point at which it should be invoked. Okay to commit,
> thus far?
This is very nice, thanks! For my taste, the comments are a bit
excessive, but I guess this is probably only me who thinks so :-)
Some comments, mainly regarding formatting:
+ while ((context = bounding_box_args()) == NULL
+ && get_line(DSC_LINE_MAX_ENFORCE) > 0)
+ ;
Please say
while ((context = bounding_box_args()) == NULL
&& get_line(DSC_LINE_MAX_ENFORCE) > 0)
;
+ while (*context == '\x20' || *context == '\t')
+ context++;
Please say
while (*context == '\x20' || *context == '\t')
context++;
+ status = (context_args("(atend)", context) == NULL)
+ ? llx = lly = urx = ury = PSBB_RANGE_IS_BAD
+ : PSBB_RANGE_AT_END;
Please say
status = (context_args("(atend)", context) == NULL)
? llx = lly = urx = ury = PSBB_RANGE_IS_BAD
: PSBB_RANGE_AT_END;
+ register size_t len = strlen(tag);
Given that recent versions of clang emit warnings if it encounters the
`register' keyword, and given that this keyword has no effect for
about 20 years, please don't use it:
size_t len = strlen(tag);
[It would be a trivial clean-up in the groff source code to remove
`register' everywhere.]
Werner
