Feature Req: 'flexible_max_line_width'

[GOvEy1nw]

I'm not sure how easy this would be to implement, but while --standard works a treat for the most part, I've found that there are often times where a final word or two end up on their own section, for example:

81
00:07:40,330 --> 00:07:43,430
They're old, old Virginia money,
plantations, shipping and industry,

82
00:07:43,610 --> 00:07:43,910
arms.

I think the ability to flexibly assign 'padding' for 'max_line_width' parameter could solve this issue.

For example, setting --flex_line_width 3 would allow a buffer of up to 3 additional words on the second line (beyond the set max_line_width value), but only if a full stop occurs within those 3 words. So in the case of:

81
00:07:40,330 --> 00:07:43,430
They're old, old Virginia money,
plantations, shipping and industry,

82
00:07:43,610 --> 00:07:43,910
arms.

--flex_line_width 3 would make it:

81
00:07:40,330 --> 00:07:43,910
They're old, old Virginia money,
plantations, shipping and industry, arms.

I hope that makes sense?

[Purfview]

This thing was already discussed in some issue.

--flex_line_width 3
I hope that makes sense?

Current splitting code doesn't look into the future, so such option doesn't fit.
About sense of such option, IMO not much, if you OK with longer lines then increase 'max_line_width' to that value, of course the issue of short cutoffs would be still present.

Theoretically something could be implemented so lines would be split at more balanced midpoint, for example it would look something like that:

81
00:07:40,330 --> 00:07:43,430
They're old, old Virginia money,
plantations, shipping

82
00:07:43,610 --> 00:07:43,910
and industry, arms.

It would solve the issue of short cutoffs, but there would still be the problem of awkwardly split subtitles, so it doesn't eliminate the need for human editing.
Probably, this can't be solved by simple code and would need some AI to do the splitting job.