Issue #469 reported that using the MonitorInputStream counted the bytes twice. This was fixed (in commit f01154a) but apparently not included in the v4.0 branch, so the bug remains in all v4 versions.
Another fix to this class (commit 554053c) to avoid counting -1 was also not merged to v4.
Unit tests for these fixes were also not merged and are missing.
Compare the MonitorInputStream blame for v3.14.5 with blame for v4.7.0.
Could the same fixes be made to v4? Ideally as 4.6.2, as we still need to support Jackson 2.
Bonus question:
There's also a MonitorInputStream class in kong.unirest.core.java - is this something different or meant as a replacement? Shouldn't the unit tests from v3 also cover this class?
Issue #469 reported that using the MonitorInputStream counted the bytes twice. This was fixed (in commit f01154a) but apparently not included in the v4.0 branch, so the bug remains in all v4 versions.
Another fix to this class (commit 554053c) to avoid counting -1 was also not merged to v4.
Unit tests for these fixes were also not merged and are missing.
Compare the MonitorInputStream blame for v3.14.5 with blame for v4.7.0.
Could the same fixes be made to v4? Ideally as 4.6.2, as we still need to support Jackson 2.
Bonus question:
There's also a MonitorInputStream class in kong.unirest.core.java - is this something different or meant as a replacement? Shouldn't the unit tests from v3 also cover this class?