Sérgio Carvalho
@sergiosgc.com
Geek com opiniões. Refugiado do Twitter. @sergio@sergiosgc.com
It's not quite there yet. I tried logging in with my handle, and got this.
October 16, 2025 at 7:56 PM
It's not quite there yet. I tried logging in with my handle, and got this.
Evoluída. A minha chamava para eu lhe abrir a porta que já estava entreaberta. No sentido fácil 🤦
No sentido difícil ainda foi preciso tempo a mostrar como fazer. Tudo ainda mais estranho porque no geral é uma cadelita bastante esperta.
Se calhar é só respeitadora da privacidade...
No sentido difícil ainda foi preciso tempo a mostrar como fazer. Tudo ainda mais estranho porque no geral é uma cadelita bastante esperta.
Se calhar é só respeitadora da privacidade...
October 11, 2025 at 10:01 AM
Evoluída. A minha chamava para eu lhe abrir a porta que já estava entreaberta. No sentido fácil 🤦
No sentido difícil ainda foi preciso tempo a mostrar como fazer. Tudo ainda mais estranho porque no geral é uma cadelita bastante esperta.
Se calhar é só respeitadora da privacidade...
No sentido difícil ainda foi preciso tempo a mostrar como fazer. Tudo ainda mais estranho porque no geral é uma cadelita bastante esperta.
Se calhar é só respeitadora da privacidade...
No cheating. Post the last picture of your pet.
July 20, 2025 at 9:56 PM
No cheating. Post the last picture of your pet.
Ainda bem que temos vários canais de informação
May 18, 2025 at 6:24 PM
Ainda bem que temos vários canais de informação
Cubro a aposta e subo com a primeira tentativa de fazer pãozinhos de queijo. Bem bons.
May 16, 2025 at 9:22 PM
Cubro a aposta e subo com a primeira tentativa de fazer pãozinhos de queijo. Bem bons.
Cubro a aposta e subo com uma meia geek
March 26, 2025 at 5:15 PM
Cubro a aposta e subo com uma meia geek
#AdventOfCode 2024
Day 6: Guard Gallivant
Actually, opted for Zig, not Nix for the next language. Nice language ecosystem. Good compiler errors, good docs, decent forum coverage. Allocators merit some opinion; later.
The problem is just a walking simulator. Brute-forced part two, shamelessly.
Day 6: Guard Gallivant
Actually, opted for Zig, not Nix for the next language. Nice language ecosystem. Good compiler errors, good docs, decent forum coverage. Allocators merit some opinion; later.
The problem is just a walking simulator. Brute-forced part two, shamelessly.
December 6, 2024 at 7:44 PM
#AdventOfCode 2024
Day 6: Guard Gallivant
Actually, opted for Zig, not Nix for the next language. Nice language ecosystem. Good compiler errors, good docs, decent forum coverage. Allocators merit some opinion; later.
The problem is just a walking simulator. Brute-forced part two, shamelessly.
Day 6: Guard Gallivant
Actually, opted for Zig, not Nix for the next language. Nice language ecosystem. Good compiler errors, good docs, decent forum coverage. Allocators merit some opinion; later.
The problem is just a walking simulator. Brute-forced part two, shamelessly.
#AdventOfCode 2024
Day 5: Print Queue
Ok, I'm through with Nim. Usually, I'd take the whole month to get an opinion over the language, but I already have. Nim is not for me. I'll explain better below.
Today's problem is simple, if you implement a custom comparator. Smooth sailing.
Day 5: Print Queue
Ok, I'm through with Nim. Usually, I'd take the whole month to get an opinion over the language, but I already have. Nim is not for me. I'll explain better below.
Today's problem is simple, if you implement a custom comparator. Smooth sailing.
