Raise CaptionReadSyntaxError for malformed SRT instead of IndexError

[eeshsaxena]

SRTReader().read() leaks a bare IndexError (or ValueError) on a few kinds of malformed SRT, rather than one of the CaptionRead* errors callers expect to catch:

from pycaption import SRTReader
SRTReader().read("1\n")               # IndexError: list index out of range
SRTReader().read("1\n00:x:01,000 --> 00:00:04,000\ntext\n")  # ValueError

The first is a caption number with no timing line after it, the second a non-numeric timestamp. A timing line missing the --> arrow, or a stamp with too few fields, hit the same paths. I guarded the timing-line lookup and wrapped the timestamp parsing so they raise CaptionReadSyntaxError like the other readers do. Added a parametrized test for the malformed cases; full suite still passes.

[OlteanuRares]

Thanks for the fix — this looks good! One small thing before we merge: could you rebase onto the latest main and update the changelog version from 2.3.8 to 2.3.9? (2.3.8 has already been released.)

[eeshsaxena]

@OlteanuRares thanks for the review. Done: rebased onto the latest main and moved the changelog entry into a new 2.3.9 section, since 2.3.8 is already released. The conflict is cleared and it's mergeable again, and the full test suite passes locally (619 passed). Ready whenever you have a chance.

[OlteanuRares]

Thanks for the rebase! I went ahead and pushed two small changes directly to your branch:

1. Rebased onto main — picks up Handle malformed SAMI without leaking ValueError/AttributeError #420 (SAMI reader fixes) and resolves the changelog conflict.
2. Added detect() bounds check — SRTReader().detect("1") was still crashing with IndexError since lines[1] is accessed without a length check. Added the two-line guard from your Raise CaptionReadSyntaxError for malformed SRT instead of IndexError/ValueError #417 plus a parametrized test.

All 24 SRT tests pass. Merging now — thanks for the contribution!