Original File

This is how your browser is rendering the text as an inlined <pre>, containing UTF-8.

अनन्त: वै वेद: ।
कार्त्स्न्या ।
1.001.01a अ॒ग्निमी॑ळे पु॒रोहि॑तं य॒ज्ञस्य॑ दे॒वमृ॒त्विज॑म् ।
1.001.01c होता॑रं रत्न॒धात॑मम् ॥
1.001.01a a̱gnim ī̍ḻe pu̱rohi̍taṁ ya̱jñasya̍ de̱vam ṛ̱tvija̍m ।
1.001.01c hotā̍raṁ ratna̱dhāta̍mam ॥

View raw in browser

From what I can tell, none of the terminals I've tried are handling this properly. Even the ones that render the script decently don't seem to align the cursor properly.

yudit

As my basis for comparison of "correct" rendering, I'm using yudit. I've not used it extensively (a few times in the past while diagnosing Unicode issues); getting it set up for the fonts and colors was a pain (today); and, I think I'd be miserable if I had to use it instead of Vim. But, Yudit is a text editor specifically designed for handling many, many scripts, various input methods mixed LTR and RTL text, etc. etc. My ~/.yudit/yudit.properties file I used for testing

(good) diacritic placement on first Devanāgarī line
55 0 36 24
(bad) Latin-1 with double-combining diacritics
155 140 24 24
(bad) vertical spacing on Latin-1 with double-combining diacritics
455 145 24 24

rxvt-unicode

Not speaking a language that uses Devanāgarī, I'm not sure how distracting you'd find the gaps in the shiro-rekha (line above the letters that connects them). But, rxvt-unicode seems to do the best with the diacritics in both writing systems.

(good) diacritics placed well
585 75 25 30
(bad) broken line above Devanāgarī
0 45 150 10

mlterm

mlterm may be useful if you're editing *only* Devanāgarī script (Lohit Hindi apparently doesn't have roman characters), but its rendering appears really squished to me, and it misses the diacritic placement in the first line.

(bad) diacritics placed incorrectly
50 0 50 25
(bad) squished Devanāgarī
110 75 150 30
(bad) no fallback for Latin characters
90 145 140 30

gnome-terminal

gnome-terminal seems unable to find some combining chars[2] (which show up as dotted circles). The Devanāgarī diacritics that it can find are in the right place (first line). But it fails on the double-diacritic on the Latin char (r with dot and macron below -- the macron is below the following char).

(good) diacritics placed correctly
95 0 50 25
(bad) missing Devanāgarī combining chars
160 60 170 30
(bad) wrong spacing for multiple combining chars on Latin chars
250 130 25 30
(bad) wrong spacing for multiple combining chars on Latin chars (should both be above the 'a')
205 160 35 30
(bad) wrong spacing for multiple combining chars on Latin chars (should both be below 'r')
720 130 35 30

etc.

konsole (KDE 4.5? -- no screenshot) omits the Devanāgarī diacritics entirely.

terminator is roughly the same as gnome-terminal.

gvim

GVim on my machine (GTK2 GUI, 7.2.442) doesn't fare so well.

With the possible exception of Yudit, no program I tried (including Firefox) on my system seemed to quite properly handle:

\U0952 DEVANAGARI STRESS SIGN ANUDATTA

(It's the source of the dotted circles in gnome-terminal.)