: bool {.closure.}): (seq[T], seq[T]) =
var split_index = 0
for i in 0..s.len()-1:
if s[i].len() == 0: split_index = i
result = (s[0..split_index-1], s[split_index+1..s.len()-1])
type compareResult = enum lessThan, equal, unknown
proc compareWithCycleGuard(lessThanTable: Table[int, seq[int]], a: int, b: int, cycleStop: int): compareResult =
if a == cycleStop: return unknown
if a == b: return equal
if not lessThanTable.hasKey(a): return unknown
if lessThanTable[a].filter( candidate => candidate == b).len() > 0: return lessThan
for candidate in lessThanTable[a]:
if lessThanTable.compareWithCycleGuard(candidate, b, cycleStop) == lessThan: return lessThan
return unknown
proc compare(lessThanTable: Table[int, seq[int]], a: int, b: int): compareResult =
if a == b: return equal
if not lessThanTable.hasKey(a): return unknown
if lessThanTable[a].filter( candidate => candidate == b).len() > 0: return lessThan
for candidate in lessThanTable[a]:
if lessThanTable.compareWithCycleGuard(candidate, b, candidate) == lessThan: return lessThan
return unknown
iterator combinations[T](s: openArray[T]): (T,T) =
for i in 0..s.len()-1:
for j in i+1..s.len()-1:
yield (s[i], s[j])
var (inputRules, inputPages) = stdin
.lines
.toSeq
.splitOn( line => line.len == 0 )
let lessThanTable =
inputRules
.map( line => line.split("|").map( parseInt ).toSeq)
.foldl(
(
var res = a
if a.hasKey(b[0]): res[b[0]].add(b[1]) else: res[b[0]] = @[b[1]]
res
)
, Table[int, seq[int]]()
)
let pages =
inputPages
.map( line => line.split(",").map( parseInt ))
echo(
pages
.filter(
line => line
.combinations
.toSeq
…](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreihfoybtgisgkemkkhcifbquugmj2lqa2ww44jnsw5hbz4lgavif2q@jpeg)
: bool {.closure.}): (seq[T], seq[T]) =
var split_index = 0
for i in 0..s.len()-1:
if s[i].len() == 0: split_index = i
result = (s[0..split_index-1], s[split_index+1..s.len()-1])
type compareResult = enum lessThan, equal, unknown
proc compareWithCycleGuard(lessThanTable: Table[int, seq[int]], a: int, b: int, cycleStop: int): compareResult =
if a == cycleStop: return unknown
if a == b: return equal
if not lessThanTable.hasKey(a): return unknown
if lessThanTable[a].filter( candidate => candidate == b).len() > 0: return lessThan
for candidate in lessThanTable[a]:
if lessThanTable.compareWithCycleGuard(candidate, b, cycleStop) == lessThan: return lessThan
return unknown
proc compare(lessThanTable: Table[int, seq[int]], a: int, b: int): compareResult =
if a == b: return equal
if not lessThanTable.hasKey(a): return unknown
if lessThanTable[a].filter( candidate => candidate == b).len() > 0: return lessThan
for candidate in lessThanTable[a]:
if lessThanTable.compareWithCycleGuard(candidate, b, candidate) == lessThan: return lessThan
return unknown
iterator combinations[T](s: openArray[T]): (T,T) =
for i in 0..s.len()-1:
for j in i+1..s.len()-1:
yield (s[i], s[j])
var (inputRules, inputPages) = stdin
.lines
.toSeq
.splitOn( line => line.len == 0 )
let lessThanTable =
inputRules
.map( line => line.split("|").map( parseInt ).toSeq)
.foldl(
(
var res = a
if a.hasKey(b[0]): res[b[0]].add(b[1]) else: res[b[0]] = @[b[1]]
res
)
, Table[int, seq[int]]()
)
let pages =
inputPages
.map( line => line.split(",").map( parseInt ))
echo(
pages
.filter(
line => line
.combinations
.toSeq…](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreigeup6jsls2p77tpra6gyuvaatlczbj663a67whjqt3mv33lpxrpm@jpeg)
December 5, 2024 at 4:29 PM
#AdventOfCode 2024
Day 5: Print Queue
Ok, I'm through with Nim. Usually, I'd take the whole month to get an opinion over the language, but I already have. Nim is not for me. I'll explain better below.
