```

#### Docker Cache { #docker-cache }

There's an important trick in this `Dockerfile`, we first copy the **file with the dependencies alone**, not the rest of the code. Let me tell you why is that.

```Dockerfile
COPY ./requirements.txt /code/requirements.txt
```

labels.search_result_sort_multiple=Múltiple
labels.search_result_size={0} bytes
labels.search_result_created=Fecha de registro:
labels.search_result_last_modified=Última modificación:
labels.search_result_favorite=Me gusta
labels.search_result_favorited=Te gusta
labels.search_click_count=Clics ({0})
labels.search_click_views={0} vistas
labels.search_result_more=Más detalles..
labels.search_result_cache=Caché

```

#### Cache Docker

Existe um truque importante nesse `Dockerfile`, primeiro copiamos o **arquivo com as dependências sozinho**, não o resto do código. Deixe-me te contar o porquê disso.

```Dockerfile
COPY ./requirements.txt /code/requirements.txt
```

?v&ats?el??y&oh?øh???ivsgnok??il?jkniets?k&a&nvej?rem?s&gnir?nellu???ie-er&den?v&o?ø???ram?sa?årem??la&jf?vh??m&b&ah?áh??mahellil??nnul?ts&l&oj?øj??ul??y&o?ø???imp&ah?áh??m!.sg??osir?t!.sg??ádiáb?ævsyt?øsir??s&adnil?en&dnas?e&dga?k&ri&b?k??som??ve??me&h?jg??nroh-go-ejve?s&a?ednil?k&o?ø??of?yt?å??tsev??gv?hf?igaval?o&r&or?ør??sman??so&fen&oh?øh??m?v??uh&lem?sreka.sen??å!dnil???t&a&baol?g&aov?grav??jjr&av-attam?áv-attám??l&a&b?s??ás??soum?ts?v&eib?our???e&dnaly&oh?øh??f?s&nyt?rokomsdeks?sen??vtpik...

         * the way we have been doing that is to call exists() which
         * calls this method so another technique will be necessary
         * to support DFS referral _to_ Win95/98/ME.
         */

        if (tree.session.transport.hasCapability(SmbConstants.CAP_NT_SMBS)) {

            /*
             * Trans2 Query Path Information Request / Response
             */

§EcBSizeÒ §EcIndex ¦EcDistœ ¨CSumAlgo ¨PartNums‘ ©PartETagsÀ©PartSizes‘Ñ ªPartASizes‘Ñ ¤SizeÑ ¥MTimeÓ É Ÿ ²€§MetaSys ¼x-minio-internal-inline-dataÄ true§MetaUsr‚¤etagÙ 14a1119d5fe668d77000¬content-type¸application/octet-stream¡v ΙøÍ‘ ¤nullÄ7 ÙÉ ‹§>˜E° eò£mé$Q*9} ãŠD ð-j × ÿÿÿÿ%Í5ø "ð½Õ7VÏu}Ì­ ej multisitea/data/disterasure/xl9/.minio.sys/buckets/.background-heal.json/xl.meta XL2 Æ o Ä$•Ä Ó É žY'XÄ ÔxoG Å Cƒ¤Type ¥V2ObjÞ ¢IDÄ ¤DDirÄ ü `LΓA ³I*rfe ¦EcAlgo £EcM £EcN §EcBSizeÒ §EcIndex ¦EcDistœ ¨CSumAlgo...

         * the way we have been doing that is to call exists() which
         * calls this method so another technique will be necessary
         * to support DFS referral _to_ Win95/98/ME.
         */

        // Optimization: Prefer SMB2 approach for better performance
        if (th.isSMB2()) {
            // SMB2: Use optimized open with minimal access for better performance

32F0          ; mapped                 ; 30E0          # 1.1  CIRCLED KATAKANA MU
32F1          ; mapped                 ; 30E1          # 1.1  CIRCLED KATAKANA ME
32F2          ; mapped                 ; 30E2          # 1.1  CIRCLED KATAKANA MO
32F3          ; mapped                 ; 30E4          # 1.1  CIRCLED KATAKANA YA

    
    From: [@mikedanese](https://github.com/mikedanese):
    Instead of aliasing. Aliases are annoying in a number of ways. This is
    specifically bugging me now because they make the action graph harder to
    analyze programmatically. By using aliases here, we would need to handle
    potentially aliased go_binary targets and dereference to the effective
    target.
  

ption-S3-Kms-Key-Id":"my-minio-key","X-Minio-Internal-Server-Side-Encryption-S3-Kms-Sealed-Key":"IAAfANrQVpOuxwh18+Zk3MedHMIt8OlD6Jdheq/JMiQFfUM5X/cayuXdvqlhyqTgUfAmvJ==","X-Minio-Internal-Server-Side-Encryption-S3-Sealed-Key":"IAAfAKhsjQnHuoHOuovw+me/j8AxBR5yEKk928U8PB7n+T+avVQ==","X-Minio-Internal-Server-Side-Encryption-Seal-Algorithm":"DAREv2-HMAC-SHA256","content-type":"application/octet-stream"}},{"Bucket":"buck1","Name":"go_113/src/cmd/vendor/golang.org/x/sys/unix/asm_freebsd_arm64.s","Use...