Significant .lyx EOL White Space Inquiry
Joel Kulesza
jkulesza at gmail.com
Sun May 31 17:30:56 UTC 2020
On Fri, May 29, 2020 at 9:49 AM Pavel Sanda <sanda at lyx.org> wrote:
> On Thu, May 28, 2020 at 10:27:54AM -0600, Joel Kulesza wrote:
> > Finally, significant EOL white space seems unidiomatic in a "source
> file,"
> > which is what a .lyx file effectively is.
> >
> > This is an expert-use situation, but *I wonder if we have a way to
> overcome
> > the need for significant EOL white space.*
>
> Well, it's all in the code, so one could certainly write patch witch moves
> the whitespace to the next line.
I understand. I was mainly asking to see whether there is a historical
reason for "needing" EOL white space, or whether it came along at some
point and has persisted.
> The issue is that any oversight in such
> patch would lead to catastrophic dataloss bugs for everyone else and one
> has to ask whether this cornercase is worth the risk.
>
Perhaps not, but it still begs the question whether steps are worth taking
to become a more idiomatic format.
> My take is that
> - if you mess with .lyx you need to respect its whitespacing and
> set your editor accordingly
>
A fair point, and that's what I've done. But, it's easy for someone to not
take this step and get burned by it at least once.
> - the real solution would be to actually do document comparison
> feature right, so one does not need to edit raw .lyx file...
>
Agreed. I see a few areas for improvement here. First, I've seen "Track
changes" flag false positives and give false negatives. So, improving
robustness in this markup would be good. Second: the ability to provide
comment/replies as wording is iterated upon would be valuable. That is the
main behavior I seek when performing document review: the ability to
provide specific actionable feedback (perhaps with discussion) on specific
components of the text. That's why a git-based review system is nice
(Gitlab, Github, Bitbucket, etc.).
Finally, improving the integration with revision tracking would be good. A
recent criticism I heard was "why do I have to count the number of
revisions back; why not just specify a commit/branch point to compare
with?". I explained why, but that is valid feedback from a git-centric
view. I wonder if specializing the version control behavior for git would
be valuable to augment the more generic interface currently available.
Thanks,
Joel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lyx.org/pipermail/lyx-devel/attachments/20200531/2be6e3ba/attachment.html>
More information about the lyx-devel
mailing list