Today's problem is simple, if you implement a custom comparator. Smooth sailing.
Day 5: Print Queue
Ok, I'm through with Nim. Usually, I'd take the whole month to get an opinion over the language, but I already have. Nim is not for me. I'll explain better below.
Today's problem is simple, if you implement a custom comparator. Smooth sailing.
#AdventOfCode 2024
Day 4: Ceres Search
Took the time to write types for coord2d, vec2d and rectangle2d. For part 2, used that to match bounding boxes, a nice save.
Writing modules in Nim is poorly documented, and what forum responses I found were arrogant orders to read the poor docs 🤦♂️
Day 4: Ceres Search
Took the time to write types for coord2d, vec2d and rectangle2d. For part 2, used that to match bounding boxes, a nice save.
Writing modules in Nim is poorly documented, and what forum responses I found were arrogant orders to read the poor docs 🤦♂️
![import sugar
import sequtils
import system
import coord2d
import vec2d
import std/tables
proc contains_word(board: Table[Coord2D, char], word: string, pos: Coord2D, dir: Vec2D): bool =
if word.len == 0: return true
if not board.hasKey(pos) or board[pos] != word[0]: return false
return board.contains_word(word[1..word.len-1], pos+dir, dir)
var input: Table[Coord2D, char] =
stdin
.lines
.toSeq
.pairs
.toSeq
.map( line => (
line
.val
.pairs
.toSeq
.filter( ch => "XMAS".find(ch.val) != -1 )
.map( ch => ( Coord2D(x: ch.key, y: line.key), ch.val ))
))
.foldl( a.concat(b) )
.toTable
echo(
input
.pairs
.toSeq
.filter( pair => pair[1] == 'X')
.map( pair =>
directions
.toSeq
.map( direction => input.contains_word("XMAS", pair[0], direction ))
.filter( r => r)
)
.foldl( a.concat(b) )
.len
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreiclovq4ct4n64f4rbkqks3vaixbfwy6ntdpovyjotkwrfvzky7cdq@jpeg)
![import sugar
import sequtils
import system
import coord2d
import vec2d
import rectangle2d
import std/tables
proc contains_word(board: Table[Coord2D, char], word: string, pos: Coord2D, dir: Vec2D): bool =
if word.len == 0: return true
if not board.hasKey(pos) or board[pos] != word[0]: return false
return board.contains_word(word[1..word.len-1], pos+dir, dir)
var input: Table[Coord2D, char] =
stdin
.lines
.toSeq
.pairs
.toSeq
.map( line => (
line
.val
.pairs
.toSeq
.filter( ch => "MAS".find(ch.val) != -1 )
.map( ch => ( Coord2D(x: ch.key, y: line.key), ch.val ))
))
.foldl( a.concat(b) )
.toTable
echo(
input
.pairs
.toSeq
.filter( pair => pair[1] == 'M')
.map( pair =>
directions
.toSeq
.map( direction => (pair[0], direction, input.contains_word("MAS", pair[0], direction )) )
.filter( r => r[2])
.map( r => bounding_box(r[0], r[0] + ("MAS".len()-1) * r[1]))
)
.foldl( a.concat(b) )
.toSeq
.foldl( (
var result = a
if a.hasKey(b):
result.inc(b)
else:
result[b] = 1
result
), CountTable[Rectangle2D]())
.values
.toSeq
.filter( freq => freq == 2)
.len
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreifsivqthsgcvnv2xq3ul5uvmhaeucmaz2r7h7izb6wlqwepfkxmoa@jpeg)
December 4, 2024 at 5:50 PM
#AdventOfCode 2024
Day 4: Ceres Search
Took the time to write types for coord2d, vec2d and rectangle2d. For part 2, used that to match bounding boxes, a nice save.
