One idea for translation is that the translation should convey "what makes x x." In other words, if a poet has a certain number of distinguishing characteristics, and these are on display in the source text, then these same characteristics should be on play in the translation.
Me moriré en París con aguacero,
un día del cual tengo ya el recuerdo.
Me moriré en París —y no me corro—
tal vez un jueves, como es hoy, de otoño.
Jueves será, porque hoy, jueves, que proso
estos versos, los húmeros me he puesto
a la mala y, jamás como hoy, me he vuelto,
con todo mi camino, a verme solo.
César Vallejo ha muerto, le pegaban
todos sin que él les haga nada;
le daban duro con un palo y duro
también con una soga; son testigos
los días jueves y los huesos húmeros,
la soledad, la lluvia, los caminos...
So, without even translating this, let's look for some characteristics.
The name of a specific bone (húmeros), where most poets would talk about bones or limbs in general. Vallejo liked very precise scientific names for things.
There's a colloquialism running through the poem, but it's not simply an imitation of how people talk, but a sort of "twisting motion." The reflexive verb of "me corro" for example. It means not, "I run" but "I accelerate." It can also mean ejaculate. "I'm in no hurry to shoot my load." ??
The grammar we taught in school would prescribe "le pegagan / todos sin que él les hiciera nada." He's mixing up the verbal tenses. (The poet already has memory of the future, in the second line.) The syntax is deliberately "roughed up."
There is a linguistic patterning: a use of six reflexive verbs in the quatrains.
The deictic situation, the here and now, is very front-and-center in this poem. The particular kind of staging of the poetic "I."
The rhythms scan, but are jerky. Enjambment is prominent. There are many pauses within a verse. It is a sonnet, but the rhymes are assonantal and irregular: AABB BAAB CCD EDD.
There is neologism and verbal play: "I prose / these verses." Soga is a rope, but also a whip and a noose. There's a verbal parallelism with two redundant noun modifiers: "los días jueves y los huesos húmeros" the Thursday days and the humeri bones.
So those are some features of this poem that make it Vallejo-like. We don't even know this unless we've read other poems by him. A good rule to follow is if there is a figure of diction, like asyndeton in the final line, that the poet has used it not accidentally.
Eshleman gets most of it right. He keeps the roughness but misses a lot of small details:
I will die in Paris, with hard dirty rain [with downpour]
one day I now remember. [why not already? That's the whole point]
I will die in Paris — and I don't run — [difference in meaning with reflexive verb?]
maybe a Thursday, like today, in autumn.
Thurday, because today, Thursday, when I prose
these lines, I have forced my humeri on [by saying "these lines" you miss the prose / verse antithesis. Why not "I prose / this verse"?]
unwillingly and, never like today have I again, [unwillingly for "a la mala": not as direct or foreful]
with all my road, seen myself alone. [missing parallelism between "me he puesto" / "me he vuelto"]
Cesar Vallejo is dead, they beat him [has died; the perfective aspect. Don't you think Vallejo used a particular aspect of the verb deliberately? Also, "they used to beat him"]
everyone, without him doing anything to them;
they hit him hard with a stick and hard
likewise with a rope; witnesses are [noose?]
the Thursdays and the humerus bones, [the Thursday days...]
the loneliness, the rain, the roads...
Is this too picky? There is no such thing, unless you think that what gives x its characteristic xness doesn't matter.