Replace with GSUB

[Weolo]

GetString(SI_SMITHING_RESEARCH_IN_PROGRESS)
Returns "Researching..." in en language.
I want a small variation of this but rather than translate this text into other languages i would like to replace the 3 dots with colon, space but I can't get the regular expression right, could someone help out?

local test1 = string.gsub(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS),'%.%.%.',': ')
local test2 = GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):gsub('%.%.%.',': ')

[Sasky]

Something to do with the string returned from GetTranslateString()
Those aren't normal periods.

Did some poking around from the /script prompt:

d(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):sub(-1,-1))
> x
 
d(string.byte("."))
46
 
d(string.byte(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):sub(-3,-3)))
226
 
d(string.byte(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):sub(-2,-2)))
128
 
d(string.byte(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):sub(-1,-1)))
166

Which is rather disgusting.

To replace it, you need to do:

GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):gsub(string.char(226,128,166),': ')

Alternatively, considering the ": " itself is rather dependent about the ellipsis being at the end, you could just do something like:

GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):sub(1,-4) .. ": "

[circonian]

Originally Posted by Weolo GetString(SI_SMITHING_RESEARCH_IN_PROGRESS) Returns "Researching..." in en language. I want a small variation of this but rather than translate this text into other languages i would like to replace the 3 dots with colon, space but I can't get the regular expression right, could someone help out? local test1 = string.gsub(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS),'%.%.%.',': ') local test2 = GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):gsub('%.%.%.',': ')

Beats me, like Sasky said they ain't normal periods.
You use string.match to grab everything up to them though, then just add whatever you want to it.

EDIT: Merlights right about ^[A-z], It should have been:

string.match(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS), "^[a-zA-Z]+")..": "
 
-- or:
string.match(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS), "^%a+")..": "

But I also just realized you said "rather than translate", GetString(SI_SMITHING_RESEARCH_IN_PROGRESS) will automatically translate it. So it really sounds like you don't want to do any of this. If you don't want it translated just do:

local researchString = "Researching: "

[votan]

Just to explain this strange characters:
The byte sequence 226,128,166 is the UTF8 encoded data for the UNICODE char U+2026 "horizontal ellipse".
LUA (including all string methods) handles strings as ANSI (one byte char), while the client prints UTF8 encoded strings correctly in UNICODE.
This makes using regex patterns a bit tricky sometimes.

[merlight]

Originally Posted by circonian [A-z]

Never, ever, in any language, pattern matching engine or whatever, use this character range. In many regular expression guides, this is given as an example of how the range operator in character classes should not be used. It matches ASCII letters, but also brackets '[' and ']', backslash '\\', caret '^', underscore '_' and backtick '`'.

[TheDepe]

Originally Posted by circonian Beats me, like Sasky said they ain't normal periods. You use string.match to grab everything up to them though, then just add whatever you want to it. string.match(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS), "^[A-z]+")..": "

local test1 = string.gsub(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS),'[.]+',': ')

Or this way..

[merlight]

Originally Posted by TheDepe local test1 = string.gsub(GetString(SI_SMITHING_RESEARCH_IN_PROGRESS),'[.]+',': ') Or this way..

That replaces a sequence of periods (ASCII 46), not Unicode ellipsis, which this topic is about.

Anyway, how would you replace "foo"?

str = str:gsub("foo", "bar")

So, do the same with what you want to replace:

str = str:gsub("…", ": ") -- ellipsis, not three periods!

Or, if you're worried about having non-ASCII characters in your code:

str = str:gsub("\226\128\166", ": ")
-- highlighting fails

NB: I will probably never stop wondering why does Lua have to differ in such stupid ways, like using decimal base where everyone else uses octal.

[Weolo]

Wow I never expected it to be a combination of all those things.
I appreciate you all taking a look... interesting stuff.

I will keep the unicode replacement in mind for future things but for this problem I will keep it simple and go with the substring to rip off the last 3 characters

GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):sub(1,-4) .. ": "

[merlight]

Originally Posted by Weolo I will keep the unicode replacement in mind for future things but for this problem I will keep it simple and go with the substring to rip off the last 3 characters GetString(SI_SMITHING_RESEARCH_IN_PROGRESS):sub(1,-4) .. ": "

That's going to bite you sooner or later. Cutting UTF-8 strings at arbitrary byte positions is never a good idea

[circonian]

Originally Posted by merlight ^[A-z] Never, ever, in any language, pattern matching engine or whatever, use this character range. In many regular expression guides, this is given as an example of how the range operator in character classes should not be used. It matches ASCII letters, but also brackets '[' and ']', backslash '\\', caret '^', underscore '_' and backtick '`'.

Eew, your right, I corrected my post.

[Sasky]

Originally Posted by merlight That's going to bite you sooner or later. Cutting UTF-8 strings at arbitrary byte positions is never a good idea

Of course, the 'proper' solution would be to make fresh translator strings. Manipulating translated or formatted strings can bite you sooner or later.

The :gsub of the 3-byte sequence anchored to the end of string is probably the most robust:

str = str:gsub("\226\128\166$", ": ")
--When highlighting fails, kill the highlighting

It falls back to the original string if something goes wrong unlike the string split.

[Weolo]

Fair enough, always happy to take on other peoples advice.
I will use the robust solution from now on.
Thanks!