Feature request: Ignore unimportant *part* lines

[piedmont]

Hi,

Would this be at all possible:

The (awesome) 'Lines to Omit' feature allows entire equivalent/unimportant lines to be omitted from the compare.

We have some cases where a line may contain both unimportant & important differences. For example the following can be equivalent:

CaptionML=ENG=Continued;
CaptionML=ENU=Continued;

However the following are definitely not equivalent:

CaptionML=ENG=Continued;
CaptionML=ENU=Carried over;

I wonder if it could be possible to specify a regular expression for part of the line to be ignored, rather than the whole line....?

Piedmont

[jeffhostetler]

Sorry, but we don't currently support that.
Omitting content within a line (or within a
fixed range of columns) has been requested
a couple of times.

If you have a specific pattern like that, you
might be able to mark the "=ENU=" and
"=ENG=" as UNIMPORTANT COMMENTS
that can then be highlighted/not-highlighted
using the menu. They will still affect the
vertical alignment of the files, but just won't
be drawn in color. It's a bit of a stretch, but
if you have a limited set of them, it might be
worth a try.

jeff

From the too much information department:
To avoid confusion, I should point out that
there's a subtle distinction within DiffMerge
between omitting content from the analysis
vs not highlighting unimportant (observed/
analyzed) changes. And unfortunately, the
word "ignore" can refer to either type. I know
I've been sloppy at times on the forum here
in my choice of words. (And can be halfway
thru answering a question before I realize I'm
doing it again.) What it sounds like you're
asking for is being able to OMIT text within
a line before it is analyzed. That I currently
can't do. The best I can do is maybe let you
mark the regions UNIMPORTANT.

Hope that helps.
j

[piedmont]

hi Jeff

That's incredibly useful - I hadn't yet come across the "Content Handling"/"Matched Contexts" section of Rulesets as a way to define equivalence. Thank you

It gives exactly what i need in file windows.

Possibly getting greedy now, but is there any way to make it apply to folder windows..?

Piedmont

[jeffhostetler]

Thanks. And glad that you were able to get that to work for you.

WRT folder windows, the best I can do is "equivalence" mode.
In each Ruleset there is a page that lets you declare that changes
in whitespace, case, and line endings are to be treated as "equivalent".
(I'm avoiding the words "ignored" and "unimportant" again.)
This is on a per-Ruleset basis.

Then in the "Equivalence Mode" page under "Folder Windows",
there are settings to tell it what level of equivalence testing
you want it to do and whether to use the per-Ruleset settings.
(It can be a tad expensive, so it might not be on by default.)

Then file pairs which only have such whitespace/case/EOL differences
are marked equivalent "~" rather than "!=". And there's a toolbar
button to let you hide them if you want.

That's the best I can do. The equivalence testing won't use the
"important/unimportant" settings. (It would have really have slowed
things down during the scanning of the tree; not saying it won't do
that in the future, but it doesn't currently.)

j