Some HTTPS links try to download the whole file?

[robertaboukhalil]

URL: https://ftp-trace.ncbi.nlm.nih.gov/ReferenceSamples/giab/data_somatic/HG008/Liss_lab/BCM_Revio_20240313/HG008-T_PacBio-HiFi-Revio_20240313_106x_GRCh38-GIABv3.bam

It seems to be downloading the whole bam file (says Pending in network tab for a few mins):

---

This is caused by the web server not properly returning Accept-Ranges: bytes from browser requests:

await fetch(url, {method:"HEAD"}).then(d => Object.fromEntries(d.headers))
// {content-length: '102324808906', content-type: 'application/x-bam', last-modified: 'Thu, 27 Jun 2024 18:31:13 GMT'}

In those cases, Emscripten will download the whole file (source)

Yet the server does support Range requests:

await fetch(url, {method:"HEAD", headers: { "Range": "bytes=0-10" }})
// ... status: 206, 

And using curl, the server returns the correct headers:

curl -I $URL
HTTP/1.1 200 OK
...
Accept-Ranges: bytes
Content-Length: 102324808906
...

To do: before loading the BAM file, make a HEAD request to check for the header