Get Sequences by Id: an unknown id pastes NCBI's error text as if it were a record #12

Claude reports: when NCBI does not know an id, efetch still answers HTTP 200 with an error text, and Get Sequences by Id treats that text as a record.

Observed: selection NOTANID999. The console prints Downloaded NOTANID999 and the text pasted into the new tab is (NCBI's own wording, letters spaced out):

Error: F a i l e d  t o  u n d e r s t a n d  i d :  N O T A N I D 9 9 9

(Fetching https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=nucleotide&id=NOTANID999&rettype=gb&retmode=text directly returns the same text with status 200, so this is not a decoding problem.) The error dialog at line 171 only runs for IOError, so the user is never told the id was wrong, and a mix of good and bad ids gives one file with error lines between the records. This was seen with the insert workaround for #4 in place, because without it the tab stays empty.

Fix: in EutilsCall.run (line 159), check the returned text (for example it must start with LOCUS) and otherwise show the error dialog and set self.result = False.

Tested on Sublime Text build 4215 (portable, Windows), plugin host Python 3.14, installed through Package Control (BioPythonUtils 2017.12.18.15.00.45, which contains the fixes for #3). Line numbers are those of BioPythonUtils.py on master (de058c4).