Compose: Dockerfile become truncated in certain circumstances

[Ovsyanka]

Description of the issue

conditions

1. Dockerfile out of the context directory
2. Dockerfile has cyrillic symbols (or other non-latin symbols I suppose)

result:

On docker-compose build Dockerfile truncated by amount of cyrillic symbols in it.

For example, if I got 7 cyrillic symbols in the Dockerfile last 7 symbols of the end of file become "cutted off".

Context information (for bug reports)

Output of docker-compose version

docker-compose version 1.25.0, build unknown
docker-py version: 4.1.0
CPython version: 3.8.1
OpenSSL version: OpenSSL 1.1.1d  10 Sep 2019

Output of docker version

Client:
 Version:           19.03.5-ce
 API version:       1.40
 Go version:        go1.13.4
 Git commit:        633a0ea838
 Built:             Fri Nov 15 03:19:09 2019
 OS/Arch:           linux/amd64
 Experimental:      false

Server:
 Engine:
  Version:          19.03.5-ce
  API version:      1.40 (minimum version 1.12)
  Go version:       go1.13.4
  Git commit:       633a0ea838
  Built:            Fri Nov 15 03:17:51 2019
  OS/Arch:          linux/amd64
  Experimental:     false
 containerd:
  Version:          v1.3.2.m
  GitCommit:        d50db0a42053864a270f648048f9a8b4f24eced3.m
 runc:
  Version:          1.0.0-rc9
  GitCommit:        d736ef14f0288d6993a1845745d6756cfc9ddd5a
 docker-init:
  Version:          0.18.0
  GitCommit:        fec3683

Output of docker-compose config

see in the "steps to reproduce"

Steps to reproduce the issue

Create context (empty) directory, docker-compose.yml and Dockerfile

docker-compose.yml

version: '3.7'

services:
  test-pass:
    build:
      context: .
      dockerfile: Dockerfile
  test-fail:
    build:
      context: context
      # path relative to the context
      dockerfile: ../Dockerfile

Dockerfile

FROM alpine

# шшш и ффф

RUN touch test

Run docker-compose build test-fail

Observed result

docker-compose build test-fail
Building test-fail
Step 1/2 : FROM alpine
 ---> cdf98d1859c1
Step 2/2 : RUN touc
 ---> Running in fc83cc1f2b69
/bin/sh: touc: not found
ERROR: Service 'test-fail' failed to build: The command '/bin/sh -c touc' returned a non-zero code: 127

The string h test\n (7 symbols) was disappeared.

This problem doesn't appear if the Dockerfile is in the context folder (look at the test-pass service).

This problem doesn't appear by using docker build.

Expected result

Expected result is succesfully created image, like by using docker build:

docker build -f Dockerfile context
Sending build context to Docker daemon  1.583kB
Step 1/2 : FROM alpine
 ---> cdf98d1859c1
Step 2/2 : RUN touch test
 ---> Using cache
 ---> bb6362bff392
Successfully built bb6362bff392

Stacktrace / full error message

N/A

[ndeloof]

thanks for reporting this issue with clear reproduction sample and initial investigation on breaking conditions. I'll investigate this issue next week

[ndeloof]

As a workaround, you can enable native CLI builder by setting COMPOSE_DOCKER_CLI_BUILD=true (I expect this to become default value in a future release)

[ndeloof]

issue (afaict) is caused by https://github.com/docker/docker-py/blob/master/docker/utils/build.py#L107-L109

tar size is computed as len(contents) == character counts, while the actual tar entry is set by contents.encode('utf-8'), which will then be truncated as some characters will be encoded on N>1 bytes.