This is easy enough (IIUC your problem): when using #nextLine while reading from a stream, all 3 EOL conventions are handled transparently, you just get the line's contents back until you are done. Then you write the lines back out with your preferred EOL convention.
On 18 Aug 2016, at 20:41, stepharo <stepharo@free.fr> wrote:
Hi
for the mooc I'm working on a srt to vtt converter.
1 00:00:07,040 --> 00:00:10,440 Hello. This week, we'll get to the heart of the matter,
2 00:00:10 600 --> 00:00:12,160 about syntax especially.
into
WEBVTT
00:00:07.040 --> 00:00:10.440 align:middle Hello. This week, we'll get to the heart of the matter,
00:00:10.600 --> 00:00:12.160 align:middle about syntax especially.
It works more or less. Now I face the problem that the files people provided me have different encodings. (I guess) because when I do not treat the input (for example withLinuxLineEndings) I get some CRs after the conversion eventhough I copy some file content and all the line ending I output are lf (or can be customizable.
I cannot apply garbage in gabrage out because the files should work.
So I thought that I should just convert first the string I read using withLinuxLineEndings so that all cr, crlf are converted into lf. But since files have different encodings I end up something to issues too many lf.
Does any of you have an idea how to handle this.
I did not find a way to know the encoding of a file (not the bom) just the file ending.
Stef