DevLys 010 to Unicode Troubleshooting & Reference Guide
Use this guide when a DevLys conversion looks partly correct but punctuation, dates, conjuncts or mixed Word text are wrong. It is a DevLys 010 troubleshooting reference based on the repository's exact mapping and regression evidence, not a second converter landing page.
Start with the exact mapping: DevLys 010
SahayakTools verifies DevLys 010 only. DevLys 020, DevLys 030 and other similarly named variants are not accepted as interchangeable evidence. A family name alone does not prove that punctuation, extended glyphs or sequence rules use the same assignments.
Do not use this 010 converter for DevLys 020, 030 or an unknown variant. Stop before conversion; a few readable words are not enough to validate the rest of the document.
Exact tested examples you can use for diagnosis
The regression suite verifies representative DevLys 010 input such as Hkkjr → भारत, d+ → क़ and {k → क्ष. It also locks the legacy @ glyph position to the visible Unicode slash /.
The date case 25@05@2016 → 25/05/2016 is tested with structured-token protection both enabled and disabled. The reverse direction is also tested: Unicode 25/05/2016 must encode as 25@05@2016 inside a DevLys 010 run and decode back exactly.
These examples are checks for a source already identified as DevLys 010. They do not certify an unknown DevLys variant merely because one sample happens to match.
Mixed Word text needs font-run evidence
A Word paragraph can contain DevLys 010 Hindi beside Calibri or Arial English, Unicode Hindi, dates and IDs. In rich paste, an explicit DevLys 010 font run is evidence for legacy decoding; normal-font runs should remain unchanged instead of being treated as DevLys bytes.
Word can split one encoded word across several inline spans. Repository regressions split verified examples such as कि, क्ष, भारत, प्रेम and कृपा at every possible encoded boundary and require the same Unicode result. Contiguous DevLys 010 spans therefore have to be interpreted as one semantic run, not converted as unrelated fragments.
If formatting was stripped before paste, or a normal visible font conflicts with a hidden Word-only legacy hint, the mapping evidence is weaker. Return to the original document and inspect the actual font runs rather than forcing a conversion.
Punctuation and structured tokens are not generic ASCII
In DevLys 010, ASCII-looking code positions can be Hindi or punctuation glyphs. Verified examples include @ → /, - → ., ] → , and A → ।. Preserving every ASCII character before decoding would therefore corrupt real legacy text.
At the same time, high-confidence URLs, email addresses, paths, dates, amounts and IDs must survive mixed-document conversion. The converter protects those structured tokens while still allowing a legacy-source @ to reach the DevLys decoder when it is the slash glyph in 25@05@2016.
For Unicode-to-DevLys output, mixed Hindi, English, numbers or unsupported symbols can require separate font runs. Use Copy for Word when raw text export is unavailable, because plain text cannot preserve those assignments.
Common failure diagnosis
Wrong throughout: re-check the source font and version before changing any settings.
Hindi is right but English changes: the paste probably lost normal-font run boundaries, or ordinary Latin text was incorrectly marked as DevLys.
Words are mostly right but matras or conjuncts break: paste from the original Word document with formatting so adjacent DevLys spans can be processed together.
Dates, paths or punctuation change: compare the exact source run. A DevLys @ can be a visible slash, while an @ inside an email address has different context.
Only a few characters fail: do not hand-correct a large document until you have ruled out a different DevLys variant or vendor mapping.
Safe Word workflow
1. Keep an untouched copy of the original Word file and verify that the affected runs explicitly use DevLys 010.
2. Paste from Word with formatting into the DevLys 010 to Unicode page. Do not flatten the whole document to plain text first when it contains mixed fonts.
3. Compare names, dates, IDs, punctuation, matras and conjuncts with a correctly rendered source.
4. Copy the reviewed Unicode result into a separate working document. For the reverse direction, use Copy for Word whenever the result depends on mixed legacy and normal font runs.
When to stop rather than guess
Stop if the source version is unknown, the visible font does not match the claimed mapping, important characters differ from the rendered original, or the document lost the font-run evidence needed to separate legacy Hindi from normal text.
Preserve the original and identify the exact font/version before continuing. A plausible-looking conversion is not reliable evidence for official, archival or print-ready text.