Re: removing the .IX macro from the ms package in 1.23 breaks old documents

[joerg van den hoff]

On 03.10.24 23:21, G. Branden Robinson wrote:
> At 2024-10-03T23:06:18+0200, joerg van den hoff wrote:
> > table of content when using ms macros (wide entries failing to line
> > break and "pushing" the page number out of line to the right instead).
> > I cannot say, however, whether this is a 1.23-related issue or whether
> > it just has escaped my attention previously (would have to revert to
> > 1.22 to check, no time right now etc.).
>
> A sample document that reproduces this issue would be really helpful.

attached is an example. I tried to condense it and reduce those MHEAD,SHEAD macros to a minimum 
(they are doing more in my actual setup). so the missing SN numbers in the resulting TOC in this
example are not a bug :).

the example should be reproducible on A4 paper (otherwise play around with point size to trigger the 
bug(s)). document should be formatted with something like

groff -ms -Tps (or -Tpdf).

to summarize the issues:

1.
the most serious one is misalignment of page number column for entries which are long but seemingly 
fail to line break early enough. there are also instances where the line break just occurs too late
w/o causing page number column misalignment (this seems to be the case when no 
further ruler char
creeps in).

2.
a related issue it seems is long entries which just fit into the line w/o misaligning the page 
number but where ruler chars are still inserted and partly, in fact, overlap with last char from the 
preceding text. possibly this is the main issue for (1) and (2): emitting one too many ruler chars
beyond end-of-ruler in other entries (plus too late line breaking).

3.
very long entries trigger line breaks but second line does not honour indent of first line. this is 
IIRC probably not a regression but has "always" happened.

issues 1) and 2) above I can not remember to have seen previously (in my actual 150p document which 
I have maintained for 25 years) so my suspicion would be that it is a regression. but I might be wrong.


HTH,
joerg



> It could form the basis of a regression test; we have about 220 of those
> now, 217 of which I wrote.  Their reason for being is to prevent the
> sorts of problems you and Deri are concerned about.
>
> Regards,
> Branden

tocbug.trf
Description: Text document