Writing modules in Nim is poorly documented, and what forum responses I found were arrogant orders to read the poor docs 🤦♂️
Day 4: Ceres Search
Took the time to write types for coord2d, vec2d and rectangle2d. For part 2, used that to match bounding boxes, a nice save.
Writing modules in Nim is poorly documented, and what forum responses I found were arrogant orders to read the poor docs 🤦♂️
#AdventOfCode 2024
Day 3: Mull It Over
Regex day. Nim has decent support. No named groups, though. Unexpected, in such a recent language.
Did a nice approach to folding with state in part 2.
Day 3: Mull It Over
Regex day. Nim has decent support. No named groups, though. Unexpected, in such a recent language.
Did a nice approach to folding with state in part 2.
![import sugar
import sequtils
import strutils
import system
import std/re
var input =
stdin
.lines
.toSeq
var regex = re"mul\((\d+),(\d+)\)"
echo(
input
.map( line => line
.findAll(regex)
.toSeq()
.map( s => (
var matches: array[0..1, string]
discard s.match(regex, matches)
matches
.map(parseInt)
))
.map( val => val[0] * val[1])
.foldl( a + b )
)
.foldl( a + b )
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreidoosqspntybjshnjybv7gsaet5snqqbo4x2ih64bvmfsyteftbye@jpeg)
![import sugar
import sequtils
import strutils
import system
import std/re
var input =
stdin
.lines
.toSeq
var regex = re"mul\((\d+),(\d+)\)|do\(\)|don't\(\)"
echo(
input
.map( line => line
.findAll(regex)
.toSeq()
.map( s => (
case s:
of "do()":
(0, 1)
of "don't()":
(0, -1)
else:
var matches: array[0..1, string]
discard s.match(regex, matches)
(matches
.map(parseInt)
.foldl( a * b )
, 0)
))
)
.foldl( a.concat(b) )
.foldl(
case b[1]:
of 1:
(a[0], 1)
of -1:
(a[0], 0)
else:
case a[1]:
of 0:
a
else:
(a[0] + b[0], 1)
, (0,1)
)[0]
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreicu7dhknun3ufyqbknadmkfrvzlieucwjttk7mpcyktklqarlotl4@jpeg)
December 3, 2024 at 11:32 AM
#AdventOfCode 2024
Day 3: Mull It Over
Regex day. Nim has decent support. No named groups, though. Unexpected, in such a recent language.
Did a nice approach to folding with state in part 2.
Day 3: Mull It Over
Regex day. Nim has decent support. No named groups, though. Unexpected, in such a recent language.
Did a nice approach to folding with state in part 2.
#AdventOfCode 2024
Day 2: Red-Nosed Reports
Not much to report. Brute-forced part 2, the input was small enough and brute forcing would be trivial to write. Nim compiler errors continue to be entirely useless. I correct my mistakes by fumbling around.
Day 2: Red-Nosed Reports
Not much to report. Brute-forced part 2, the input was small enough and brute forcing would be trivial to write. Nim compiler errors continue to be entirely useless. I correct my mistakes by fumbling around.
