Ignore large block of text

[AlexFeren]

Hello!
Fortigate configuration file has a large block of text containing PEM-formatted certificate and key, which get updated at every backup.

For example:

Code: Select all

        set private-key "-----BEGIN OPENSSH PRIVATE KEY-----
b3BlbnNzaC1rZXktdjEAAAAACmFlczI1Ni1jdHIAAAAGYmNyeXB0AAAAGAAAABCGR2b7mr
:
:
qcRimIuw==
-----END OPENSSH PRIVATE KEY-----"

I'd like ExamDiff ignore it. How can I do?
R's, Fern.

[psguru]

If these lines have consistent line numbers (say from line 1 to line 7), you can ignore then by line ranges (under Options | Text Comparison).

[AlexL]

You publushed your question in the forum of ExamDiff Pro, but the question wording mentions ExamDiff.
I'll assume that you meant ExamDiff Pro.
In this case, I would try the following:
Menu -> Tools -> Options... -> Text Comparison -> Ignore lines -> Line ranges
And specify there the exact amount of lines to ignore, I think it's 7 lines in your case.
By the way, when you have entered that configuration menu, you can use <F1> button to get help about format of specifying the number (or even range) of ignored lines.
Don't forget to remove this setting when you compare other files which do not need such ignoration. Settings are saved.

[AlexFeren]

Hello! No, the line numbers for these blocks of text aren't static - varying each time the configuration backup is generated. Moreover, their length isn't static - hence PEM format specifies the delimiters. Moreover, there are multiples of these spread out through the configuration file...

[psguru]

Then another option is to use the ignore regular expressions option, if you can come with one that is. These can be multi-line.

[AlexFeren]

> Then another option is to use the ignore regular expressions option, if you can come with one that is. These can be multi-line.

Surely, a ExamDiff can make this easier by implementing "Ignore multi-line text bounded by start and end text" feature?

[psguru]

Yes, it would be but it's never come up until now. For now, you can create a new document type and define your own "comments" for it using a regular expression, like this for C comments:

Code: Select all

/\\*.*?\\*/

[AlexFeren]

Hi... Can you propose the regular expression given the specification in previous post?

[psguru]

Are there unique start/end words that delimit your "comments"? If so, use something like:

Code: Select all

word1.*?word2

[AlexFeren]

Are there unique start/end words that delimit your "comments"?

This:

Code: Select all

        set private-key "-----BEGIN OPENSSH PRIVATE KEY-----
b3BlbnNzaC1rZXktdjEAAAAACmFlczI1Ni1jdHIAAAAGYmNyeXB0AAAAGAAAABCGR2b7mr
:
:
qcRimIuw==
-----END OPENSSH PRIVATE KEY-----"

[AlexFeren]

For now, you can create a new document type and define your own "comments" for it using a regular expression, like this for C comments:

Suggestions?

[psguru]

I gave you a suggestion earlier: use a regex like this

Code: Select all

word1.*?word2

where word1/2 are unique starting/ending texts of the ignored block.

[JeremyNicoll]

Perhaps another option would be to write a simple 'plugin' which strips out the blocks of lines
you're not interested in. Possibly you'd want the plugin to leave a single line where each such
block had been saying "There was a key definition here." or even (if the number of them
matters) "Key definition nnnnn was here."

[AlexFeren]

I gave you a suggestion earlier: use a regex like this

Code: Select all

word1.*?word2

where word1/2 are unique starting/ending texts of the ignored block.

Hi! I already answered that - showing exact syntax of the block and no - it don't work. This warrants a dedicated feature in ExamDiff.

[psguru]

Please provide screenshots illustrating the issue.