Using Regex Find/Change to change by Unicode

[oston]

Many thanks for all this help.
These are old files that I am given to work with. Indesign handles then with no problems and I can handle the problem of these old fonts in InDesign.

I just thought that I was doing something wrong with these searches in Sigil, but apparently not, since you say that supporting \x{0101} would need a code change.
I tried Find: \x01x01, but I must have misunderstood your workaround because that did not work.

But no need to do anything further with this. no need for support for \x{0101} just for this one case.
Thanks, everyone for all the help and suggestions.

[KevinH]

Yes, I will add support for \x{BLAH} where BLAH is either a string of 2, 4, or 6 hex chars.
With 6, I will convert it to uint32 and use fromUcs4 to build the replacement string.

[oston]

Originally Posted by KevinH

Yes, I will add support for \x{BLAH} where BLAH is either a string of 2, 4, or 6 hex chars.
With 6, I will convert it to uint32 and use fromUcs4 to build the replacement string.

Thanks, very much, Kevin.

[KevinH]

Okay, I just pushed a fix for this to master.

The following *replacement* escapes will now be recognized:

\x{hh}
\x{hhhh}
\x{hhhhhh}

where h is any hex character (0-9, a-f, A-F).

Note 1: the 6 hex character version is there to support the extended plane above the BMP allowing the first two hex chars to be less than or equal to "10" to indicate the plane. This may result in QChar pairs as needed to properly encode the entire value.

Note 2: using < 2, 3, 5, >6 hex chars will result in that escape code being ignored and treated just like any other string of characters so remember to prepend a "0" when needed to fit one of the required formats.

[oston]

Thanks, Kevin. I installed V 2.1, but the replace still replaces á with the string \x{0101}, rather than with ā.
See the screen capture.
Is there something else I needed to do?

[KevinH]

The fix was pushed to master just now. If you build your own from source you can see and test the results. If not, I can post a CI build of Sigil I just made if you let me know what platform, assuming Windows or Mac.

[oston]

Thanks, Kevin. I am not familiar with the process to build my own executable. I am running Sigil on Windows 11.

[KevinH]

I will create a pre-release CI build for you to test with.

[DNSB]

Originally Posted by oston

But when I try to replace with ā using the \x{0101} the á is replaced with the string \x{0101}

With the current Sigil version, you can use the \x{0101] in the find field but you will need to use the replacement character in the replace field.

Find: \x[0101]
Replace: ā

deleted comment about building from source

[KevinH]

And yes just putting the actual character in the replace field instead of its unicode value in hex works even before this change.

Either way .. here is a pre-release automated build of Sigil master for Windows. I have uploaded it to my personal github repo here:

https://github.com/kevinhendricks/Bu...r_sigil_1.0.0/

You need to grab the file called:

sigil-installer.zip

Then unzip it and you should see the Sigil Windows installer.

Please do report back if this works for you.


And anyone else on Windows that is willing to test Sigil and report back issues before our next release, feel free to grab and use the same one.

[Turtle91]

Originally Posted by DNSB

With the current Sigil version, you can use the \x{0101] in the find field but you will need to use the replacement character in the replace field.

Find: \x[0101]
Replace: ā

deleted comment about building from source

That's kinda what I was referring to in my previous post. Couldn't the find and replace be done using "normal", string search, mode instead of "regex"??

Find: á
Replace: ā

or

Find: \x[0101]
Replace: ā

as DNSB suggested.

It's always better to make Sigil consistent with itself because it reduces confusion, but sometimes simple gets the job done better/faster/stronger.

[KevinH]

Yes that normal replacement would work as well.

[DiapDealer]

Originally Posted by KevinH

Okay, I just pushed a fix for this to master.

The following *replacement* escapes will now be recognized:

\x{hh}
\x{hhhh}
\x{hhhhhh}

where h is any hex character (0-9, a-f, A-F).

Note 1: the 6 hex character version is there to support the extended plane above the BMP allowing the first two hex chars to be less than or equal to "10" to indicate the plane. This may result in QChar pairs as needed to properly encode the entire value.

Note 2: using < 2, 3, 5, >6 hex chars will result in that escape code being ignored and treated just like any other string of characters so remember to prepend a "0" when needed to fit one of the required formats.

Working great on Windows and Linux!

[oston]

Originally Posted by KevinH

Either way .. here is a pre-release automated build of Sigil master for Windows. I have uploaded it to my personal github repo here:

https://github.com/kevinhendricks/Bu...r_sigil_1.0.0/

...

Please do report back if this works for you.

Installed V 2.2, Kevin. The search and replace with {HHHH} works perfectly.
Many thanks indeed,

[KevinH]

Great. Glad to hear it is working.
Thanks to all who tested and reported back.