![import sugar
import sequtils
import strutils
import system
proc seqDeltas(s: seq[int]): seq[int] =
var res: seq[int]
for index in 0..(s.len-1):
if index > 0:
res.add(s[index] - s[index-1])
return res
var input =
stdin
.lines
.toSeq
.map( line => line
.split(" ")
.map(part => part.strip())
.filter( part => part != "")
.map(parseInt)
)
echo(
input
.map( seqDeltas )
.map( deltaRow => (
deltaRow
.map( delta => (delta, abs(delta) > 0 and abs(delta) < 4))
.foldl( (b[0], a[1] and b[1] and a[0]/abs(a[0]) == b[0]/abs(b[0]) ) )
))
.filter( a => a[1] )
.len()
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreieganssdijuikcif5pt3sagbzhcf63r2tq5onqoz4nhkr6bshkopu@jpeg)
![import sugar
import sequtils
import strutils
import system
proc seqDeltas(s: seq[int]): seq[int] =
var res: seq[int]
for index in 0..(s.len-1):
if index > 0:
res.add(s[index] - s[index-1])
return res
var input =
stdin
.lines
.toSeq
.map( line => line
.split(" ")
.map(part => part.strip())
.filter( part => part != "")
.map(parseInt)
)
var deltas: seq[seq[seq[int]]]
for row in input:
var candidates: seq[seq[int]]
candidates.add(seqDeltas(row))
for i in 0..row.len-1:
var changedRow = row
changedRow.delete(i)
candidates.add(seqDeltas(changedRow))
deltas.add(candidates)
echo(
deltas
.map( puzzle => # seq[seq[int]]
puzzle.map( candidateDeltas => # seq[int]
candidateDeltas
.map( delta => (delta, abs(delta) > 0 and abs(delta) < 4))
.foldl( (b[0], a[1] and b[1] and a[0]/abs(a[0]) == b[0]/abs(b[0]) ) )
)
.foldl( a or b[1], false)
)
.filter( a => a )
.len()
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreidwfvctmlj2hdrzcorfams72uhdigbt24gmemljs7smclcftye4lq@jpeg)
December 2, 2024 at 1:02 PM
#AdventOfCode 2024
Day 2: Red-Nosed Reports
Not much to report. Brute-forced part 2, the input was small enough and brute forcing would be trivial to write. Nim compiler errors continue to be entirely useless. I correct my mistakes by fumbling around.
Day 2: Red-Nosed Reports
Not much to report. Brute-forced part 2, the input was small enough and brute forcing would be trivial to write. Nim compiler errors continue to be entirely useless. I correct my mistakes by fumbling around.
#AdventOfCode 2024
Day 1: Historian Hysteria
This year I'm trying my hand at Nim (nim-lang.org)
So far, I realized I'm spoiled by how good rustc error reporting is. Nim is like a C++ compiler on drugs. Thoroughly opaque for the uninitiated.
But hey, got it done...
github.com/sergiosgc/Ad...
Day 1: Historian Hysteria
This year I'm trying my hand at Nim (nim-lang.org)
So far, I realized I'm spoiled by how good rustc error reporting is. Nim is like a C++ compiler on drugs. Thoroughly opaque for the uninitiated.
But hey, got it done...
github.com/sergiosgc/Ad...
![import sugar
import sequtils
import strutils
import system
import std/algorithm
var (leftList, rightList) = stdin
.lines
.toSeq
.map( line => line
.split(" ", 2)
.map(part => part.strip())
.filter( part => part != "")
.map(parseInt)
)
.map( line => (line[0], line[1]) )
.unzip
leftList.sort()
rightList.sort()
echo(
leftList
.zip(rightList)
.toSeq
.map( row => ( abs(row[0] - row[1]) ) )
.foldl( a+b )
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreig7yxahkbr4tzbqztb3qrvlukgpmbxyheeefujmr5s2pcsipgisdy@jpeg)
![import sugar
import sequtils
import strutils
import system
var (leftList, rightList) = stdin
.lines
.toSeq
.map( line => line
.split(" ", 2)
.map( part => part.strip() )
.filter( part => part != "" )
.map(parseInt)
)
.map( line => (line[0], line[1]) )
.unzip
echo(
leftList
.map( left => (
left * rightList.filter( right => left == right ).len()
))
.foldl( a + b )
)](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:r5onmqjmslku7nkbfrdoeaoj/bafkreigtq6i3bng3p3mzdltxv77tmacfa7skcocx6dqemq4zxvsbsvpgd4@jpeg)
December 1, 2024 at 3:28 PM
#AdventOfCode 2024
Day 1: Historian Hysteria
This year I'm trying my hand at Nim (nim-lang.org)
So far, I realized I'm spoiled by how good rustc error reporting is. Nim is like a C++ compiler on drugs. Thoroughly opaque for the uninitiated.
But hey, got it done...
github.com/sergiosgc/Ad...
Day 1: Historian Hysteria
This year I'm trying my hand at Nim (nim-lang.org)
So far, I realized I'm spoiled by how good rustc error reporting is. Nim is like a C++ compiler on drugs. Thoroughly opaque for the uninitiated.
But hey, got it done...
github.com/sergiosgc/Ad...
Não é identificação que se quer. É associação. Mostras este a 100 pessoas, mais de 90 dizem Canadá como primeira palavra que lhes vem à cabeça.
Honestamente, alguma vez um círculo amarelo pode sequer sonhar chegar a esta posição top-of-mind? Faltam tantas características para estar sequer perto...
Honestamente, alguma vez um círculo amarelo pode sequer sonhar chegar a esta posição top-of-mind? Faltam tantas características para estar sequer perto...
April 7, 2024 at 2:04 PM
Não é identificação que se quer. É associação. Mostras este a 100 pessoas, mais de 90 dizem Canadá como primeira palavra que lhes vem à cabeça.
Honestamente, alguma vez um círculo amarelo pode sequer sonhar chegar a esta posição top-of-mind? Faltam tantas características para estar sequer perto...
Honestamente, alguma vez um círculo amarelo pode sequer sonhar chegar a esta posição top-of-mind? Faltam tantas características para estar sequer perto...
Outra análise comparativa que demonstra facilmente que a aproximação foi má é face aos nossos congéneres:
April 7, 2024 at 12:10 PM
Outra análise comparativa que demonstra facilmente que a aproximação foi má é face aos nossos congéneres:
Entretanto, na realidade
March 1, 2024 at 3:18 PM
Entretanto, na realidade
Estranho assumirem assim que têm falhas de segurança no isolamento de aplicações ao nível do sistema operativo...
February 16, 2024 at 10:45 AM
Estranho assumirem assim que têm falhas de segurança no isolamento de aplicações ao nível do sistema operativo...
8601 was worth it just for the xkcd.
February 6, 2024 at 1:05 AM
8601 was worth it just for the xkcd.
Esta proposta é a que Berlim usou, e funciona. No caso deles, foram aos 30% de oferta pública.
Atenção que não é subsidiar a oferta. É participar directamente no mercado.
Atenção que não é subsidiar a oferta. É participar directamente no mercado.
February 4, 2024 at 12:47 PM
Esta proposta é a que Berlim usou, e funciona. No caso deles, foram aos 30% de oferta pública.
Atenção que não é subsidiar a oferta. É participar directamente no mercado.
Atenção que não é subsidiar a oferta. É participar directamente no mercado.
If you see this, post your favourite fictional vehicle.
January 31, 2024 at 8:27 AM
If you see this, post your favourite fictional vehicle.
Dito na Assembleia da República:
January 27, 2024 at 1:56 AM
Dito na Assembleia da República:
Day 13: Point of Incidence
Some days I shouldn't touch code. Stepped into the mud with a wrong data structure for the problem, waded through mud onto the solution. Awful code, though it works. Part 2 is O(n^2), I know. I just don't care.
github.com/sergiosgc/Ad...
Some days I shouldn't touch code. Stepped into the mud with a wrong data structure for the problem, waded through mud onto the solution. Awful code, though it works. Part 2 is O(n^2), I know. I just don't care.
github.com/sergiosgc/Ad...
December 13, 2023 at 5:51 PM
Day 13: Point of Incidence
Some days I shouldn't touch code. Stepped into the mud with a wrong data structure for the problem, waded through mud onto the solution. Awful code, though it works. Part 2 is O(n^2), I know. I just don't care.
github.com/sergiosgc/Ad...
Some days I shouldn't touch code. Stepped into the mud with a wrong data structure for the problem, waded through mud onto the solution. Awful code, though it works. Part 2 is O(n^2), I know. I just don't care.
github.com/sergiosgc/Ad...